1 2011-01-13 00:16:47 lzsaver has joined
2 2011-01-13 00:22:15 lzsaver has left ()
3 2011-01-13 00:27:53 <hacim> anyone seen stats for a nvidia Geforce fx5700? its a bit old, so maybe not so useful
4 2011-01-13 00:28:19 <ArtForzZz> iirc < 8xxx series isnt supported at all
5 2011-01-13 00:29:08 darrob has quit (Remote host closed the connection)
6 2011-01-13 00:29:31 <ArtForzZz> and even if it were, it'd be slow as fuck
7 2011-01-13 00:31:02 darrob has joined
8 2011-01-13 00:31:30 <ArtForzZz> 4 pixel shaders @ 425 MHz... about 256 khash/second
9 2011-01-13 00:33:17 <ArtForzZz> so about a factor of 2000 slower than a modern high-end GPU
10 2011-01-13 00:43:39 <lfm> art is correct gforce older than 8000 series is not supported for cuda/opencl
11 2011-01-13 00:44:10 <Diablo-D3> yes, its not
12 2011-01-13 00:45:51 <ArtForzZz> well, you can do GPGPU without a GPGPU toolkit using plain ole pixel shaders
13 2011-01-13 00:46:06 TD has joined
14 2011-01-13 00:46:10 <ArtForzZz> but it's really pointless as old hardware is sloooow
15 2011-01-13 00:46:43 <lfm> hed have to write his own sofware and such projects are very tricky
16 2011-01-13 00:47:22 <Diablo-D3> lfm: no
17 2011-01-13 00:47:27 <Diablo-D3> the hardware is incapable of doing it
18 2011-01-13 00:47:42 <ArtForzZz> errr.. no
19 2011-01-13 00:48:19 <ArtForzZz> whoops, yes
20 2011-01-13 00:48:24 <lfm> diablo isnt there some tricks with pixel shaders and stuff that people used to do before cuda?
21 2011-01-13 00:48:39 <ArtForzZz> yep
22 2011-01-13 00:48:52 <Diablo-D3> lfm: not anything useful here
23 2011-01-13 00:49:09 <ArtForzZz> iirc you still need GL2.x capability
24 2011-01-13 00:50:23 <lfm> well ya, pretty pointless. if he is really despirate and short of money he could get a gforce 8200 or 8400 for like $30 and play that way, still kinda pointless if he actually wants to make a profit but he might learn some of the concepts
25 2011-01-13 00:51:01 <ArtForzZz> I have the slight feeling a used HD5450 might be better value :P
26 2011-01-13 00:51:18 <lfm> ya ok
27 2011-01-13 00:51:34 <ArtForzZz> "only" about as fast as a modern high end CPU
28 2011-01-13 00:51:45 <ArtForzZz> but thats not really bad considering it's a 20W card
29 2011-01-13 00:52:01 <lfm> about the same as a 8400 then actually
30 2011-01-13 00:52:51 <ArtForzZz> maybe even a bit lower
31 2011-01-13 00:53:05 <ArtForzZz> and it should be significantly faster for 3D, too
32 2011-01-13 00:53:34 <ArtForzZz> kind of a problem if you're stuck on AGP though
33 2011-01-13 00:53:57 * sipa wonders why his perfectly reachable and connected node isn't shown on http://maps.google.com/maps?q=https://smsz.net/btcStats/bitcoin.kml
34 2011-01-13 00:54:06 <lfm> I clocked a 8400 gs at 1500 khash/s and I think its like 15 watss, no fan
35 2011-01-13 00:54:11 <lfm> watts
36 2011-01-13 00:54:36 <ArtForzZz> a 5450 should do ~12Mh/s
37 2011-01-13 00:55:25 <davex__> is this bs, ati radeon 5970 can do 580,000 khash?
38 2011-01-13 00:55:27 <davex__> http://174.143.149.98/wiki/doku.php?id=bitcoin_miners
39 2011-01-13 00:55:45 <lfm> davex for $35?
40 2011-01-13 00:56:08 <ArtForzZz> slightly OCed, yep
41 2011-01-13 00:56:27 <ArtForzZz> most 5970s can hit 600Mhash no problem
42 2011-01-13 00:56:59 <ArtForzZz> I'm averaging ~625Mhash per 5970
43 2011-01-13 00:56:59 <davex__> hmm. guess they're $700, though.
44 2011-01-13 00:57:07 <ArtForzZz> nowadays, yep
45 2011-01-13 00:57:18 <ArtForzZz> they used to be $500 2 months or so ago
46 2011-01-13 00:57:23 <davex__> haha
47 2011-01-13 00:57:29 <xelister> bitcoin power
48 2011-01-13 00:57:33 <ArtForzZz> EOL power
49 2011-01-13 00:57:43 <xelister> +bitcoin power
50 2011-01-13 00:58:05 <ArtForzZz> the GSM unlocking guys also love 5970s ;)
51 2011-01-13 00:58:13 <lfm> soon have a 6970 to see what price that is
52 2011-01-13 00:58:16 <xelister> they are not like, DDR 1 chips that you just /have/ to have to fix old machines that are "does not breake - do not fix it" for years
53 2011-01-13 00:58:39 <xelister> ArtForzZz: and TI calculators, I found on freenet
54 2011-01-13 00:59:06 <ArtForzZz> err, we already have a 6970, dual-GPU 6xxx is 6990
55 2011-01-13 00:59:19 <xelister> radeons would be my favorite gfx card ever, if not they uber-ghay software drivers shit crap bullshit big piece of garbage
56 2011-01-13 00:59:23 <ArtForzZz> and it's supposed to be released in february
57 2011-01-13 00:59:27 <ArtForzZz> yep
58 2011-01-13 00:59:38 <ArtForzZz> fglrx sucks donkey cock
59 2011-01-13 00:59:51 <xelister> I have a CHK@ for that, ArtForzZz ;)
60 2011-01-13 01:00:12 <luke-jr> tcatm: Assuming X8 by default is terrible. Your change in fact does something *totally* out of spec for hexadecimal amounts (where Xn is 16^n).
61 2011-01-13 01:00:21 <ArtForzZz> rule 34 also applies to drivers?
62 2011-01-13 01:00:25 <luke-jr> tcatm: I modified it to default to X8 for decimal, and (true) X4 for hexadecimal.
63 2011-01-13 01:01:47 <xelister> I wonder will they ever release an upgrade to drivers, that remains compatible with ubuntu, fixes the bugs esp from VT switching (and I discovered 1 new type of bug, this one forces always HARD sysrq-B RESET, THANKS)
64 2011-01-13 01:02:01 <ArtForzZz> unlikely
65 2011-01-13 01:02:06 <xelister> while also working fine with OpenCL (low cpu usage, fine with 2 cpus)
66 2011-01-13 01:02:11 <xelister> gpus
67 2011-01-13 01:02:23 <ArtForzZz> actually losing OpenCL helps a lot there
68 2011-01-13 01:02:23 <xelister> ArtForzZz: you think they will not fix their shit, not even in like a month or two?
69 2011-01-13 01:02:33 <ArtForzZz> yep
70 2011-01-13 01:02:39 <xelister> why :[
71 2011-01-13 01:02:51 <xelister> Amd took over, they should introduce quality'n'shite
72 2011-01-13 01:03:18 <ArtForzZz> iirc their plan is "release specs, let the community figure it out"
73 2011-01-13 01:03:34 <xelister> just specs? new driver will take like a yaer
74 2011-01-13 01:03:45 <xelister> wtf why not release source of current driver?!
75 2011-01-13 01:04:08 <ArtForzZz> 3rd party IP
76 2011-01-13 01:04:25 <ArtForzZz> I think the open radeon -dev tree already has prelim support for 68xx
77 2011-01-13 01:04:33 <lfm> to embarassed to let people see shit code
78 2011-01-13 01:04:40 <ArtForzZz> yep
79 2011-01-13 01:04:41 <xelister> are they stupid, they don't see how GPU computing community would love them for that? hint hint, scientists/hackers/security experts/advanced programming/math guys usually do not use crap like windows or mac
80 2011-01-13 01:05:00 <ArtForzZz> real coders use CAL IL
81 2011-01-13 01:05:07 <xelister> :O
82 2011-01-13 01:05:14 <xelister> FUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
83 2011-01-13 01:05:38 <xelister> ArtForzZz: what third party IP.. they bought Ati
84 2011-01-13 01:05:51 <xelister> also what, they can sell but can not sell the code to like FOSS ?
85 2011-01-13 01:05:56 <ArtForzZz> and ATI licensed that crap from other companies
86 2011-01-13 01:05:57 <xelister> or to FSF or smth
87 2011-01-13 01:06:13 <ArtForzZz> also probably fear of patent suits
88 2011-01-13 01:06:19 <xelister> they need a serious bitchslap in their fucking head to come to their senses
89 2011-01-13 01:06:47 <tcatm> luke-jr: okay, but from now in the spec won't change anymore
90 2011-01-13 01:06:47 <xelister> in the end, they will loose money for not publishing drviers that would open way to be loved <3 obvious choice amongst all cluster computing users
91 2011-01-13 01:06:49 <ArtForzZz> 3D rendering is a patent minefield nearly as bad as video encoding
92 2011-01-13 01:07:12 <xelister> well then just release it in EU or something, USA can suck our collective dick
93 2011-01-13 01:07:16 <lfm> ati was hardware outfit, their drivers always sucked even since ati wonder bios where I had to reset 5 times to get it to start
94 2011-01-13 01:07:20 <luke-jr> tcatm: I'm making one final change I think
95 2011-01-13 01:07:47 <xelister> luke-jr: haha yea I rememver All-In-Wonder cards. they REALLY MADE ME WONDER how such crap can crash my game each 15 minutes
96 2011-01-13 01:07:58 <luke-jr> tcatm: I am modifying it such that X<digits> is always required in the URI, but that implementations should handle the X8/X4 defaults
97 2011-01-13 01:08:19 <xelister> only good thing from Ati, is that we can say, that linux driver support can be actually BETTER then on windows ;)
98 2011-01-13 01:08:45 <tcatm> luke-jr: does that change the code for parseAmount?
99 2011-01-13 01:09:04 <ArtForzZz> ?
100 2011-01-13 01:09:31 <ArtForzZz> their windows drivers more-or-less work for accelerated video decoding
101 2011-01-13 01:09:31 <luke-jr> tcatm: not since the X4 change, no
102 2011-01-13 01:09:41 <tcatm> luke-jr: ok
103 2011-01-13 01:09:42 <ArtForzZz> which can't be said of fglrx
104 2011-01-13 01:09:43 <luke-jr> m[9] ? Math.pow(16, parseInt(m[9], 16)) : 0x10000
105 2011-01-13 01:09:47 <luke-jr> that is the last changed line
106 2011-01-13 01:09:57 <xelister> wikileaks should just leak the code that may or may not be the actuall code from Ati, and Ati should release statement they will not sue for using it no matter is it their code or not ... then probably they are off the hook?
107 2011-01-13 01:10:16 <tcatm> luke-jr: https://github.com/tcatm/bitcoin-js-remote/commit/c27a3149aeaa5038e28ecd1d09310dc70587aea7
108 2011-01-13 01:10:20 <luke-jr> ArtForzZz: the free drivers are almost supporting video decoding and even DirectX
109 2011-01-13 01:10:28 <xelister> "it's not our code that is infiringing imaginary patents - it is some anon's driver"
110 2011-01-13 01:10:56 <ArtForzZz> wtf does "almost" supporting video decoding do?
111 2011-01-13 01:11:27 <luke-jr> ArtForzZz: it's a WIP, and should work soon ;P
112 2011-01-13 01:11:36 <ArtForzZz> wonder how well that'll work
113 2011-01-13 01:11:57 <ArtForzZz> iirc the problem is AMd can't realse docs on UVD because it's pretty tightly coupled to HDCP crap
114 2011-01-13 01:11:58 <lfm> means it wont lock up more than once per hour?
115 2011-01-13 01:12:23 <luke-jr> oh
116 2011-01-13 01:12:28 <luke-jr> I think it's more along the lines of vdpau
117 2011-01-13 01:12:43 <lfm> sigh, oh ya hdcp, thatd fould em up
118 2011-01-13 01:12:51 <luke-jr> tcatm: it would be much easier to contribute if you use gitorious or something free âº
119 2011-01-13 01:12:59 <ArtForzZz> I think I read about the effort, dont they pretty much emulate that using shaders?
120 2011-01-13 01:13:14 <luke-jr> tcatm: do you need my help with fonts?
121 2011-01-13 01:13:19 <tcatm> luke-jr: fonts?
122 2011-01-13 01:13:24 <ArtForzZz> = once again, DRM crazyness annoying consumers
123 2011-01-13 01:13:34 <luke-jr> tcatm: yeah, you probably need a webfont or something, for most phones
124 2011-01-13 01:13:37 <xelister> heh HDCP is shite
125 2011-01-13 01:13:44 <ArtForzZz> not to mention HDCP is completely. fucking. pointless.
126 2011-01-13 01:13:49 <tcatm> luke-jr: on android it looks fine
127 2011-01-13 01:13:54 <xelister> it makes me regret buying any original and never do that mistake again
128 2011-01-13 01:13:55 <luke-jr> tcatm: tried a TBC amount?
129 2011-01-13 01:13:58 <lfm> digit restrictions mismanagement
130 2011-01-13 01:14:01 <xelister> I was like... should I buy or pirate?
131 2011-01-13 01:14:07 <tcatm> luke-jr: yep
132 2011-01-13 01:14:22 <xelister> hmm.. rememer the last time I played LEGAL MOVIE and was forced to watch like 10 commercialls with no-skip? fuuuuuuuuu
133 2011-01-13 01:14:25 <luke-jr> tcatm: (39*4194304).formatBitcoin()
134 2011-01-13 01:14:30 <ArtForzZz> most fun part: HDCP doesnt even stop someone from capturing the uncompressed stream
135 2011-01-13 01:14:41 <luke-jr> tcatm: try amount=163577856X0
136 2011-01-13 01:14:58 <tcatm> luke-jr: Oh, i'm not using that formatBitcoin yet as it would need a little more work. I'm converting everything to BTC for now
137 2011-01-13 01:15:03 <ArtForzZz> so instead of grabbing the DVI pairs between computer and display I grab the *same* data off the LVDS link between display controller board and panel
138 2011-01-13 01:15:09 <luke-jr> tcatm: k, that avoids the font issue :p
139 2011-01-13 01:15:21 <xelister> ArtForzZz: btw the keys for HDCP are long-leaked
140 2011-01-13 01:15:26 <xelister> as you probbl know
141 2011-01-13 01:15:28 <ArtForzZz> yep
142 2011-01-13 01:15:36 <ArtForzZz> but even without that it's fucking pointless
143 2011-01-13 01:15:46 <lfm> art that voids your warantee! oh dear!
144 2011-01-13 01:15:51 <ArtForzZz> they built a locked gate without a fence
145 2011-01-13 01:16:02 <xelister> ;)
146 2011-01-13 01:16:24 <ArtForzZz> yeah, the lock is now broken, but there never was a fucking fence
147 2011-01-13 01:16:25 <tcatm> luke-jr: can you develop iPhone apps?
148 2011-01-13 01:16:30 <xelister> how exactly would you catch it without the keys?
149 2011-01-13 01:16:39 <luke-jr> tcatm: no idea, no interest since you can't write free apps for iPhone
150 2011-01-13 01:16:44 <lfm> xelister inside the TV
151 2011-01-13 01:17:03 <ArtForzZz> well, you always have the analog hole
152 2011-01-13 01:17:09 <xelister> lfm: so inside TV there is a blackbox decoder, and from it clear line path in LVDS format to display?
153 2011-01-13 01:17:20 <ArtForzZz> yep
154 2011-01-13 01:17:51 <luke-jr> but you can't rip it compressed that way
155 2011-01-13 01:17:53 <lfm> prolly even has a dip plug or edge connector to tap into
156 2011-01-13 01:17:55 <xelister> ArtForzZz: all anti pirace is stupid, eventually I can capture it off my fucking eye - like stick HDTV camera in front of panel (and dont give crap about lost details)
157 2011-01-13 01:17:56 <ArtForzZz> clock, 4 pairs each R/G/B pixel data, h/vsync
158 2011-01-13 01:17:58 <luke-jr> you have to recompress it, which is usually lossy
159 2011-01-13 01:18:10 <ArtForzZz> yes
160 2011-01-13 01:18:21 <lfm> luke rip[ it uncompressed do your own compression
161 2011-01-13 01:18:25 <ArtForzZz> but thats the thing, HDCP protects the uncompressed stream
162 2011-01-13 01:18:30 <Diablo-D3> [08:10:36] <ArtForzZz> most fun part: HDCP doesnt even stop someone from capturing the uncompressed stream
163 2011-01-13 01:18:39 <Diablo-D3> it also doesnt stop them from unecrypting it
164 2011-01-13 01:18:54 <xelister> HDCP doesn't stop from anything
165 2011-01-13 01:18:59 <ArtForzZz> thats the point
166 2011-01-13 01:19:03 <Diablo-D3> no I mean
167 2011-01-13 01:19:07 <ArtForzZz> it's DRM without purpose
168 2011-01-13 01:19:07 <Diablo-D3> HDCP was broken
169 2011-01-13 01:19:19 <xelister> well, perhaps from cops to say OHHH NOO that PEDORIST is haxoring the copyright, protect them, fire!
170 2011-01-13 01:19:21 TD has quit (Quit: TD)
171 2011-01-13 01:19:22 <luke-jr> tcatm: can we kill x-btc yet? :D
172 2011-01-13 01:19:22 <ArtForzZz> HDCPs DESIGN was already broken
173 2011-01-13 01:19:26 <Diablo-D3> ArtForzZz: well
174 2011-01-13 01:19:32 <Diablo-D3> COPYING the stream is fine and all
175 2011-01-13 01:19:43 <tcatm> luke-jr: We'll simple not use it :)
176 2011-01-13 01:19:49 <ArtForzZz> actually thats what HDCP was supposed to protect against
177 2011-01-13 01:19:50 <xelister> ArtForzZz: it though stops technicall illeterates from coping. They can not even set VCR clock
178 2011-01-13 01:19:50 <Diablo-D3> but who wants to download a torrent of 9000gb of video?
179 2011-01-13 01:20:09 <Diablo-D3> it was supposed to stop digital ripping out of the DVI jack
180 2011-01-13 01:20:17 <ArtForzZz> honestly, the loss isn't that bad
181 2011-01-13 01:20:23 <xelister> but not warez group, so it leaks to torrent, and there almos everyone can get it
182 2011-01-13 01:20:31 <xelister> btw I know few cops, they pirate warez like hell
183 2011-01-13 01:20:37 <ArtForzZz> only thing you lose over "normal" recoding is that you do 2 color space conversions
184 2011-01-13 01:20:41 <xelister> stupid dead laws, they should be removed
185 2011-01-13 01:20:51 <lfm> ya blue ray keys that they cant protect either
186 2011-01-13 01:20:53 <Diablo-D3> ArtForzZz: HDCP is lossless though
187 2011-01-13 01:20:57 <Diablo-D3> well
188 2011-01-13 01:20:58 <Diablo-D3> generally
189 2011-01-13 01:21:03 <Diablo-D3> DRM is fucked
190 2011-01-13 01:21:05 <ArtForzZz> so?
191 2011-01-13 01:21:08 <Diablo-D3> sony _almost_ had it
192 2011-01-13 01:21:10 <Diablo-D3> on the ps3
193 2011-01-13 01:21:20 <Diablo-D3> but why the fuck build a hypervisor that doesnt employ nx?
194 2011-01-13 01:21:21 <xelister> a robbery in MY COPYRIGHT PRIV KEYS??? more likelly then you think
195 2011-01-13 01:21:24 <ArtForzZz> not really
196 2011-01-13 01:21:32 <ArtForzZz> locked down hardware != DRM
197 2011-01-13 01:21:36 <brocktice> Here's a problem:
198 2011-01-13 01:21:48 <Diablo-D3> and why the fuck use ECDSA keys where the random nonce is used more than once?
199 2011-01-13 01:21:52 <ArtForzZz> it's the reverse of DRM
200 2011-01-13 01:22:02 <brocktice> how do you securely encrypt something, but ship millions of devices that can decrypt it without giving away the secret key?
201 2011-01-13 01:22:12 <Diablo-D3> brocktice: burn it to hardware
202 2011-01-13 01:22:14 <ArtForzZz> you can't
203 2011-01-13 01:22:19 <brocktice> exactly
204 2011-01-13 01:22:20 <Diablo-D3> which IBM did.
205 2011-01-13 01:22:26 <Diablo-D3> the problem is
206 2011-01-13 01:22:30 <brocktice> They're fucked before they even start.
207 2011-01-13 01:22:32 <Diablo-D3> the software part of the chain could be violated
208 2011-01-13 01:22:47 <ArtForzZz> and given enough time and resources you could probably crack a IBM crypto module, too
209 2011-01-13 01:22:59 <Diablo-D3> the _only_ thing that fucked sony was their security hypervisor allowed bugs to be used
210 2011-01-13 01:23:00 <brocktice> Oh sure
211 2011-01-13 01:23:11 <brocktice> Diablo-D3: The only thing that fucked them to this point.
212 2011-01-13 01:23:13 <lfm> they can only try to hide keys deep in firmware/hardware and hope no one figures out how to get at it too soon
213 2011-01-13 01:23:18 <ArtForzZz> yep
214 2011-01-13 01:23:23 <brocktice> If that hadn't been discovered, there would have been something else.
215 2011-01-13 01:23:31 <Diablo-D3> brocktice: yes, but no one would have known they had broken ECDSA keys without it
216 2011-01-13 01:23:36 <brocktice> caveat: I'm not a sony ps3 DRM expert.
217 2011-01-13 01:23:40 <Diablo-D3> they used a buffer overflow exploit to inject their own code in
218 2011-01-13 01:23:46 <Diablo-D3> NX would have stopped that dead.
219 2011-01-13 01:23:48 <brocktice> The DRM just keeps it from using unsigned software right?
220 2011-01-13 01:23:52 <Diablo-D3> yes
221 2011-01-13 01:23:58 <brocktice> That's different from trying to encrypt media
222 2011-01-13 01:23:59 <Diablo-D3> its just a security signature
223 2011-01-13 01:24:02 <ArtForzZz> yep
224 2011-01-13 01:24:03 <Diablo-D3> not true, brocktice
225 2011-01-13 01:24:04 <lfm> prolly no one would have tried to break ps3 if they had keept supporting linux on ps3
226 2011-01-13 01:24:08 <ArtForzZz> it's DRM-in-reverse
227 2011-01-13 01:24:08 <brocktice> Yeah much like trying to root android phones
228 2011-01-13 01:24:25 <brocktice> The stock bootloaders are finicky about signatures on updates.
229 2011-01-13 01:24:39 <brocktice> But, people find a way around.
230 2011-01-13 01:24:44 <Diablo-D3> the software part of the bootloader chain is AES encrypted
231 2011-01-13 01:24:57 <Diablo-D3> the IBM bootrom has the key to unlock the first stage of the bootloader
232 2011-01-13 01:25:18 <Diablo-D3> and then the first stage unlocks the second stage, and the second stage can insert and remove keys as needed, via updates
233 2011-01-13 01:25:20 <brocktice> Well, it did buy them a few years, right?
234 2011-01-13 01:25:26 <Diablo-D3> thats all they have to do
235 2011-01-13 01:25:28 <brocktice> that was probably all they expected.
236 2011-01-13 01:25:29 <Diablo-D3> is buy a few years
237 2011-01-13 01:25:36 <ArtForzZz> so it takes a while to find someone with too much time and access to a FIB station
238 2011-01-13 01:25:46 <Diablo-D3> the number of years they need to buy needs to exceed the numbers of years they intend on selling it
239 2011-01-13 01:25:59 <Diablo-D3> which if the PS2 was any indication, it would have been ten years
240 2011-01-13 01:25:59 <ArtForzZz> you can't securely store anything in a chip really
241 2011-01-13 01:26:05 <Diablo-D3> ArtForzZz: no
242 2011-01-13 01:26:07 <Diablo-D3> the chip didnt fail
243 2011-01-13 01:26:10 <Diablo-D3> the software did.
244 2011-01-13 01:26:11 <ArtForzZz> I know
245 2011-01-13 01:26:16 <xelister> ArtForzZz: even trusted computing chip got facked
246 2011-01-13 01:26:19 <Diablo-D3> so Im not going to fault IBM here
247 2011-01-13 01:26:22 <ArtForzZz> but guess how a lot of sat smart cards get cracked?
248 2011-01-13 01:26:31 <Diablo-D3> well, they start attacking the chips there
249 2011-01-13 01:26:34 <ArtForzZz> yep
250 2011-01-13 01:26:37 <Diablo-D3> IBM however knows how to build this shit properly
251 2011-01-13 01:26:39 * brocktice still uses his PS2
252 2011-01-13 01:26:46 <Diablo-D3> they have employed similar shit on milspec harware
253 2011-01-13 01:26:47 <luke-jr> what happens when they figure out how to combine verify/decrypt?
254 2011-01-13 01:26:48 <brocktice> I see no need for a PS3
255 2011-01-13 01:26:49 <Diablo-D3> its generally shit proof
256 2011-01-13 01:26:50 <ArtForzZz> you can't protect a chip alone from offline attacks
257 2011-01-13 01:26:56 <Diablo-D3> luke-jr: er, they do
258 2011-01-13 01:26:58 <ArtForzZz> it's physically impossible
259 2011-01-13 01:27:10 <Diablo-D3> ArtForzZz: you "can"
260 2011-01-13 01:27:11 <xelister> ArtForzZz: it can scream for help
261 2011-01-13 01:27:12 <ArtForzZz> you NEED a independent power source
262 2011-01-13 01:27:20 <brocktice> ArtForzZz: unless you build some kind of mechanical self-destruct.
263 2011-01-13 01:27:21 <Diablo-D3> one arcade machine did it for the lulz
264 2011-01-13 01:27:25 <ArtForzZz> how can you scream when you have ... no power?
265 2011-01-13 01:27:26 <brocktice> But even that could be protected maybe
266 2011-01-13 01:27:28 <xelister> ArtForzZz: it can scream for help using micro atomic powerplant
267 2011-01-13 01:27:31 <brocktice> err, worked around
268 2011-01-13 01:27:32 <Diablo-D3> they had an encrypted rom chip set
269 2011-01-13 01:27:39 <luke-jr> Diablo-D3: key failure = meltdown?
270 2011-01-13 01:27:42 <xelister> or, just
271 2011-01-13 01:27:48 <xelister> give 1 guard with each device
272 2011-01-13 01:27:50 <Diablo-D3> the key was stored in _ram_ on the cpu
273 2011-01-13 01:27:53 <ArtForzZz> which is exactly what the ibm crypto modules do, they have their own internal battery
274 2011-01-13 01:27:54 <Diablo-D3> the cpu had a built in battery
275 2011-01-13 01:27:55 <luke-jr> IIRC any recent SIM card destroys itself if you try to clone it
276 2011-01-13 01:27:58 <xelister> want to watch TV @ home -> guard comes in with the TV
277 2011-01-13 01:28:02 <Diablo-D3> unplug the chip
278 2011-01-13 01:28:05 <xelister> commencing entertainment time!
279 2011-01-13 01:28:06 <Diablo-D3> the battery is also unplugged
280 2011-01-13 01:28:18 <ArtForzZz> yes, they try to play games with light-sensitive transistors on modern chips to deter decapping
281 2011-01-13 01:28:19 <Diablo-D3> its a very old trick
282 2011-01-13 01:28:28 <Diablo-D3> theres tons of fucking shit to do
283 2011-01-13 01:28:32 <ArtForzZz> yep
284 2011-01-13 01:28:38 <xelister> but eventually the hacking wins
285 2011-01-13 01:28:40 <xelister> ?
286 2011-01-13 01:28:42 <ArtForzZz> and tons of ways to bypass it
287 2011-01-13 01:28:42 <Diablo-D3> the IBM cells work well here because the initial part of the chain is on the chip
288 2011-01-13 01:28:43 <Diablo-D3> as in
289 2011-01-13 01:28:48 <Diablo-D3> you can no longer boot the chip at all
290 2011-01-13 01:28:59 <Diablo-D3> it doesnt know what to do when you turn it on
291 2011-01-13 01:29:17 <Diablo-D3> it'd be like starting a normal x86 box with a blank bios
292 2011-01-13 01:29:20 <ArtForzZz> iirc the cell also has the "keys to the kingdom" in ROM
293 2011-01-13 01:29:27 <xelister> you can have chip auto-destroy when it looses power reserves?
294 2011-01-13 01:29:34 <Diablo-D3> ArtForzZz: no
295 2011-01-13 01:29:40 <Diablo-D3> only do the first stage bootloader
296 2011-01-13 01:29:46 <ArtForzZz> so?
297 2011-01-13 01:29:49 <Diablo-D3> *to
298 2011-01-13 01:29:56 <ArtForzZz> aaand?
299 2011-01-13 01:30:05 <Diablo-D3> and the bootloader cant read the keys from the cpu.
300 2011-01-13 01:30:11 <ArtForzZz> huh?
301 2011-01-13 01:30:27 <Diablo-D3> theres no way to access the keys inside the cell from anything.
302 2011-01-13 01:30:57 <lfm> hardware debuggers?
303 2011-01-13 01:31:01 <Diablo-D3> lfm: nope
304 2011-01-13 01:31:04 <ArtForzZz> lower level
305 2011-01-13 01:31:26 <Diablo-D3> lfm: the cell also prevents you from tampering with sony's security hypervisor that runs in the one SPE
306 2011-01-13 01:31:31 <Diablo-D3> you cant hardware debug it
307 2011-01-13 01:31:36 ybit has joined
308 2011-01-13 01:32:06 <ArtForzZz> if someone can build it, someone can take it apart
309 2011-01-13 01:32:13 <Diablo-D3> the problem is
310 2011-01-13 01:32:21 <Diablo-D3> we no longer have to crack the cell
311 2011-01-13 01:32:25 <Diablo-D3> the bootloader chain is broken
312 2011-01-13 01:32:29 <lfm> i spoze they disable hardware debug on consumer chips, then we have to wait for insider leaks
313 2011-01-13 01:32:33 <ArtForzZz> yep, it's broken
314 2011-01-13 01:32:40 <Diablo-D3> lfm: no, theres no insider leak
315 2011-01-13 01:32:51 <Diablo-D3> lfm: once the SPE goes into supervisor mode, it cant be taken out
316 2011-01-13 01:32:55 <lfm> someone has a backup of the key
317 2011-01-13 01:32:57 <ArtForzZz> it might take a few years and a few M$, but it's certainly possible to RE hardware
318 2011-01-13 01:33:03 <Diablo-D3> lfm: its not a key
319 2011-01-13 01:33:10 <Diablo-D3> lfm: it simply ignores hardware debug commands
320 2011-01-13 01:33:23 <Diablo-D3> the hardware is designed this way
321 2011-01-13 01:33:24 <ArtForzZz> unless you find the FF storing the "SPE in supervisor mode" bit, create a contact with a FIB and bond it out
322 2011-01-13 01:33:39 <lfm> diablo how do they make new versions of firmware? someone has a key dont they?
323 2011-01-13 01:33:39 <Diablo-D3> ArtForzZz: well, you'd have to open the chip
324 2011-01-13 01:33:45 <ArtForzZz> err... yes
325 2011-01-13 01:33:47 <Diablo-D3> lfm: depends which part
326 2011-01-13 01:33:49 <ArtForzZz> thats generally how you attack crypto processors
327 2011-01-13 01:34:09 <Diablo-D3> lfm: sony cant update the first part of the bootloader without IBM
328 2011-01-13 01:34:13 <Diablo-D3> lfm: and they havent updated it
329 2011-01-13 01:34:19 <Diablo-D3> and updating it wont fix the problem
330 2011-01-13 01:34:41 <lfm> anything they can update we can use to boot our stuff if it is leaked
331 2011-01-13 01:34:48 <Diablo-D3> yes
332 2011-01-13 01:34:59 <Diablo-D3> the real problem is
333 2011-01-13 01:35:02 <ArtForzZz> ;;bc,mtgox
334 2011-01-13 01:35:02 <gribble> {"ticker":{"high":0.3996,"low":0.3176,"vol":39878,"buy":0.33,"sell":0.3996,"last":0.3996}}
335 2011-01-13 01:35:03 <Diablo-D3> to resecure this
336 2011-01-13 01:35:11 <Diablo-D3> we need to replace every ps3 AND every printed game
337 2011-01-13 01:35:20 <Diablo-D3> because we now know the key for game drm
338 2011-01-13 01:35:37 <Diablo-D3> because we can decrypt the part of the firmware that manages game drm
339 2011-01-13 01:35:41 <ArtForzZz> yep
340 2011-01-13 01:35:52 <Diablo-D3> all because of _one_ bug
341 2011-01-13 01:35:54 <Diablo-D3> no nx.
342 2011-01-13 01:36:02 <ArtForzZz> iirc we can now decrypt the whole chain down to the 0bl
343 2011-01-13 01:36:14 <Diablo-D3> yup
344 2011-01-13 01:36:35 <xelister> what is the bug against which NX is needed here?
345 2011-01-13 01:36:38 <Diablo-D3> you can make sony's first boot loader stage directly load a new OS
346 2011-01-13 01:36:40 <Diablo-D3> and sony cant stop you
347 2011-01-13 01:36:48 <ArtForzZz> yep
348 2011-01-13 01:36:54 <Diablo-D3> it'd have full access to the hardware
349 2011-01-13 01:37:05 <lfm> replace hypervisor then
350 2011-01-13 01:37:10 <brocktice> osnap
351 2011-01-13 01:37:11 <Diablo-D3> lfm: it never loads
352 2011-01-13 01:37:15 <brocktice> almost hit 0.4 on mtgox
353 2011-01-13 01:37:20 <ArtForzZz> this is the stuff that normally LOADS the hypervisor
354 2011-01-13 01:37:22 <Diablo-D3> the security hypervisor is in the second bootloader stage
355 2011-01-13 01:37:28 <lfm> k
356 2011-01-13 01:37:38 <Diablo-D3> the security hypervisor only does important shit like manage game drm
357 2011-01-13 01:37:52 <Diablo-D3> it doesnt do anything as part of the OS security chain
358 2011-01-13 01:37:55 <ArtForzZz> and iirc it even doesnt do that properly
359 2011-01-13 01:37:58 <Diablo-D3> it just needs to not leak shit
360 2011-01-13 01:37:59 <lfm> and it sandboxed the old linux
361 2011-01-13 01:38:09 <ArtForzZz> yep, and failed at it ;)
362 2011-01-13 01:38:18 <Diablo-D3> lfm: yes, OtherOS ran as a hypervisor guest, same way Game OS does
363 2011-01-13 01:38:31 <Diablo-D3> the hardware managed the SPE protection
364 2011-01-13 01:38:34 <Diablo-D3> so there wasnt a real issue
365 2011-01-13 01:39:02 <Diablo-D3> IBM must be pissed though
366 2011-01-13 01:39:14 <ArtForzZz> why?
367 2011-01-13 01:39:15 <Diablo-D3> at no point did they drop the ball
368 2011-01-13 01:39:32 <Diablo-D3> and they're going to have to end up making it harder for solution vendors to fuck shit up
369 2011-01-13 01:39:38 <ArtForzZz> yep
370 2011-01-13 01:39:54 <Diablo-D3> Sony only fucked up because IBM didnt supply the whole fucking thing
371 2011-01-13 01:40:00 <Diablo-D3> which now they'll essentially have to do
372 2011-01-13 01:40:06 <Diablo-D3> because the Sonys of the world cant be trusted to do it right
373 2011-01-13 01:40:23 <ArtForzZz> I still dont get how you can botch ECDSA so badly
374 2011-01-13 01:40:28 <Diablo-D3> ArtForzZz: the random key
375 2011-01-13 01:40:30 <Diablo-D3> er
376 2011-01-13 01:40:33 <Diablo-D3> the random part of the key
377 2011-01-13 01:40:38 <Diablo-D3> k or whatever
378 2011-01-13 01:40:41 <ArtForzZz> yeah, the nonrandom nonce
379 2011-01-13 01:40:43 <Diablo-D3> its supposed to be a nonce
380 2011-01-13 01:40:45 <Diablo-D3> it wasnt.
381 2011-01-13 01:40:57 <ArtForzZz> yep
382 2011-01-13 01:40:57 <Diablo-D3> someone didnt fucking check their shit
383 2011-01-13 01:41:03 <ArtForzZz> it was a nalways
384 2011-01-13 01:41:08 <Diablo-D3> lol
385 2011-01-13 01:41:12 <Diablo-D3> it was the number 4.
386 2011-01-13 01:41:19 <Diablo-D3> chosen fairly by a die.
387 2011-01-13 01:41:33 <ArtForzZz> but serously, why fucking implement ECDSA sig GENERATION yourself?
388 2011-01-13 01:41:45 <Diablo-D3> sometimes I just want to punch someone's blood out
389 2011-01-13 01:41:51 <ArtForzZz> crypto libraries exist for a purpose
390 2011-01-13 01:41:55 <Diablo-D3> ArtForzZz: well
391 2011-01-13 01:41:57 <Diablo-D3> like I said
392 2011-01-13 01:42:05 <ArtForzZz> NIH syndrome I guess
393 2011-01-13 01:42:07 <Diablo-D3> IBM is going to have to handhold on future shit
394 2011-01-13 01:42:25 <Diablo-D3> IBM has a fucking history of doing shit perfectly.
395 2011-01-13 01:42:26 <ArtForzZz> yep
396 2011-01-13 01:42:40 <Diablo-D3> and you know what
397 2011-01-13 01:42:50 <Diablo-D3> I'd even buy into a trusted platform of IBM's if I needed one
398 2011-01-13 01:43:04 <Diablo-D3> there is huge security implications in not having shit that breaks at a drop of a hat
399 2011-01-13 01:43:08 <ArtForzZz> well, they generally do what they're supposed to
400 2011-01-13 01:43:21 <xelister> TRADE| 50 @ $0.395
401 2011-01-13 01:43:22 <xelister> huhuhu
402 2011-01-13 01:43:49 <ArtForzZz> though I had to implement some ugly hacks for enterprise IBM hardware
403 2011-01-13 01:44:19 <lfm> ya IBM was pretty evil when they were on top with 370 etc
404 2011-01-13 01:45:10 <lfm> like ebcidic! really!
405 2011-01-13 01:51:54 redMBA has joined
406 2011-01-13 01:59:50 int0x27h has quit (Changing host)
407 2011-01-13 01:59:50 int0x27h has joined
408 2011-01-13 02:02:56 Sirius has joined
409 2011-01-13 02:04:48 nevezen has quit (Quit: leaving)
410 2011-01-13 02:06:02 nevezen has joined
411 2011-01-13 02:07:33 kermit has joined
412 2011-01-13 02:08:08 darrob has quit (Ping timeout: 240 seconds)
413 2011-01-13 02:13:23 15SABC9TM has joined
414 2011-01-13 02:13:55 15SABC9TM is now known as dwdollar
415 2011-01-13 02:14:25 zylche has quit (Ping timeout: 255 seconds)
416 2011-01-13 02:14:53 zylche has joined
417 2011-01-13 02:14:54 zylche has quit (Excess Flood)
418 2011-01-13 02:15:27 zylche has joined
419 2011-01-13 02:19:59 davex__ has quit (Remote host closed the connection)
420 2011-01-13 02:22:48 davex__ has joined
421 2011-01-13 02:23:45 EvanR has joined
422 2011-01-13 02:23:54 darrob has joined
423 2011-01-13 02:28:00 <luke-jr> is Genjix ever here?
424 2011-01-13 02:29:41 <nanotube> ;;seen genjix
425 2011-01-13 02:29:41 <gribble> genjix was last seen in #bitcoin-dev 3 days, 4 hours, 38 minutes, and 3 seconds ago: <genjix> ic
426 2011-01-13 02:29:47 <nanotube> yes
427 2011-01-13 02:32:03 <EvanR> "if you hire a hooker and pay her in bitcoins, is it technically hooking"
428 2011-01-13 02:33:29 <lfm> if you hire a hooker and pay here in buttons, is it technicly protitution?
429 2011-01-13 02:33:39 <EvanR> yes, the next question
430 2011-01-13 02:34:14 <lfm> same thing
431 2011-01-13 02:34:24 <EvanR> ok now whats the answer ;)
432 2011-01-13 02:34:39 <lfm> yes
433 2011-01-13 02:34:49 <EvanR> what about legally
434 2011-01-13 02:35:22 <EvanR> bitcoin legal advise, 10000BTC per hour
435 2011-01-13 02:35:34 <lfm> prolly depends on local law. they try not to prosecute husbands that put their wife on an allowance
436 2011-01-13 02:35:41 * luke-jr notes the Trade page *does* list a lawyer
437 2011-01-13 02:35:50 <EvanR> lol
438 2011-01-13 02:36:21 devon_hillard has quit (Quit: Leaving)
439 2011-01-13 02:36:36 <luke-jr> lfm: that is clearly not prostitution, since the wife retains the marital debt even if she is given no allowance
440 2011-01-13 02:37:01 <EvanR> marriage is like grand prostitution
441 2011-01-13 02:37:18 <lfm> but if the husband cuts off the allowance if the wife cuts off the nookie, is that prostitution?
442 2011-01-13 02:37:36 <luke-jr> wife can't "cut off the nookie"
443 2011-01-13 02:37:48 <lfm> depends on the jurisdiction
444 2011-01-13 02:37:52 <luke-jr> no
445 2011-01-13 02:38:11 <luke-jr> a wife is bound under natural law to have relations so long as the husband's request is reasonable
446 2011-01-13 02:38:13 <luke-jr> (and vice versa)
447 2011-01-13 02:38:24 <lfm> luke-jr we are talking civl law here, not biblical law
448 2011-01-13 02:38:30 <luke-jr> natural law applies to all
449 2011-01-13 02:38:44 <lucky> huh?
450 2011-01-13 02:38:48 <lucky> says who? that's stupid.
451 2011-01-13 02:38:52 <lfm> you are deluded Luke
452 2011-01-13 02:39:00 <luke-jr> that's part of the nature of marriage
453 2011-01-13 02:39:09 <lfm> in your dreams
454 2011-01-13 02:39:52 <lucky> ... says who?
455 2011-01-13 02:40:12 <luke-jr> lucky: the infallible Catholic Church ofc
456 2011-01-13 02:40:36 <lucky> uh... huh.
457 2011-01-13 02:40:38 <lfm> infallible? dont make us laugh
458 2011-01-13 02:40:53 <luke-jr> lfm: why not? you gonna die if you laugh?
459 2011-01-13 02:41:15 devon_hillard has joined
460 2011-01-13 02:41:19 devon_hillard has quit (Read error: Connection reset by peer)
461 2011-01-13 02:44:04 <lfm> luke-jr you make it sound like you are one of those people who think a husband cannot rape his own wife by definition?
462 2011-01-13 02:45:12 <luke-jr> lfm: that is correct, although it does not excuse battery of course; if a wife refuses to meet her obligations, it doesn't justify the husband being abusive
463 2011-01-13 02:46:13 <luke-jr> 2 wrongs don't make a right
464 2011-01-13 02:46:25 <lfm> well luke-jr I guess you have been asleep for the last 30 years or so cuz that would be you only excuse for thinking that way
465 2011-01-13 02:47:04 <luke-jr> lfm: natural law and morality never changes
466 2011-01-13 02:47:10 <nevezen> bitcoin's dev & technical discussion forum is an interesting read.. now only starting on the second page :)
467 2011-01-13 02:47:31 <lfm> morality changes all the time. you just arnt paying attention
468 2011-01-13 02:48:17 <lfm> as for "natural law" there isnt much natural about marriage and other similar human activities
469 2011-01-13 02:54:16 <luke-jr> lfm: nope
470 2011-01-13 02:54:26 <luke-jr> nothing perfect changes
471 2011-01-13 02:56:41 <lfm> luke, you're a hopeless troll. I appologize to the rest of the channel for getting roped into your stupid debate
472 2011-01-13 03:08:27 OneFixt has quit (Read error: Connection reset by peer)
473 2011-01-13 03:08:41 OneFixt has joined
474 2011-01-13 03:08:43 OneFixt has quit (Changing host)
475 2011-01-13 03:08:43 OneFixt has joined
476 2011-01-13 03:10:17 <kiba> finally
477 2011-01-13 03:10:25 <kiba> now all I need to do is sell my article for bitcoin
478 2011-01-13 03:10:49 nevezen has quit (Quit: leaving)
479 2011-01-13 03:12:20 <EvanR> too many L names flying around
480 2011-01-13 03:15:08 <luke-jr> lfm: just the facts
481 2011-01-13 03:17:47 <luke-jr> anyhow⦠anything wrong with this? Mushkin 2x4GB DDR3 1333 + AMD Phenom II X4 955 + Gigabyte GA-890GPA-UD3H
482 2011-01-13 03:25:28 <Keefe> asking whether any of that is not compatible with each other?
483 2011-01-13 03:26:18 <luke-jr> whether it sounds like a good desktop
484 2011-01-13 03:26:30 <luke-jr> $344
485 2011-01-13 03:26:55 <Keefe> that's subjective
486 2011-01-13 03:27:13 <Keefe> for most people, the very cheapest stuff newegg sells is just fine
487 2011-01-13 03:29:14 <Keefe> i.e. 1gb ddr3, sempron 140, cheap ddr3 board
488 2011-01-13 03:49:46 mtgox has quit (Ping timeout: 276 seconds)
489 2011-01-13 04:02:21 ybit has left ()
490 2011-01-13 04:05:45 dust__ has joined
491 2011-01-13 04:09:57 niekie has joined
492 2011-01-13 04:22:22 <xelister> TRADE| 4.18 @ $0.4
493 2011-01-13 04:22:23 <xelister> !!!
494 2011-01-13 04:23:59 <Cusipzzz> painting the tape obv....less than 2 USD :p
495 2011-01-13 04:28:55 lucky- has joined
496 2011-01-13 04:29:36 lucky has quit (Read error: Connection reset by peer)
497 2011-01-13 04:31:25 <kupo> is it ok to run 0.3.19 ?
498 2011-01-13 04:31:29 <kupo> sorry 0.3.15 ?
499 2011-01-13 04:33:53 <nanotube> kupo: iirc it's fine, no security fixes in .15+ that i remember. but best to run latest if you can.
500 2011-01-13 04:34:22 <kupo> ok ty
501 2011-01-13 04:34:29 <kupo> how do I join these pools btw?
502 2011-01-13 04:34:37 <nanotube> see instructions on the ,,pool page
503 2011-01-13 04:34:37 <gribble> No fancy GPU farm, and don't want to wait for months for a block gen? Join the mining pool! http://mining.bitcoin.cz/
504 2011-01-13 04:35:41 <kupo> ack
505 2011-01-13 04:36:14 <nanotube> have fun. :)
506 2011-01-13 04:36:48 * nanotube pings nathan7 nameless| Xunie for ops. :)
507 2011-01-13 04:39:43 lucky- is now known as lucky
508 2011-01-13 04:39:43 <kiba> so the price of bitcoin reached 0.40 USD
509 2011-01-13 04:39:43 lucky has quit (Changing host)
510 2011-01-13 04:39:43 lucky has joined
511 2011-01-13 04:39:48 <kiba> wee!
512 2011-01-13 04:40:08 <kiba> trying to sell my interview for bitcoin...
513 2011-01-13 04:40:15 <kiba> interview article...
514 2011-01-13 04:41:45 darrob has quit (Ping timeout: 240 seconds)
515 2011-01-13 04:44:10 darrob has joined
516 2011-01-13 04:44:24 * nameless !~root@weowntheinter.net|pings nanotube for money
517 2011-01-13 04:44:43 <kupo> hiya nanotube
518 2011-01-13 04:44:49 * kiba pings bencoder
519 2011-01-13 04:44:56 <Cusipzzz> 5770 for 120$ good deal?
520 2011-01-13 04:45:12 <Cusipzzz> at these rates, even i may have to get into mining
521 2011-01-13 04:46:16 <nanotube> nameless|: haha
522 2011-01-13 04:46:20 <nanotube> kiba: pong
523 2011-01-13 04:48:04 <nanotube> kupo: sup?
524 2011-01-13 04:48:59 <xelister> kiba: interview about what
525 2011-01-13 04:50:14 <kiba> MT's QBitcoin project
526 2011-01-13 04:50:40 <kupo> nanotube: nvm
527 2011-01-13 04:51:14 <nanotube> k
528 2011-01-13 04:54:00 grondilu has joined
529 2011-01-13 04:54:15 <grondilu> nice spike in price on mtgox !
530 2011-01-13 04:54:30 <nanotube> yep
531 2011-01-13 04:54:37 <nanotube> it's been all the rage around here. :)
532 2011-01-13 04:55:16 * Cusipzzz rages
533 2011-01-13 04:55:36 * nanotube rages against fiat currency heh
534 2011-01-13 04:56:38 * Cusipzzz cries at his unexecuted mtgox buy orders :/
535 2011-01-13 04:56:51 <nanotube> heh
536 2011-01-13 04:56:52 nevezen has joined
537 2011-01-13 04:58:54 <xelister> RAGGGEEEON
538 2011-01-13 04:59:49 <xelister> rageon 6999. Now with new, better drivers! Don't worry about lack of multi-GUI OpenCL, any OpenCL or Linux support, why bother when you have cool x64 anti-aliasing, woo
539 2011-01-13 05:00:02 <xelister> *gpu
540 2011-01-13 05:00:29 <kiba> yay!
541 2011-01-13 05:00:37 <kiba> a 27% rise in the price of bitcoin in one day!
542 2011-01-13 05:02:09 <newsham> bubble bobble
543 2011-01-13 05:02:33 <newsham> bid/ask spread is huge
544 2011-01-13 05:02:58 <newsham> 0.34 to 0.40
545 2011-01-13 05:03:40 <newsham> 0.34 to 0.414
546 2011-01-13 05:04:06 <Cusipzzz> when people wake up and seethe offer at > .4, they will move bids up...
547 2011-01-13 05:04:35 <kiba> I predict we end up at 0.50
548 2011-01-13 05:04:36 <kiba> tommorow
549 2011-01-13 05:04:49 <nanotube> kiba: whatcha willing to bet? :)
550 2011-01-13 05:05:02 <Cusipzzz> i'll take some of that as well =)
551 2011-01-13 05:05:02 <kiba> no bet
552 2011-01-13 05:05:03 <hacim> 5870's are about $200-250, i bet they are pretty decent still compared to a 5970
553 2011-01-13 05:05:16 <nanotube> kiba: hehe
554 2011-01-13 05:05:36 <newsham> there's a lot of shares up between 0.414 and 0.5
555 2011-01-13 05:05:40 <nanotube> Cusipzzz: that said, there are only 2000+ worth of asks between here and .5
556 2011-01-13 05:05:58 <newsham> at least 2kbtc
557 2011-01-13 05:06:01 <Cusipzzz> 1000 USD...then a brick wall at .51
558 2011-01-13 05:06:30 <nanotube> mm, well, 'brick wall' in the amount of 10kusd. considering recent volumes, not so much.
559 2011-01-13 05:06:37 <Cusipzzz> ya maybe
560 2011-01-13 05:06:46 <nanotube> the main question is... will the miners wake up and be like zomg, maybe now is a good time to sell before the bubble bursts. :)
561 2011-01-13 05:07:13 <Cusipzzz> hope so...voltatility is a good thing...and not straight up :p
562 2011-01-13 05:07:21 <Keefe> if they put in asks around 0.4, it won't affect the market much
563 2011-01-13 05:07:57 <newsham> i dont think 0.4 is the peak of hte bubble
564 2011-01-13 05:08:01 <nanotube> well, depends on if the buying pressure keeps up.
565 2011-01-13 05:08:30 * hacim wonders where the buying pressure comes from
566 2011-01-13 05:08:41 * Cusipzzz eyes EvanR suspiciously
567 2011-01-13 05:08:49 <hacim> he's offline!
568 2011-01-13 05:08:54 <kiba> w00t!
569 2011-01-13 05:08:58 <kiba> we reach 2 million dollars
570 2011-01-13 05:09:09 <kiba> now, we're a 2 million dollars economy!
571 2011-01-13 05:09:14 <nevezen> is there a maximum amount of bitcoins that can be generated at the moment?
572 2011-01-13 05:09:28 <nanotube> nevezen: can you clarify your question?
573 2011-01-13 05:09:29 <kiba> 21 million
574 2011-01-13 05:09:37 <nanotube> nevezen: what do you mean by 'at the moment' ?
575 2011-01-13 05:09:47 <newsham> nev: 300 per hr?
576 2011-01-13 05:09:53 <nevezen> before generating becomes irrelevant, if ever..
577 2011-01-13 05:10:00 <kiba> 6 blocks per hours, translate into 300
578 2011-01-13 05:10:04 <kiba> 300 * 24
579 2011-01-13 05:10:13 <newsham> not quite a maximum (could be a lot more, if a bunch of miners came online quickly before difficulty got adjusted again) but roughly
580 2011-01-13 05:10:15 <Cusipzzz> some blocks were really long today, like 45 mins+
581 2011-01-13 05:10:31 <kiba> around 7200 bitcoins
582 2011-01-13 05:10:49 <nevezen> so what's the current difficulty at the moment? is there a scale?
583 2011-01-13 05:10:56 <nanotube> ;;bc,stats
584 2011-01-13 05:10:58 <gribble> Current Blocks: 102392 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 423 blocks | Next Difficulty In About: 2 days, 12 hours, 51 minutes, and 54 seconds | Next Difficulty Estimate: 18046.79146339
585 2011-01-13 05:11:02 <nanotube> nevezen: --^
586 2011-01-13 05:11:04 <nevezen> ahh
587 2011-01-13 05:11:48 <nevezen> will the difficulty just continue increasing to infinity?
588 2011-01-13 05:12:13 <nanotube> it will continue adjusting to keep block gen to about 1 per 10 min.
589 2011-01-13 05:12:13 <newsham> difficulty is adjusted to target a specific block rate.
590 2011-01-13 05:12:30 <newsham> so if a bunch of miners go offline, it could come down
591 2011-01-13 05:12:39 <newsham> but more likely it will grow as bitcoin popularity grows
592 2011-01-13 05:12:51 <nevezen> oh, so it's not based on the number of current existing blocks
593 2011-01-13 05:13:17 <nanotube> no
594 2011-01-13 05:13:18 <tcatm> http://tcatm.github.com/bitcoin-js-remote/ updated js-remote with QR and bitcoin: URI support
595 2011-01-13 05:13:19 <newsham> its based on how fast people have been mining recently
596 2011-01-13 05:13:23 <nanotube> ;;bc,wiki difficulty
597 2011-01-13 05:13:24 <gribble> https://en.bitcoin.it/wiki/Difficulty | Jan 3, 2011 ... Difficulty is a measure of how difficult it is to find a new block compared to ... Current difficulty, as output by BitCoin's getDifficulty. ...
598 2011-01-13 05:13:48 <nevezen> more time in the wiki then for me :)
599 2011-01-13 05:17:13 <nanotube> tcatm: ooh nice. :)
600 2011-01-13 05:19:02 <hacim> i'm guessing most miners are using 5970's but since everyone who had 5870's switched to 5970's when they came out, those are much cheaper, if half the h/sec
601 2011-01-13 05:19:34 <nanotube> well, you also have to consider hpW
602 2011-01-13 05:19:39 <nevezen> is there a maximum amount of USD bitcoin can hold?
603 2011-01-13 05:19:50 <nanotube> nevezen: bitcoin doesn't hold any usd
604 2011-01-13 05:19:57 <hacim> nanotube: hpW is how much juice its going to cost?
605 2011-01-13 05:19:57 <nanotube> any more than an oz of gold holds usd
606 2011-01-13 05:20:07 <nanotube> hacim: yea, hash per watt
607 2011-01-13 05:20:07 <hacim> tcatm: thats pretty sweet!
608 2011-01-13 05:20:11 <nevezen> I mean if you wanted to spend 10 million dollars, converted to btc
609 2011-01-13 05:20:41 <nanotube> ah, well you can look at current market price to get an idea ,,bc,mtgox
610 2011-01-13 05:20:42 <gribble> {"ticker":{"high":0.405,"low":0.3176,"vol":40439,"buy":0.34,"sell":0.414,"last":0.405}}
611 2011-01-13 05:20:43 <Cusipzzz> it would be difficult to buy that many
612 2011-01-13 05:20:57 <nanotube> nevezen: but you couldn't really buy many at that price before driving up price
613 2011-01-13 05:20:59 <hacim> hum, I'm surprised the hpW aren't on http://174.143.149.98/wiki/doku.php?id=bitcoin_miners
614 2011-01-13 05:21:06 <nevezen> it was just an example
615 2011-01-13 05:21:09 <nanotube> ;;bc,wiki mining hardware comparison
616 2011-01-13 05:21:09 <gribble> https://en.bitcoin.it/wiki/Mining_Hardware_Comparison | Jan 3, 2011 ... Mining Hardware Comparison. From Bitcoin. Jump to: navigation, search. This page is a stub, you can help by expanding it. ...
617 2011-01-13 05:21:14 <nanotube> hacim: --^ they are here
618 2011-01-13 05:21:24 <hacim> oh thats a better page
619 2011-01-13 05:21:31 <x6763> nevezen: http://www.bitcoin.org/sites/default/files/bitcoin.pdf - Satoshi's (bitcoin creator) whitepaper
620 2011-01-13 05:21:59 <hacim> slightly more Mhash/W
621 2011-01-13 05:22:22 <hacim> err, slightly less
622 2011-01-13 05:23:47 * hacim tries to find cost of electricity locally
623 2011-01-13 05:24:17 <nanotube> look at your electric bill. ;)
624 2011-01-13 05:24:40 <hacim> looks like 19.98cents/kWH
625 2011-01-13 05:24:50 <newsham> see also http://aws.amazon.com/free/
626 2011-01-13 05:25:27 <hacim> so if a 5870 can do 1.668Mhash/Watt and that is my cost of electricity
627 2011-01-13 05:25:36 <hacim> hmm someone must have made a calculator for this already
628 2011-01-13 05:25:48 Cusipzzz has quit ()
629 2011-01-13 05:25:50 <hacim> newsham: are people really using amazon for mining?
630 2011-01-13 05:25:59 grubles has joined
631 2011-01-13 05:26:44 <newsham> dunno
632 2011-01-13 05:28:43 gavinandresen has quit (Quit: gavinandresen)
633 2011-01-13 05:29:07 <nanotube> hacim: not that i know of... but you can calculate for yourself. running it for 24 hours will be ,,(math calc 24*188/1000) kw which will cost you ,,(math calc 24*188/10000*.1998) usd
634 2011-01-13 05:29:07 <gribble> 4.512
635 2011-01-13 05:29:08 <gribble> 0.09014976
636 2011-01-13 05:29:57 <nanotube> er whoops, extra zero in second one. .9014 it will cost per day
637 2011-01-13 05:29:59 <hacim> what is the 188?
638 2011-01-13 05:30:07 <nanotube> 188 watts that the card uses
639 2011-01-13 05:30:13 <nanotube> (see table)
640 2011-01-13 05:30:29 <hacim> oh right, duh
641 2011-01-13 05:31:03 <hacim> ,,(math calc 24*294/10000*.01998)
642 2011-01-13 05:31:04 <gribble> 0.014097888
643 2011-01-13 05:31:19 <grubles> which card are we talking about?
644 2011-01-13 05:31:28 <nanotube> hacim: /1000 not /10000
645 2011-01-13 05:31:40 <hacim> ,,(math calc 24*294/1000*.1998)
646 2011-01-13 05:31:40 <gribble> 1.4097888
647 2011-01-13 05:31:42 <nanotube> grubles: ati 5870
648 2011-01-13 05:32:05 <hacim> err no, thats off again, that should be 14cents/day
649 2011-01-13 05:32:12 <grubles> ah, i have a 5970
650 2011-01-13 05:32:40 dust__ has quit (Ping timeout: 265 seconds)
651 2011-01-13 05:33:42 <nanotube> grubles: 5970 is 294W stock. so make your calculations accordingly. :)
652 2011-01-13 05:34:11 <nanotube> hacim: no, it's 1.4 usd per day.
653 2011-01-13 05:34:23 <nanotube> to run the 5970 24/7 at your cost of power.
654 2011-01-13 05:34:39 <nanotube> ;;bc,calc 600000
655 2011-01-13 05:34:40 <gribble> The average time to generate a block at 600000 Khps, given current difficulty of 16307.48285682 , is 1 day, 8 hours, 25 minutes, and 33 seconds
656 2011-01-13 05:34:52 <grubles> mines oc'd to 885mhz core with 1.162v so it has to be more than that
657 2011-01-13 05:34:53 newsham has quit (Remote host closed the connection)
658 2011-01-13 05:35:02 <nanotube> hacim: and you'll gen a block about once every 1.5 days, which is worth about 20 usd. so heftily profitable.
659 2011-01-13 05:35:10 <nanotube> grubles: yea probably.
660 2011-01-13 05:35:16 <hacim> actually, i was wrong about my cost of power, its ,,(math calc 24*188/1000*.109362)
661 2011-01-13 05:35:17 <gribble> 0.493441344
662 2011-01-13 05:35:26 <nanotube> nice
663 2011-01-13 05:35:29 <hacim> perday for the 5870
664 2011-01-13 05:35:55 <hacim> and ,,(math calc 24*294/1000*.109362) for the 5970
665 2011-01-13 05:35:56 <gribble> 0.771658272
666 2011-01-13 05:36:04 <grubles> +60000 k/hash/s with the overclock
667 2011-01-13 05:36:57 <hacim> nanotube: a block about once every 1.5 days for the 5870 or the 5970?
668 2011-01-13 05:36:59 <grubles> totaling ~660 m/hash/s
669 2011-01-13 05:37:46 * hacim tries to figure out where you got 600000 Khps from
670 2011-01-13 05:38:48 <grubles> i think you misread, its 60k
671 2011-01-13 05:39:20 <nanotube> hacim: it's 600mhps for the 5970, about half that for the 5870
672 2011-01-13 05:39:25 <nanotube> (assuming a bit of overclocking)
673 2011-01-13 05:39:26 * hacim is confused by the grubles and gribles :)
674 2011-01-13 05:39:33 <nanotube> haha yea
675 2011-01-13 05:39:46 <grubles> a stock 5970 i think is 500 m/hash/s
676 2011-01-13 05:39:58 <hacim> table says 535.06
677 2011-01-13 05:40:09 <grubles> yeah sounds right
678 2011-01-13 05:40:51 <nanotube> yep
679 2011-01-13 05:41:02 <nanotube> that's stock. people routinely oc it to around 600 though
680 2011-01-13 05:41:17 <nanotube> but ... ,,(bc,calc 535000)
681 2011-01-13 05:41:17 <gribble> The average time to generate a block at 535000 Khps, given current difficulty of 16307.48285682 , is 1 day, 12 hours, 21 minutes, and 56 seconds
682 2011-01-13 05:41:23 <nanotube> still about 1.5 days. :)
683 2011-01-13 05:41:37 <hacim> ,,(bc,calc 313.65)
684 2011-01-13 05:41:38 <gribble> The average time to generate a block at 313.65 Khps, given current difficulty of 16307.48285682 , is 7 years, 4 weeks, 1 day, 13 hours, 36 minutes, and 9 seconds
685 2011-01-13 05:41:44 <hacim> oops
686 2011-01-13 05:41:48 <grubles> lol
687 2011-01-13 05:41:55 <hacim> ,,(bc,calc 313650)
688 2011-01-13 05:41:56 <gribble> The average time to generate a block at 313650 Khps, given current difficulty of 16307.48285682 , is 2 days, 14 hours, 1 minute, and 46 seconds
689 2011-01-13 05:42:07 <hacim> wow, basically 2x
690 2011-01-13 05:42:27 <grubles> with my crossfire 5770s thats how long it would take
691 2011-01-13 05:42:28 <hacim> 2x the time, 2x the Mhash/sec, 2x the money
692 2011-01-13 05:42:35 LobsterMan has quit (Remote host closed the connection)
693 2011-01-13 05:43:01 <hacim> grubles: isn't the 5770's ,,(bc,calc 156830) ?
694 2011-01-13 05:43:02 <gribble> The average time to generate a block at 156830 Khps, given current difficulty of 16307.48285682 , is 5 days, 4 hours, 3 minutes, and 18 seconds
695 2011-01-13 05:43:37 <grubles> basically know that a 5970 has 3200 stream processors, half that for the 5870, half again for the 5770, so 1600 for the 5870 and 800 for the 5770.
696 2011-01-13 05:44:01 <grubles> i had 2 in crossfire :p
697 2011-01-13 05:44:10 <hacim> ah
698 2011-01-13 05:44:18 <hacim> i wonder what people are able to overclock their 5770's at
699 2011-01-13 05:44:26 <hacim> err, 5870s
700 2011-01-13 05:44:52 <grubles> not sure, i've heard 2 5870s in crossfire overclock and perform better than the 5970 though
701 2011-01-13 05:45:07 <nanotube> but also eat more power
702 2011-01-13 05:45:27 <grubles> but i got my 5970 in order to save room for another one to have quadfire
703 2011-01-13 05:45:30 <grubles> indeed
704 2011-01-13 05:46:05 LobsterMan has joined
705 2011-01-13 05:46:05 LobsterMan has quit (Changing host)
706 2011-01-13 05:46:05 LobsterMan has joined
707 2011-01-13 05:46:19 <hacim> what is crossfire/quadfire?
708 2011-01-13 05:46:52 <grubles> having more than one graphics card run in parallel
709 2011-01-13 05:46:58 <lfm> hacim ati multi-gpu system
710 2011-01-13 05:47:04 <hacim> ah
711 2011-01-13 05:47:12 <grubles> you can have 2,3, or 4 cards
712 2011-01-13 05:47:13 redMBA has quit (Ping timeout: 260 seconds)
713 2011-01-13 05:47:22 <hacim> so i'm guessing everyone is waiting for the next gen card to come out
714 2011-01-13 05:47:32 <LobsterMan> does anyone have any experience overclocking nvidia cards?
715 2011-01-13 05:47:55 <grubles> i think even the 6970 wont be as good as the 5970
716 2011-01-13 05:48:08 <hacim> what about the 6990 amd?
717 2011-01-13 05:48:09 <nanotube> LobsterMan: probably same general rules apply: be careful, monitor temps, etc.
718 2011-01-13 05:48:17 <grubles> in terms of bitcoin mining
719 2011-01-13 05:48:25 <LobsterMan> my temps are fine, but my driver crashed if i took my clock too high
720 2011-01-13 05:48:25 <hacim> http://www.fudzilla.com/graphics/item/21487-amd-to-have-hd-6990-dual-cards-in-february?utm_source=twitterfeed&utm_medium=twitter
721 2011-01-13 05:48:25 <LobsterMan> lol
722 2011-01-13 05:48:40 <grubles> samething happens with ati cards
723 2011-01-13 05:48:42 <nevezen> I use rivatuner..
724 2011-01-13 05:48:45 <nanotube> hacim: i think none of the new cards will be better mining wise... that's what i've been hearing
725 2011-01-13 05:48:51 <LobsterMan> i'm using rivatuner too
726 2011-01-13 05:48:51 <grubles> probably means your voltage is too low
727 2011-01-13 05:48:53 <hacim> interesting
728 2011-01-13 05:49:02 <hacim> nanotube: what are you running?
729 2011-01-13 05:49:06 <nevezen> but I max out my fanspeed..
730 2011-01-13 05:49:09 <LobsterMan> i've got 2 gtx275's running stock @ 684mhz, i took them to 760mhz and my driver crashed
731 2011-01-13 05:49:11 annodomini has joined
732 2011-01-13 05:49:23 <nanotube> hacim: mining wise - nothing. my comp is a laptop.
733 2011-01-13 05:49:40 <nanotube> LobsterMan: so... up the voltage if you can... or just don't take them up so high. :)
734 2011-01-13 05:49:44 <grubles> lobsterman, have you increased the voltage for your cards
735 2011-01-13 05:50:03 <LobsterMan> can i do that in rivatuner? or would i need to do that in the bios?
736 2011-01-13 05:50:55 <nevezen> I think rivatuner can do it..
737 2011-01-13 05:51:20 <LobsterMan> my temps are fine, even at 760MHz (around 82), but i guess something broke because my display went black for a few secs and then i got a warning that the driver crashed but it had recovered, then my cards went into some low power mode until i rebooted...
738 2011-01-13 05:51:35 <grubles> lobsterman, did you just set it 760 or gradually get to that speed
739 2011-01-13 05:52:21 <nevezen> is 100'C too high for an 8800GT?
740 2011-01-13 05:52:23 <LobsterMan> i went from 684 to like 715, then 730, then 760
741 2011-01-13 05:52:42 <LobsterMan> "Display driver nvlddmkm stopped responding and has successfully recovered."
742 2011-01-13 05:53:05 <grubles> yeah thats what happens usually when your overclock is unsuccessful
743 2011-01-13 05:53:09 <lfm> i think 100 c is like the limit for most any chip
744 2011-01-13 05:53:14 <grubles> or you get the blue screen
745 2011-01-13 05:53:26 <LobsterMan> i'm running at 733 now and all seems ok
746 2011-01-13 05:53:34 <LobsterMan> like i said my temps were fine at 760 so i thought it was ok
747 2011-01-13 05:53:36 <grubles> yeah 760 might just be too high
748 2011-01-13 05:53:51 <grubles> if you havent increased the volts
749 2011-01-13 05:54:13 <hacim> nanotube: was wondering because I'm only running linux, and not sure if I can overclock without windows
750 2011-01-13 05:54:25 <hacim> i'm only on a laptop myself, but I have a old shuttle pc sitting here off...
751 2011-01-13 05:54:33 <nanotube> hacim: ah well, i'm only running linux too. but i don't really have any tips for you overclocking-wise. :)
752 2011-01-13 05:54:43 <lfm> hacim how old?
753 2011-01-13 05:54:48 <grubles> hacim, what graphics card do you have
754 2011-01-13 05:54:53 <nanotube> nevezen: from what i hear... yes 100c is too high. you should aim for the 70s
755 2011-01-13 05:55:19 <LobsterMan> nvidia cards run kind of hot
756 2011-01-13 05:55:26 <hacim> grubles: i have an ancient card (geforce fx5700), wont even work for a gpu miner... but I am looking at getting a 5870 for about $150
757 2011-01-13 05:55:32 <nevezen> mine is currently at 92, stock speed but the fan is on 100%
758 2011-01-13 05:55:44 <hacim> lfm: the shuttle is a bit old, i'm not sure if it would even take the card
759 2011-01-13 05:55:54 <nanotube> hacim: ooh, 150, that's way less than half the price. may make sense to get two of those rather than one 5970
760 2011-01-13 05:56:01 <nanotube> (or even one of those)
761 2011-01-13 05:56:02 <nevezen> I think the minimum requirement for CUDA is an 8800GT
762 2011-01-13 05:56:08 <nevezen> or openCL
763 2011-01-13 05:56:27 <lfm> hacim ya if its one of those real small shuttle boxes it might be too small for many cards
764 2011-01-13 05:56:35 <hacim> thats what i am worried about
765 2011-01-13 05:56:42 <hacim> nanotube: yeah, not a bad price... but only one avail
766 2011-01-13 05:56:53 <hacim> lfm: it is one of the small ones unfortunately
767 2011-01-13 05:56:56 <LobsterMan> my card is one of those "oc" cards that came stock with a slightly higher clock speed, i guess that should be taken into account too
768 2011-01-13 05:57:11 <grubles> a 5870 might be too big for a shuttle case
769 2011-01-13 05:57:27 <nanotube> unless you run open-case. :)
770 2011-01-13 05:57:57 <grubles> you should hear how loud my 5970 is with an open air case
771 2011-01-13 05:58:15 <nanotube> noise smoise. :)
772 2011-01-13 05:58:24 <nanotube> stick it in the basement where it's nice and cool, and forget about it.
773 2011-01-13 05:58:33 <hacim> if you have a basement :)
774 2011-01-13 05:58:44 <grubles> lol thats where it is right now
775 2011-01-13 05:58:55 <grubles> its like a little blow dryer
776 2011-01-13 05:59:05 <LobsterMan> will poclbm make use of the shader clock as well as the core clock?
777 2011-01-13 05:59:19 <nevezen> heh, mine is kinda' loud too
778 2011-01-13 05:59:52 <grubles> lobsterman, i believe the memory clock doesnt affect the hashes per second
779 2011-01-13 06:00:00 <LobsterMan> yeah that it doesn't seem to
780 2011-01-13 06:00:09 <LobsterMan> but i can adjust both the core clock and shader clock with rivatuner
781 2011-01-13 06:00:24 <grubles> right
782 2011-01-13 06:00:29 <hacim> hm noise could be a problem
783 2011-01-13 06:00:31 <grubles> oh i see
784 2011-01-13 06:00:48 <grubles> is there a way to link the clocks to be the same
785 2011-01-13 06:00:53 <lfm> tuning for bitcoins may make it slopw for graphics
786 2011-01-13 06:01:00 <LobsterMan> yeah a little checkbox for "link clocks"
787 2011-01-13 06:01:10 <LobsterMan> only the core clock seems to affect hashrate as best i can see...
788 2011-01-13 06:01:21 <grubles> hmm ok
789 2011-01-13 06:01:32 <grubles> i cant even set the shader clock with my card so i didnt know
790 2011-01-13 06:01:41 <grubles> it must be linked i guess
791 2011-01-13 06:02:36 soultcer has quit (Remote host closed the connection)
792 2011-01-13 06:04:23 <grubles> lobsterman, do you know about -v and -w flags for the py miner
793 2011-01-13 06:04:29 <LobsterMan> yeah
794 2011-01-13 06:04:41 <grubles> did they help at all
795 2011-01-13 06:04:50 <LobsterMan> -v seems to make the miner crash for me, but i set -w 256 and i get a slight increase in hashrate
796 2011-01-13 06:05:09 <hacim> $349 is not too bad for a 5970
797 2011-01-13 06:05:32 <grubles> where is that happening
798 2011-01-13 06:05:36 <nanotube> hacim: where are you getting that? last i saw 5970s were going for 500+
799 2011-01-13 06:05:46 acous has quit (Ping timeout: 260 seconds)
800 2011-01-13 06:05:50 <hacim> i found one on a local craigslist
801 2011-01-13 06:06:02 <LobsterMan> grubles the miner crashes after like 10s if i try to use -v at all
802 2011-01-13 06:06:07 <grubles> i would definitely take that offer up if i had the cash
803 2011-01-13 06:06:22 <nanotube> hacim: sounds like a great offer - assuming card is in working order. :)
804 2011-01-13 06:06:41 <nanotube> i'd get it while you have the chance.
805 2011-01-13 06:07:05 AAA_awright has quit (Read error: Connection reset by peer)
806 2011-01-13 06:07:21 <grubles> lobsterman, interesting, i had a gtx 260 and im pretty sure -v worked on it
807 2011-01-13 06:07:31 AAA_awright has joined
808 2011-01-13 06:07:35 <grubles> lobsterman, are your drivers updated
809 2011-01-13 06:07:41 <hacim> nanotube: heh, i dont even know if I have the hardware to run it, much less a place to put it where it wont be annoyingly loud
810 2011-01-13 06:07:46 <nanotube> hehe
811 2011-01-13 06:07:49 <LobsterMan> i have the most recent official, non beta nvidia drivers
812 2011-01-13 06:07:57 <nanotube> hmm, also there's one on ebay too: http://cgi.ebay.com/Sapphire-Radeon-HD-5970-ATI-2GB-GDDR5-Video-Card-/170590174729?pt=PCC_Video_TV_Cards&hash=item27b7f77e09
813 2011-01-13 06:08:28 <grubles> not bad at all
814 2011-01-13 06:08:43 <hacim> yeah I was just looking at that one too
815 2011-01-13 06:08:53 <hacim> its a little cheaper than the one I was talking about
816 2011-01-13 06:09:07 <hacim> now all you are going to jump on it :P
817 2011-01-13 06:09:22 <grubles> lol
818 2011-01-13 06:09:59 Diablo-D3 has quit (Ping timeout: 240 seconds)
819 2011-01-13 06:09:59 <grubles> not i
820 2011-01-13 06:10:15 <nanotube> well not me - i have an old com in a good case, but it'd need a new mobo+cpu+ram+psu combo to take the 5970. which would have me out at least another 300
821 2011-01-13 06:10:57 <hacim> yeah, i'm guessing that might be my situation too
822 2011-01-13 06:11:12 <grubles> yeah im pretty sure id need a new psu for another
823 2011-01-13 06:11:15 <hacim> probably those don't go in agp slots
824 2011-01-13 06:11:37 <nanotube> yea these are pci-e
825 2011-01-13 06:12:03 <nanotube> hehe, that card's got more ram than my main comp...
826 2011-01-13 06:12:13 <hacim> yeah me too :P
827 2011-01-13 06:12:31 <nevezen> man, I thought 512MB video ram was a monster
828 2011-01-13 06:12:37 <nevezen> but now we're up to 2GB?
829 2011-01-13 06:12:56 <nanotube> ya... crazy times
830 2011-01-13 06:13:06 <grubles> and terabytes of storage
831 2011-01-13 06:13:08 <nanotube> bitcoins are up to .4 usd... video ram is up to 2gb...
832 2011-01-13 06:13:48 <nevezen> I suppose 2GB isn't the max we've seen..
833 2011-01-13 06:14:11 <hacim> yeah, i dont have pci-e
834 2011-01-13 06:14:16 <grubles> there is a 4gb 5970 its $1000 though
835 2011-01-13 06:14:20 <nevezen> ouch
836 2011-01-13 06:14:22 <grubles> the asus one
837 2011-01-13 06:14:31 <grubles> yeah
838 2011-01-13 06:14:35 <hacim> thats pricey
839 2011-01-13 06:15:04 <nanotube> and the extra ram is useless for mining...
840 2011-01-13 06:15:13 <grubles> yeah
841 2011-01-13 06:15:18 <nevezen> I'm curious.. I don't suppose people are running bitcoin clients on tesla clusters, are they?
842 2011-01-13 06:15:33 <grubles> possibly
843 2011-01-13 06:15:37 <nanotube> they could... but teslas aren't as good as atis
844 2011-01-13 06:16:10 <nevezen> I'm behind like about 2 years on components.. :(
845 2011-01-13 06:16:15 <grubles> how would someone cluster say 4 computers each with 4 5970s?
846 2011-01-13 06:16:19 <nevezen> I know nothing about ATI cards
847 2011-01-13 06:16:29 <grubles> with a pool?
848 2011-01-13 06:16:49 <nanotube> grubles: same as you'd cluster 2 computers with 1 5970 each.
849 2011-01-13 06:17:06 <nanotube> just run miner processes on each, and gave it fetch work from the pool
850 2011-01-13 06:17:15 <grubles> in my experience the pooled miners dont have as high of hashes/s than the py miner for instance
851 2011-01-13 06:17:19 <lfm> grubles miners let you specify the ip address for the server
852 2011-01-13 06:17:38 <nanotube> you can run the py miner and just give it the pool address. the miner is exactly the same.
853 2011-01-13 06:17:44 <nanotube> see the ,,pool page for details.
854 2011-01-13 06:17:44 <gribble> No fancy GPU farm, and don't want to wait for months for a block gen? Join the mining pool! http://mining.bitcoin.cz/
855 2011-01-13 06:18:03 <nanotube> at least, that pool. puddinpop's pool does use separate software
856 2011-01-13 06:18:17 <grubles> im talking about the client software
857 2011-01-13 06:18:40 <nanotube> yes, same miner client.
858 2011-01-13 06:18:46 <lfm> grubles yes the bitcoin client is a server to miners
859 2011-01-13 06:18:57 <grubles> with the rpc miner i had less hashes/s than the py miner
860 2011-01-13 06:19:22 <nanotube> what's the 'rpc miner' ?
861 2011-01-13 06:19:46 <nevezen> from the point of joining that pool, when did you start generating coins?
862 2011-01-13 06:19:49 <grubles> http://www.bitcoin.org/smf/index.php?topic=2444.0 here
863 2011-01-13 06:20:03 <lfm> they're all rpc miners but I think hhe means pudnpop's miners
864 2011-01-13 06:20:18 <grubles> yes his
865 2011-01-13 06:20:20 <nanotube> ah those are puddinpop's miners
866 2011-01-13 06:20:28 <nanotube> but you can use the py miner with the slush pool
867 2011-01-13 06:20:38 <nanotube> see that pool page
868 2011-01-13 06:20:55 <grubles> yep
869 2011-01-13 06:21:11 <hacim> i wonder what mhash/sec could be achieved with amazon
870 2011-01-13 06:21:11 <lfm> ifaik all the miners use the same protocol and can be used with any of the sevrers, bitcind or a pool
871 2011-01-13 06:21:14 <nanotube> just saying... if you have a 5970 you don't really need to use the pool. :)
872 2011-01-13 06:21:30 <grubles> lol indeed
873 2011-01-13 06:21:43 <nanotube> lfm: unless something changed in puddinpop's miner... i think his pool server does something special... ?
874 2011-01-13 06:21:58 <nanotube> hacim: probably not enough to pay for itself.
875 2011-01-13 06:22:04 <lfm> nanotube its still compatible afaik
876 2011-01-13 06:22:15 <grubles> so with two seperate computers, you would just set bitcoin.exe -server on one and connect all the py miners to it?
877 2011-01-13 06:22:23 <nanotube> grubles: yes
878 2011-01-13 06:22:34 <lfm> grubles exactly
879 2011-01-13 06:22:43 <grubles> i see
880 2011-01-13 06:23:08 <nanotube> lfm: mm i guess something did change, then. heh
881 2011-01-13 06:23:15 <hacim> nanotube: yeah, i was just wondering if anyone has tried it and what hash/sec they get. curious about what it ends up costing
882 2011-01-13 06:23:44 <nanotube> hacim: someone did... but not at these prices. :)
883 2011-01-13 06:24:11 * hacim searched the forum for the word 'amazon' and surprisingly found nothing
884 2011-01-13 06:24:24 <lfm> nanotube oh ya, pudinpop has a whole new set of miners from his original one, the new ones work with stadard getwork calls
885 2011-01-13 06:24:58 <grubles> hacim, you could have 2 top of the line 1000 dollar cpus and they wouldnt get half the hashes/s as a 100 dollar gpu
886 2011-01-13 06:24:59 <hacim> oh, search doesn't work unless you are logged in
887 2011-01-13 06:25:05 <nanotube> lfm: ah ic
888 2011-01-13 06:25:29 grondilu has quit (Quit: leaving)
889 2011-01-13 06:25:34 <hacim> grubles: amazon has gpu clusters
890 2011-01-13 06:25:52 <nevezen> what does google apps have? :)
891 2011-01-13 06:26:01 <grubles> oh, i didnt know that :D
892 2011-01-13 06:26:05 <hacim> http://aws.amazon.com/ec2/hpc-applications/
893 2011-01-13 06:26:05 <xelister> hacim: amazon builds clusters on shitty nvidias
894 2011-01-13 06:26:28 <hacim> 2 x NVIDIA Tesla âFermiâ M2050 GPUs
895 2011-01-13 06:27:06 <hacim> not even listed on the chart
896 2011-01-13 06:30:21 <nanotube> hacim: yea ask ArtForzZz what those can do, hps wize
897 2011-01-13 06:30:23 <nanotube> wise
898 2011-01-13 06:30:35 <grubles> seems like those have only 448 stream processors
899 2011-01-13 06:30:56 <lfm> ecc memory
900 2011-01-13 06:31:40 soultcer has joined
901 2011-01-13 06:31:48 <grubles> gtx 275 have 240
902 2011-01-13 06:31:49 <nanotube> so... maybe like 90mhps?
903 2011-01-13 06:32:24 LobsterMan has quit (Remote host closed the connection)
904 2011-01-13 06:32:35 <grubles> nvidia are better at floating point operations which bitcoin does none of i believe
905 2011-01-13 06:32:41 <nanotube> right
906 2011-01-13 06:32:52 <hacim> i'm trying to figure out how much it would cost to run a gpu miner on an amazon instance
907 2011-01-13 06:33:05 <xelister> hacim: it is not proffitable I can bet
908 2011-01-13 06:33:08 <nanotube> hacim: they don't seem to have the prices on that right up front anywhere.
909 2011-01-13 06:33:13 <nanotube> xelister: even at .4/btc ?
910 2011-01-13 06:33:20 <grubles> i think i saw it was ~$.04 an hour
911 2011-01-13 06:33:32 <xelister> hm
912 2011-01-13 06:33:34 <nanotube> mmm ok
913 2011-01-13 06:33:50 <xelister> ok but how much dicking around with that?
914 2011-01-13 06:33:53 <grubles> or a couple grand a year
915 2011-01-13 06:34:06 <nanotube> so .04 an hour, you get two teslas at 90mhps each, you also get some number of 'cpu compute units' whatever those are, say another 10 mhps
916 2011-01-13 06:34:15 <nanotube> so 190 mhps total.
917 2011-01-13 06:34:16 <xelister> you just get ssh login with X to account on linux where you can use OpenCL and all is installed?
918 2011-01-13 06:34:18 <nanotube> ;;bc,calc 190000
919 2011-01-13 06:34:19 <gribble> The average time to generate a block at 190000 Khps, given current difficulty of 16307.48285682 , is 4 days, 6 hours, 23 minutes, and 52 seconds
920 2011-01-13 06:34:25 <lfm> ;;bc,calc 90000
921 2011-01-13 06:34:26 <gribble> The average time to generate a block at 90000 Khps, given current difficulty of 16307.48285682 , is 1 week, 2 days, 0 hours, 10 minutes, and 23 seconds
922 2011-01-13 06:34:52 <xelister> 150 coins... 45 usd
923 2011-01-13 06:34:55 <nanotube> so in about 4 days, you gen a block on average, which gives you 20 usd at current prices
924 2011-01-13 06:35:03 <nanotube> (to be generous)
925 2011-01-13 06:35:12 <nanotube> ;;math calc 4*24*.04
926 2011-01-13 06:35:12 <gribble> 3.84
927 2011-01-13 06:35:21 <xelister> so 90M or 190M ?
928 2011-01-13 06:35:21 <lfm> 90 mhash/s or 190 mhash/s??
929 2011-01-13 06:35:23 <nanotube> and cost to run on amazon is 3.84 usd for the 4 days
930 2011-01-13 06:35:32 <nanotube> lfm: 90 per one tesla. there are two teslas in one unit
931 2011-01-13 06:35:41 <lfm> oh ok
932 2011-01-13 06:35:42 <nanotube> and i added another 10mhps for the cpu compute units included
933 2011-01-13 06:35:57 <nanotube> (just ballpark, cuz i have no idea what their 'compute units' are...
934 2011-01-13 06:35:59 <hacim> so it seems profitable
935 2011-01-13 06:36:02 <nanotube> so it seems that it is profitable
936 2011-01-13 06:36:03 <nanotube> yes
937 2011-01-13 06:36:06 <nanotube> unless we're missing something.
938 2011-01-13 06:36:23 <hacim> i've been meaning to try amazon, maybe I should give it a whirl
939 2011-01-13 06:36:33 <nanotube> you get 20 usd per 3.84 usd investment... with no up front costs of hardware
940 2011-01-13 06:36:52 <nanotube> hacim: give it a try and let us know. :) specifically what their cost per hour turns out to be, and what the hps performance is
941 2011-01-13 06:36:53 LobsterMan has joined
942 2011-01-13 06:36:53 LobsterMan has quit (Changing host)
943 2011-01-13 06:36:53 LobsterMan has joined
944 2011-01-13 06:36:53 <nanotube> heh
945 2011-01-13 06:36:55 <hacim> if we didn't miss something
946 2011-01-13 06:36:59 <grubles> every 4 days though
947 2011-01-13 06:37:04 <nanotube> yea, maybe it's not .04 per hour.
948 2011-01-13 06:37:16 <nanotube> grubles: 4 days at .04 per hour is only 3.84
949 2011-01-13 06:37:32 <nanotube> and you can join a pool, to smooth out your generation :)
950 2011-01-13 06:37:57 <nanotube> well, i was rather generous at .4 usd/btc. the highest bid is only ,,bc,mtgox
951 2011-01-13 06:37:57 <gribble> {"ticker":{"high":0.405,"low":0.3176,"vol":40283,"buy":0.3401,"sell":0.414,"last":0.405}}
952 2011-01-13 06:37:58 <xelister> can you join a pool and send fake hashes to "smooth out" your generation ;) ?
953 2011-01-13 06:38:02 <nanotube> .34, not .4
954 2011-01-13 06:38:15 <nanotube> xelister: no... that's been discussed at length.
955 2011-01-13 06:38:21 <xelister> wait, actually what the clients send, some easier hashes?
956 2011-01-13 06:38:28 <nanotube> pools are pretty good at weeding out fakes.
957 2011-01-13 06:38:30 <xelister> like >>4 the difficulty ?
958 2011-01-13 06:38:44 <lfm> hi is 0.405
959 2011-01-13 06:38:46 <nanotube> yea, slush pool asks for diff=1 hashes as shares
960 2011-01-13 06:39:21 <grubles> i suppose it is do-able with the rising price of bitcoin, but a block every 4 days is not going to net you a lot of coins
961 2011-01-13 06:39:28 <hacim> cluster GPU instances are actually $2.10/hr
962 2011-01-13 06:39:33 <grubles> and then the difficulty will probably rise which means you pay more for your cluster time
963 2011-01-13 06:39:36 <xelister> nanotube: so what is keeping users from sending easy hashes as agreed, but if they hit a winnig hard hash then just keep it? is it because they do not own the data needed to use the nonce?
964 2011-01-13 06:40:05 <hacim> thats $201.60 for four days
965 2011-01-13 06:40:34 <lfm> xelister right
966 2011-01-13 06:40:40 <xelister> so you get 20 USD for 200 USD invested
967 2011-01-13 06:40:44 <hacim> http://aws.amazon.com/ec2/#pricing
968 2011-01-13 06:40:59 acous has joined
969 2011-01-13 06:41:19 soultcer has quit (Ping timeout: 240 seconds)
970 2011-01-13 06:41:19 <annodomini> xelister: Right. The hash is the hash of the whole block, including the nonce. The whole block includes the coin minting transaction. You can't change the block after the fact; that invalidates the hash.
971 2011-01-13 06:41:53 LobsterMan has quit (Remote host closed the connection)
972 2011-01-13 06:41:53 <nanotube> xelister: yes the hash contains the pool's key in the 50btc gen tx. keeping the block is useless to the client.
973 2011-01-13 06:42:32 <nanotube> hacim: haha, 2.10 per hour
974 2011-01-13 06:42:44 <nanotube> a couple orders of magnitude off on the cost there we are hehe
975 2011-01-13 06:42:45 <nevezen> not worth the investment then?
976 2011-01-13 06:42:47 <hacim> i thought that didn't make sense
977 2011-01-13 06:42:58 <nanotube> ;;math calc 24*4*2.10
978 2011-01-13 06:42:58 <gribble> 201.6
979 2011-01-13 06:43:10 <nanotube> yea
980 2011-01-13 06:43:32 <nanotube> i guess bitcoins need to go up in price to about 4 usd/btc in order for it to be profitable
981 2011-01-13 06:43:45 <nanotube> (while difficulty stays the same)
982 2011-01-13 06:44:00 <nevezen> this faintly reminds me back when I was running dnet
983 2011-01-13 06:44:47 annodomini has quit (Quit: annodomini)
984 2011-01-13 06:44:55 LobsterMan has joined
985 2011-01-13 06:45:00 <nanotube> hacim: well, it was a nice thought. hehe
986 2011-01-13 06:45:29 <LobsterMan> hmm
987 2011-01-13 06:45:42 <LobsterMan> is 'PCIE spread spectrum' likely to have any significant impact?
988 2011-01-13 06:45:45 <LobsterMan> it was auto in my bios
989 2011-01-13 06:45:46 <nanotube> maybe if they were offering 2x5970 on the gpu clusters instead of the teslas.
990 2011-01-13 06:45:48 <LobsterMan> but i just set it to disabled
991 2011-01-13 06:47:32 <hacim> nanotube: yeah, i was all ready to give them a shot too :)
992 2011-01-13 06:47:41 <grubles> it says something about making it easier to overclock in my mobo's manual
993 2011-01-13 06:47:49 <hacim> there is no way I'm dropping $210 for four days of computing :)
994 2011-01-13 06:47:53 <grubles> overclock a gpu i mean
995 2011-01-13 06:48:19 <nevezen> yes, disable spread spectrum when overclocking (the cpu)
996 2011-01-13 06:49:37 <hacim> looks like someone tried this: http://www.bitcoin.org/smf/index.php?topic=1334.msg26185#msg26185
997 2011-01-13 06:49:42 <hacim> 53481 khash/s
998 2011-01-13 06:50:12 <xelister> hacim: I wonder who buys such clusters thou
999 2011-01-13 06:50:21 <hacim> seriously
1000 2011-01-13 06:50:40 LobsterMan has quit (Remote host closed the connection)
1001 2011-01-13 06:50:47 acous has quit (Quit: Computer has gone to sleep.)
1002 2011-01-13 06:52:04 <nevezen> scientific folks?
1003 2011-01-13 06:52:28 <nevezen> run their simulations during the day. let it run bitcoin during the wee hours of night
1004 2011-01-13 06:52:35 <nanotube> hacim: well, for the both of them, 112 mhps.
1005 2011-01-13 06:52:42 <nanotube> but still, quite a bit worse than we thought. :)
1006 2011-01-13 06:53:17 <nanotube> they do have a 'prepaid' cluster version, where you pay 5000 usd or so for a whole year.
1007 2011-01-13 06:53:27 <nanotube> at that rate, though... it's cheaper to just make your own. :)
1008 2011-01-13 06:54:12 <hacim> yeah, for real
1009 2011-01-13 06:54:43 <grubles> and you cant even play games on it :p
1010 2011-01-13 06:54:54 <nanotube> heh
1011 2011-01-13 06:55:38 <xelister> or warm your house
1012 2011-01-13 06:56:00 <hacim> FPGA doesn't help bitcoin?
1013 2011-01-13 06:56:24 <nanotube> there's been discussion of that as well. they cost a lot.
1014 2011-01-13 06:56:40 <nanotube> art did design and order a bunch of asics, though...
1015 2011-01-13 06:56:42 <xelister> hacim: ArtForzZz is building his own processor thing
1016 2011-01-13 06:56:46 <nanotube> we're all waiting to see how that turns out.
1017 2011-01-13 06:56:50 <nevezen> or you can hack your own
1018 2011-01-13 06:56:55 <nevezen> http://nsa.unaligned.org/
1019 2011-01-13 06:57:36 <grubles> nevezen, interesting....
1020 2011-01-13 07:01:27 <hacim> where can I find info about ArtForzZz's efforts?
1021 2011-01-13 07:01:47 <mrb_> NSA@home is old... the guy claims about 3.3 billion MD5 hashes/sec, which matches the performance of a single HD 5830 ($170)
1022 2011-01-13 07:02:39 <mrb_> and with a higher power consumption (240W) than a 5830 (175W)
1023 2011-01-13 07:04:17 <mrb_> current FPGAs offer maybe a bit better perf/W but worse perf/$ than GPUs on SHA256 workloads (Bitcoin)
1024 2011-01-13 07:04:33 <mrb_> 22:28 < grubles> nvidia are better at floating point operations which bitcoin does none of i believe
1025 2011-01-13 07:05:30 <mrb_> not exactly true. Nvidia has better double-precision fp perf, ATI GPUs kick the ass of Nvidia GPUs on single precision fp workloads just like on integer workloads
1026 2011-01-13 07:06:41 <mrb_> because Fermi can do 1 dp fp op every 2 cycles, whereas ATI can do 1 dp fp op only on 1 of the 4 ALUs (69xx) or 1 of the 5 ALUs (68xx/5xxx)
1027 2011-01-13 07:09:11 molecular has quit (Ping timeout: 255 seconds)
1028 2011-01-13 07:09:59 molecular has joined
1029 2011-01-13 07:10:37 <lfm> and 57xx and lower cant do any dp floating ops
1030 2011-01-13 07:12:54 FreeMoney has joined
1031 2011-01-13 07:13:07 <grubles> i stand corrected :)
1032 2011-01-13 07:14:54 <lfm> figuring out nvidia dp float support is at least as screwy
1033 2011-01-13 07:15:40 <lfm> basiclly 260 and later have dp but not 260m or any of the |m| (mobile) models
1034 2011-01-13 07:16:01 <mrb_> yep, 57xx can't do dp. but 58xx and 59xx can of course. so grubles is still incorrect :)
1035 2011-01-13 07:18:14 * mrb_ realizes he didn't understand the meaning of 'i stand corrected'
1036 2011-01-13 07:19:59 <xelister> mrb_: yeah once my buddy said that
1037 2011-01-13 07:20:08 <xelister> later I found out he took lots of blue pills
1038 2011-01-13 07:22:38 <nevezen> it's still a neat way to crunch numbers
1039 2011-01-13 07:23:19 <lfm> i think the whole opencl/cuda stuff is kinda messy myself
1040 2011-01-13 07:24:08 kisom_dev has quit (Ping timeout: 276 seconds)
1041 2011-01-13 07:25:03 <xelister> opencl?
1042 2011-01-13 07:25:23 james_ has quit (Ping timeout: 272 seconds)
1043 2011-01-13 07:26:14 kisom_dev has joined
1044 2011-01-13 07:27:29 james has joined
1045 2011-01-13 07:27:31 <lfm> gpu architechture is still in it infancy and gpu software support (OpenCL compiler etc) is worse
1046 2011-01-13 07:27:57 james is now known as Guest80686
1047 2011-01-13 07:30:42 <nevezen> more khash/s in the next few years with the same hardware? :p
1048 2011-01-13 07:31:18 <sipa> i doubt it
1049 2011-01-13 07:31:19 <lfm> naw they'll keep barging ahead with both hardware and software
1050 2011-01-13 07:31:25 <sipa> maybe a few %
1051 2011-01-13 07:32:12 <nevezen> I'm sure we'll find new uses for old hardware..
1052 2011-01-13 07:32:12 LobsterMan has joined
1053 2011-01-13 07:32:13 LobsterMan has quit (Changing host)
1054 2011-01-13 07:32:13 LobsterMan has joined
1055 2011-01-13 07:32:32 <lfm> the gpu will get closer integration with the cpu sorta like the fpu (8087 line) did
1056 2011-01-13 07:32:59 <nevezen> they tried that with pinetrail I think
1057 2011-01-13 07:33:04 <nevezen> in netbooks
1058 2011-01-13 07:33:19 <lfm> just a first attempt, its not the last
1059 2011-01-13 07:35:57 <LobsterMan> heh i think i broke something <_<
1060 2011-01-13 07:36:26 <LobsterMan> every time i try to start furmark now....my drivers crash, even with all oc'ing turned off
1061 2011-01-13 07:36:29 <LobsterMan> http://www.ozone3d.net/benchmarks/fur/
1062 2011-01-13 07:36:44 reQunix has joined
1063 2011-01-13 07:36:51 <lfm> do a full power down?
1064 2011-01-13 07:36:55 <LobsterMan> several times
1065 2011-01-13 07:37:10 <LobsterMan> maybe something got corrupted somewhere along the line
1066 2011-01-13 07:37:34 <LobsterMan> it seems like it crashes when furmark tries to read the clock speeds
1067 2011-01-13 07:37:42 <lfm> well overcloacking can do permanent damage
1068 2011-01-13 07:37:46 <nathan7> fur?
1069 2011-01-13 07:38:02 <LobsterMan> it starts up for a second, then when it would normally show me the clock speeds the program crashes, then my drivers crash
1070 2011-01-13 07:38:06 <LobsterMan> (see my link for furmark.....)
1071 2011-01-13 07:38:50 <LobsterMan> idk....rivatuner is working fine, and everest reports the correct clock speeds, and my system is still stable
1072 2011-01-13 07:40:08 <lfm> try some game or opengl screensaver?
1073 2011-01-13 07:40:18 <LobsterMan> i've tried counterstrike and it works fine
1074 2011-01-13 07:40:28 <LobsterMan> only furmark is crashing which is strange
1075 2011-01-13 07:40:31 <lfm> ok furmark is broken
1076 2011-01-13 07:41:01 necrodearia has quit (Read error: Operation timed out)
1077 2011-01-13 07:41:05 <LobsterMan> actually i have one more thing i think i can try....
1078 2011-01-13 07:41:05 <LobsterMan> i disabled pcie spread spectrum in bios....i wonder if that could be causing this; that seems to be when it started crashing
1079 2011-01-13 07:41:37 <lfm> geez that 4 way in your link seems to have no cooling for the middle cards!
1080 2011-01-13 07:41:50 * LobsterMan reboots one last time to see if anything is different with it set back to auto......lol
1081 2011-01-13 07:41:54 <LobsterMan> wish me luck
1082 2011-01-13 07:42:04 <lfm> bye
1083 2011-01-13 07:42:12 LobsterMan has quit (Remote host closed the connection)
1084 2011-01-13 07:42:57 <mrb_> I doubt high-end GPUs will ever be integrated in CPUs
1085 2011-01-13 07:43:53 HarryS has quit (Ping timeout: 260 seconds)
1086 2011-01-13 07:44:10 necrodearia has joined
1087 2011-01-13 07:44:29 <lfm> well they mioght tie em in closer than pcie bus like on the hyperlinks or something
1088 2011-01-13 07:45:00 <mrb_> graphics manufacturer are already constrained by the 300W TDP envelope of a double-width card. putting this on a single CPU socket would be an even smaller envelope
1089 2011-01-13 07:45:39 soultcer has joined
1090 2011-01-13 07:45:55 <lfm> smaller feature sizes will help power/heat limits
1091 2011-01-13 07:46:03 kiba` has joined
1092 2011-01-13 07:46:09 <mrb_> high-end Intel CPUs already have integrated PCIe lanes... how much 'closer' can you get?
1093 2011-01-13 07:46:11 kiba has quit (Read error: Operation timed out)
1094 2011-01-13 07:46:40 <xelister> if the balls arent touching, it's not compact
1095 2011-01-13 07:46:48 <lfm> hyperlinks like those used between cpu chips
1096 2011-01-13 07:47:55 <lfm> for better shareing of memory at least
1097 2011-01-13 07:48:19 <lfm> access to each other's memory
1098 2011-01-13 07:49:12 <mrb_> it's the Jevons paradox: make the die more power-efficient, and the manufacturers will simply find a way to put the 'saved' energy to new use, eg. packing more ALUs on their dies
1099 2011-01-13 07:49:21 sgornick has joined
1100 2011-01-13 07:49:52 <mrb_> and get back right to the previous power consumption levels
1101 2011-01-13 07:49:59 FreeMoney has quit (Quit: Page closed)
1102 2011-01-13 07:50:02 <lfm> could also be related to the peter principle
1103 2011-01-13 07:50:08 <mrb_> that's why all GPUs have plateau'd at 300W over the past 3 years
1104 2011-01-13 07:50:20 <xelister> it was common PSU limit?
1105 2011-01-13 07:50:21 LobsterMan has joined
1106 2011-01-13 07:50:22 LobsterMan has quit (Changing host)
1107 2011-01-13 07:50:22 LobsterMan has joined
1108 2011-01-13 07:50:36 <mrb_> despite constant improvement of the perf/W
1109 2011-01-13 07:50:42 <LobsterMan> meh i think furmark just sucks
1110 2011-01-13 07:50:43 <LobsterMan> lol
1111 2011-01-13 07:50:50 <mrb_> 300W is the max authorized by the PCIe specs
1112 2011-01-13 07:50:54 <mrb_> for double-width cards
1113 2011-01-13 07:50:59 <lfm> every watt has to be cooled too
1114 2011-01-13 07:51:23 <mrb_> 75W from the slot, 75W from the 6-pin power connector, and 150W from the 8-pin one
1115 2011-01-13 07:52:28 <lfm> so cant they use two 8 pins
1116 2011-01-13 07:53:00 <sipa> wow, 75W from the PCIe slot?
1117 2011-01-13 07:53:03 <mrb_> electrically speaking they could, but they would be violating the specs
1118 2011-01-13 07:53:40 <LobsterMan> ;;bc,stats
1119 2011-01-13 07:53:42 <gribble> Current Blocks: 102416 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 399 blocks | Next Difficulty In About: 2 days, 9 hours, 18 minutes, and 3 seconds | Next Difficulty Estimate: 18111.47884977
1120 2011-01-13 07:53:43 <mrb_> manufacturers do this in-house though for their GTX 595 prototypes (2 x 8-pin)
1121 2011-01-13 07:53:58 <lfm> gross amount of power anyway. they need to do more with less
1122 2011-01-13 07:54:31 <LobsterMan> ;;bc,calc 109100
1123 2011-01-13 07:54:34 <gribble> The average time to generate a block at 109100 Khps, given current difficulty of 16307.48285682 , is 1 week, 0 days, 10 hours, 19 minutes, and 40 seconds
1124 2011-01-13 07:54:45 <mrb_> ArtForz says the GTX 595 can't be dual-GF110 because of the excessive power consumption but I think it will. Nvidia will simply reduce the voltage and clocks until it its under 300W
1125 2011-01-13 07:54:51 <lfm> part of the reason the P4 sucked was it took too much power
1126 2011-01-13 07:54:56 <mrb_> s/its/fits/
1127 2011-01-13 07:56:30 <lfm> lobsterman so its running now?
1128 2011-01-13 07:59:13 <mrb_> same thing for CPUs. they have plateau'd at ~130W because that's 5.4 Amp on each wire of the 4-pin ATX12V connector which is dedicated to the CPU. You can do up to 6, 7 Amps without the wires heating too much but it requires 16 gauge which most PSUs don't have
1129 2011-01-13 08:05:41 <lfm> and intel got its arse kicked when amd was able to make the athlon 64 more efficient than the p4. big customers went amd in droves
1130 2011-01-13 08:06:07 soultcer has quit (Quit: ZNC - http://znc.sourceforge.net)
1131 2011-01-13 08:06:39 larsivi has joined
1132 2011-01-13 08:07:13 soultcer has joined
1133 2011-01-13 08:07:50 <mrb_> yep.
1134 2011-01-13 08:08:23 TheAncientGoat has joined
1135 2011-01-13 08:17:02 molecular has quit (Ping timeout: 240 seconds)
1136 2011-01-13 08:17:46 molecular has joined
1137 2011-01-13 08:18:50 <LobsterMan> lfm seems to be working
1138 2011-01-13 08:19:07 <lfm> wtg
1139 2011-01-13 08:19:33 <LobsterMan> from 684->720MHz and stable
1140 2011-01-13 08:21:12 <lfm> Average interval last 100 blocks: 7.66 min
1141 2011-01-13 08:21:32 andrew12 has quit (Read error: Connection reset by peer)
1142 2011-01-13 08:21:38 <lfm> someone is poring on the steam
1143 2011-01-13 08:28:41 davex__ has quit (Ping timeout: 265 seconds)
1144 2011-01-13 08:28:58 slush has quit (Quit: Leaving.)
1145 2011-01-13 08:35:41 hhgh has joined
1146 2011-01-13 08:36:03 hhgh has left ()
1147 2011-01-13 08:36:37 TD_ has joined
1148 2011-01-13 08:37:38 eca has joined
1149 2011-01-13 08:38:20 davex__ has joined
1150 2011-01-13 08:38:21 EvanR has quit (Ping timeout: 265 seconds)
1151 2011-01-13 08:38:34 TD_ has quit (Client Quit)
1152 2011-01-13 08:42:13 <LobsterMan> lfm this is the error i kept getting earlier
1153 2011-01-13 08:42:16 <LobsterMan> "Display driver nvlddmkm stopped responding and has successfully recovered."
1154 2011-01-13 08:42:20 <LobsterMan> http://pastebin.com/ztabL6UW
1155 2011-01-13 08:42:47 eca has left ()
1156 2011-01-13 08:42:49 <LobsterMan> apparently the nvidia cards have some sort of built in failsafe/watchdog, and something keeps triggering it when i don't think it should be triggered...
1157 2011-01-13 08:43:22 <lfm> I think you know more about that than I do
1158 2011-01-13 08:43:40 TD has joined
1159 2011-01-13 08:50:30 HarryS has joined
1160 2011-01-13 08:58:41 LobsterMan has quit (Remote host closed the connection)
1161 2011-01-13 09:01:06 LobsterMan has joined
1162 2011-01-13 09:01:07 LobsterMan has quit (Changing host)
1163 2011-01-13 09:01:07 LobsterMan has joined
1164 2011-01-13 09:01:21 satamusic_ has joined
1165 2011-01-13 09:03:22 satamusic has quit (Ping timeout: 240 seconds)
1166 2011-01-13 09:14:11 altamic has joined
1167 2011-01-13 09:23:19 darkskiez has joined
1168 2011-01-13 09:24:39 TD has quit (Remote host closed the connection)
1169 2011-01-13 09:24:45 TD has joined
1170 2011-01-13 09:30:03 darrob has quit (Disconnected by services)
1171 2011-01-13 09:30:11 darrob has joined
1172 2011-01-13 09:33:23 TheAncientGoat has quit (Ping timeout: 276 seconds)
1173 2011-01-13 09:37:32 <joe_1> party
1174 2011-01-13 09:37:45 <joe_1> bitcoin has been around 35 to 40 cents per coin all day
1175 2011-01-13 09:42:27 dirtyfilthy has joined
1176 2011-01-13 09:44:17 <dirtyfilthy> if two miners create two blocks of the same previous block, what makes one block chain more difficult than the other?
1177 2011-01-13 09:45:00 * niekie slowly waves at the channel.
1178 2011-01-13 09:45:04 <sipa> dirtyfilthy: nothing
1179 2011-01-13 09:45:09 <niekie> Hi guys, anyone remember me? :)
1180 2011-01-13 09:45:14 <sipa> they will have the same difficulty
1181 2011-01-13 09:45:21 <sipa> but blocks will get built up on both
1182 2011-01-13 09:45:33 <sipa> and typically, sooner on one than on the other
1183 2011-01-13 09:45:54 <sipa> and at that point the chain with the splitted block + one in is more "difficulty" than the one with only the split block
1184 2011-01-13 09:46:49 <lfm> dirtyfilthy you find out when the block AFTER that comes along and it will follow on from the winner
1185 2011-01-13 09:47:32 <sipa> and if the successor blocks to both arrive at approximately the same time, the block after that will determine the winner :)
1186 2011-01-13 09:47:38 <dirtyfilthy> ah right
1187 2011-01-13 09:47:44 <dirtyfilthy> ok
1188 2011-01-13 09:48:23 <dirtyfilthy> so after a chain split the transactions get merged somehow to the longest chain?
1189 2011-01-13 09:48:26 <lfm> its tree shapped, each block referances its predecessor and the longest branch becomes the trunc, the rest of the branches are abandoned
1190 2011-01-13 09:48:34 larsivi has quit (Ping timeout: 240 seconds)
1191 2011-01-13 09:48:40 larsivi has joined
1192 2011-01-13 09:49:07 <lfm> transactions lost to dropped branches get re-issued to be included in the main branch
1193 2011-01-13 09:49:18 <dirtyfilthy> ahhhh ok right
1194 2011-01-13 09:49:23 <MT`AwAy> or will be already included in the main branch
1195 2011-01-13 09:50:19 TheAncientGoat has joined
1196 2011-01-13 09:50:48 <dirtyfilthy> been reading about http://en.wikipedia.org/wiki/Linked_Timestamping and trying to get my head around it
1197 2011-01-13 09:50:51 <dirtyfilthy> thanks heaps
1198 2011-01-13 09:57:23 larsivi has quit (Ping timeout: 240 seconds)
1199 2011-01-13 09:58:20 dirtyfilthy has quit (Ping timeout: 240 seconds)
1200 2011-01-13 09:58:56 larsivi has joined
1201 2011-01-13 10:02:43 joe_8 has joined
1202 2011-01-13 10:03:25 joe_1 has quit (Ping timeout: 265 seconds)
1203 2011-01-13 10:04:40 joe_8 is now known as joe_1
1204 2011-01-13 10:04:59 altamic has quit (Ping timeout: 272 seconds)
1205 2011-01-13 10:05:31 altamic has joined
1206 2011-01-13 10:10:24 reQunix has quit (Quit: Computer has gone to sleep.)
1207 2011-01-13 10:14:00 larsivi has quit (Quit: No Ping reply in 180 seconds.)
1208 2011-01-13 10:14:20 larsivi has joined
1209 2011-01-13 10:21:41 AAA_awright has quit (Read error: Connection reset by peer)
1210 2011-01-13 10:24:54 larsivi has quit (Ping timeout: 260 seconds)
1211 2011-01-13 10:25:06 slush has joined
1212 2011-01-13 10:29:24 larsivi has joined
1213 2011-01-13 10:31:32 larsivi has quit (Read error: Operation timed out)
1214 2011-01-13 10:33:00 <UukGoblin> hi
1215 2011-01-13 10:33:15 <UukGoblin> just got a used 5970 which I'm pretty sure was flashed
1216 2011-01-13 10:33:22 <UukGoblin> and I'm getting
1217 2011-01-13 10:33:37 <UukGoblin> ERROR - Get temperature failed for Adapter 2 - ATI Radeon HD 5900 Series
1218 2011-01-13 10:35:31 <Keefe> how many cards in that system currently?
1219 2011-01-13 10:35:40 <lfm> possibly the temp is not hooked up on that board
1220 2011-01-13 10:36:07 <Keefe> can you read the clocks, just not the temps?
1221 2011-01-13 10:36:16 <lfm> 0, 1, 2?
1222 2011-01-13 10:37:28 <Keefe> yes, "Adapter 2" implies at least 3 chips. i just wonder how many more
1223 2011-01-13 10:40:20 <lfm> 0 is the cpu tho right? and 5970 is 2 adapters so maybe no more
1224 2011-01-13 10:41:15 altamic has quit (Ping timeout: 276 seconds)
1225 2011-01-13 10:42:00 <UukGoblin> Keefe, there were 2 cards, I'm gonna try with just the 1 soon
1226 2011-01-13 10:42:25 <UukGoblin> nope, 0 and 1 are the cores on the first 5970, 2 and 3 are the cores on the second 5970
1227 2011-01-13 10:43:01 <lfm> oh is your cpu single core?
1228 2011-01-13 10:43:01 <UukGoblin> basically, after flashing the bios and changing the voltage, I can't read the temperature off the card anymore
1229 2011-01-13 10:43:22 <UukGoblin> cpu does not appear in the OpenCL device list
1230 2011-01-13 10:43:33 <lfm> it did in mine
1231 2011-01-13 10:43:45 <lfm> linux 64?
1232 2011-01-13 10:44:41 <Keefe> UukGoblin: could you read temps before flashing? and can you still read clocks now?
1233 2011-01-13 10:45:09 <UukGoblin> lfm, linux64 indeed, sdk 2.1
1234 2011-01-13 10:45:47 <lfm> same! hmmm
1235 2011-01-13 10:46:10 <UukGoblin> Keefe, there were two cases when it happened. First when I flashed my card (increased voltage on it), and second when I bought this used card. In the first case yes, when I flashed back to the backup of the original bios, I could read the temps fine again
1236 2011-01-13 10:46:24 <UukGoblin> ubuntu 10.04 [lucid]
1237 2011-01-13 10:47:38 altamic has joined
1238 2011-01-13 10:48:12 <UukGoblin> ;;bc,stats
1239 2011-01-13 10:48:15 <gribble> Current Blocks: 102432 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 383 blocks | Next Difficulty In About: 2 days, 7 hours, 19 minutes, and 20 seconds | Next Difficulty Estimate: 18081.13229629
1240 2011-01-13 10:49:07 <sipa> today's total hashing speed estimate: 138GH/s
1241 2011-01-13 10:50:37 <Keefe> i recently had problems with "aticonfig --initial" when i made changes to the arrangement of cards. it wouldn't remove stale entries from xorg.conf, so i deleted xorg.conf and had aticonfig create a new one fresh
1242 2011-01-13 10:50:56 <UukGoblin> Keefe, I'll try that, it might be that. I forgot to recreate xorg.conf
1243 2011-01-13 10:51:10 <Keefe> i was getting similar errors not being able to read temps and clocks
1244 2011-01-13 10:51:20 <UukGoblin> ok, thanks, I'll try that.
1245 2011-01-13 10:51:50 <UukGoblin> it's gonna take a while as the monkey that's doing the hardware bits is not responding on the IM
1246 2011-01-13 10:53:55 larsivi has joined
1247 2011-01-13 10:54:17 <UukGoblin> hm, another matter - is it theoretically possible for nodes to only posses a truncated block chain and still function properly?
1248 2011-01-13 10:54:43 <UukGoblin> as in, not store transactions that happened looong ago in order to save space and scale better/
1249 2011-01-13 10:54:48 <sipa> sure, with lower correctness guarantees
1250 2011-01-13 10:55:35 <UukGoblin> (the truncation is to the early blocks, not the late ones of course)
1251 2011-01-13 10:56:36 <UukGoblin> hmm, if you created a special kind of document that has current balances of all accounts, then signed it, and sent out the signature as a transaction to the bitcoin network which would then be accepted in a block...
1252 2011-01-13 10:56:49 <UukGoblin> special transaction...
1253 2011-01-13 10:57:03 <UukGoblin> one that would be verified by the block generator...
1254 2011-01-13 10:57:07 <UukGoblin> hrm
1255 2011-01-13 10:57:12 <sipa> the bitcoin network doesn't know about balances
1256 2011-01-13 10:57:17 <sipa> only about transactions
1257 2011-01-13 10:58:15 <UukGoblin> ah, true...
1258 2011-01-13 10:58:32 <sipa> but transactions are verified in blocks anyway
1259 2011-01-13 10:58:54 <sipa> so if you only verify whether each block is a valid successor to the previous one, you have already quite some guarantees
1260 2011-01-13 10:59:02 <UukGoblin> well, so if a node had a truncated block chain, it couldn't validate future transactions...
1261 2011-01-13 10:59:22 <sipa> i believe that's what TD's android client does
1262 2011-01-13 10:59:37 <UukGoblin> it could only send new transactions and dumbly accept blocks...
1263 2011-01-13 11:00:03 <sipa> yes, exactly
1264 2011-01-13 11:00:15 <sipa> but it can verify the blocks without verifying the transactions in them
1265 2011-01-13 11:00:21 <TD> yes
1266 2011-01-13 11:00:27 <TD> that's indeed what my code does
1267 2011-01-13 11:00:29 <sipa> so it needs to trust the ones who send the blocks to him
1268 2011-01-13 11:00:44 <sipa> to only contain valid and verified transactions
1269 2011-01-13 11:01:11 <UukGoblin> right
1270 2011-01-13 11:01:17 <sipa> TD: you keep the complete block chain, or only the last few blocks, to be able to correct after a split?
1271 2011-01-13 11:01:58 <TD> well the current code i have on my laptop keeps the block chain but the way i plan to evolve it is to keep around a block locator in long term storage only
1272 2011-01-13 11:02:04 <TD> so a truncated block chain as UukGoblin said
1273 2011-01-13 11:02:52 <UukGoblin> http://www.aaronsw.com/weblog/squarezooko
1274 2011-01-13 11:03:32 larsivi has quit (Ping timeout: 240 seconds)
1275 2011-01-13 11:08:49 <niekie> Hey UukGoblin :)
1276 2011-01-13 11:08:57 <niekie> LTNS, hehe.
1277 2011-01-13 11:09:11 <omglolbbq1> should i post link to free site that just accepts donations in the 'we accept bitcoins' topic? or is it more for sales?
1278 2011-01-13 11:09:16 <UukGoblin> hi niekie
1279 2011-01-13 11:10:58 <necrodearia> wow comments @ http://blog.mises.org/12367/should-the-quantity-of-money-be-increased/
1280 2011-01-13 11:11:05 <necrodearia> Peter Surda and frank
1281 2011-01-13 11:13:24 <necrodearia> search keyword "bitcoin"
1282 2011-01-13 11:14:39 reQunix has joined
1283 2011-01-13 11:15:02 <UukGoblin> "i'm not a layman in this field - I have a phd from an Artificial Intelligence research group and know more about nodes and networks of nodes and such than most." LOLOL
1284 2011-01-13 11:15:12 <UukGoblin> I /love/ when people bring up their academic background
1285 2011-01-13 11:15:34 <UukGoblin> to try to prove to someone that they know what they're talking about
1286 2011-01-13 11:18:15 <joe_1> well at least he admits that he would be happy to be proved wrong
1287 2011-01-13 11:18:18 <UukGoblin> Keefe, thanks, it seems that regenerating the config /AND/ rebooting solved the problem
1288 2011-01-13 11:19:03 <Keefe> :)
1289 2011-01-13 11:19:27 <niekie> Heh, the Mining Pool is currently at 13370 Mhash/s :P
1290 2011-01-13 11:21:25 <omglolbbq1> lol UukGoblin, epic talk there :)
1291 2011-01-13 11:24:33 <Keefe> just one pool operator already controls 10% of the network?
1292 2011-01-13 11:27:01 <UukGoblin> Keefe, just one miner controls about 20% of it
1293 2011-01-13 11:27:11 <UukGoblin> need. more. growth.
1294 2011-01-13 11:27:23 * UukGoblin orders 2 gazillions more of 5970s ;-]
1295 2011-01-13 11:27:55 <Keefe> i'll be selling a 5970 soon
1296 2011-01-13 11:28:03 <UukGoblin> why?
1297 2011-01-13 11:28:34 grubles has quit (Ping timeout: 240 seconds)
1298 2011-01-13 11:28:34 <Keefe> ordered one more than i needed
1299 2011-01-13 11:28:43 <UukGoblin> whereabouts are you?
1300 2011-01-13 11:28:54 <Keefe> 6 5970's in 2 boxes is enough for me for now
1301 2011-01-13 11:29:02 <Keefe> southern california
1302 2011-01-13 11:29:08 <Keefe> usa
1303 2011-01-13 11:29:30 <UukGoblin> too far
1304 2011-01-13 11:29:36 * UukGoblin london
1305 2011-01-13 11:29:51 <Keefe> mm, was going to guess UK
1306 2011-01-13 11:30:50 <UukGoblin> Uuk is just the polish spelling of Ook actually
1307 2011-01-13 11:30:54 <Keefe> but if the cards are extra expensive there, it might still be worthwhile to buy mine and have it shipped fedex
1308 2011-01-13 11:30:57 <UukGoblin> I should probably change my nick some day
1309 2011-01-13 11:31:17 <UukGoblin> how much do you want for yours?
1310 2011-01-13 11:31:36 <UukGoblin> (in $, EUR, GBP or BTC)
1311 2011-01-13 11:31:55 <Keefe> 460 US$ worth of BTC, plus shipping
1312 2011-01-13 11:32:28 <UukGoblin> hmm it's not bad actually
1313 2011-01-13 11:32:40 <Keefe> expecting to have the card tomorrow. should be a new sealed sapphire
1314 2011-01-13 11:33:36 <Keefe> today actually. 3:30am here
1315 2011-01-13 11:33:50 <UukGoblin> allright, that's about 1300 BTC by my calculation
1316 2011-01-13 11:33:55 <UukGoblin> plus shipping
1317 2011-01-13 11:34:53 <Keefe> sounds about right
1318 2011-01-13 11:40:37 altamic has quit (Ping timeout: 255 seconds)
1319 2011-01-13 11:41:10 <UukGoblin> Keefe, ok, figure out shipping to London plz and I'm happy to buy it if it's allright :-]
1320 2011-01-13 11:42:41 <UukGoblin> niekie, btw, if you appear on #bitcoin-otc, I'm happy to give you a positive rating
1321 2011-01-13 11:42:57 altamic has joined
1322 2011-01-13 11:43:20 <Keefe> hmm, shipping 6 lb to london by fedex (2 business day delivery) costs about $125
1323 2011-01-13 11:44:07 <Keefe> i don't know the actual weight and box size right now though. will calc more accurately when it gets here
1324 2011-01-13 11:44:35 larsivi has joined
1325 2011-01-13 11:45:44 <UukGoblin> eek, that's a bit much actually
1326 2011-01-13 11:46:16 <UukGoblin> probably still a good deal for me though
1327 2011-01-13 11:47:19 grubles has joined
1328 2011-01-13 11:47:53 <Keefe> would you want the sapphire box. might save a bit to just ship the card in the anti-static bag in a fedex box (with foam peanuts of course)
1329 2011-01-13 11:48:00 <Keefe> ?*
1330 2011-01-13 11:48:30 <UukGoblin> nope, I wouldn't
1331 2011-01-13 11:54:09 <UukGoblin> Keefe, $520 total is what I can pay unfortunately
1332 2011-01-13 11:54:26 <UukGoblin> stuff on ebay is a bit cheaper than that
1333 2011-01-13 11:54:29 <UukGoblin> (locally)
1334 2011-01-13 11:55:55 <Keefe> i see. that probably won't be enough for me
1335 2011-01-13 11:56:21 noagendamarket has joined
1336 2011-01-13 11:56:37 <UukGoblin> check parcelforce, they might be a bit cheaper
1337 2011-01-13 12:00:17 <niekie> UukGoblin: ah. Thanks. Hadn't heard of that :)
1338 2011-01-13 12:02:06 <niekie> Been out of the Bitcoin stuff for quite a while. Heh.
1339 2011-01-13 12:02:33 <niekie> Now I'm trying to pick up mining a bit again for fun. Can't get any GPU workers to run though at the moment :(
1340 2011-01-13 12:02:51 <niekie> Well, I did get one to run for a small bit, but then it segfaulted.
1341 2011-01-13 12:03:12 <UukGoblin> try DiabloMiner
1342 2011-01-13 12:03:21 <niekie> That's exactly the one that segfaulted, hehe.
1343 2011-01-13 12:03:24 <UukGoblin> it works well on ATI / opencl with SDK 2.1
1344 2011-01-13 12:03:26 <UukGoblin> oh
1345 2011-01-13 12:03:54 <UukGoblin> feel free to bug Diablo-D3 here then ;-) he'll probably say it's your fault though ;-]
1346 2011-01-13 12:03:55 <omglolbbq1> does running bitcoin client without generating help the network in any way? verifying blockchain or anything..?
1347 2011-01-13 12:04:03 <niekie> UukGoblin: hehe. ;)
1348 2011-01-13 12:04:12 <niekie> Oh well, I've got 4 workers running :)
1349 2011-01-13 12:06:39 <joe_1> omglolbbq1 yes it provides moral support
1350 2011-01-13 12:06:40 <niekie> omglolbbq1: AFAIK, it will still pass blocks to other clients without generating.
1351 2011-01-13 12:06:53 <niekie> Which is kind of helping the network :)
1352 2011-01-13 12:07:06 <omglolbbq1> ok
1353 2011-01-13 12:08:36 larsivi has quit (Ping timeout: 265 seconds)
1354 2011-01-13 12:15:44 RazielZ has joined
1355 2011-01-13 12:16:45 larsivi has joined
1356 2011-01-13 12:21:08 larsivi has quit (Ping timeout: 240 seconds)
1357 2011-01-13 12:22:58 altamic has quit (Ping timeout: 240 seconds)
1358 2011-01-13 12:24:10 TD has quit (Ping timeout: 250 seconds)
1359 2011-01-13 12:25:40 devon_hillard has joined
1360 2011-01-13 12:29:12 mtgox has joined
1361 2011-01-13 12:30:50 larsivi has joined
1362 2011-01-13 12:36:57 larsivi has quit (Ping timeout: 276 seconds)
1363 2011-01-13 12:44:07 larsivi has joined
1364 2011-01-13 12:53:54 TD has joined
1365 2011-01-13 12:53:57 altamic has joined
1366 2011-01-13 12:58:04 altamic has quit (Ping timeout: 240 seconds)
1367 2011-01-13 12:59:36 altamic has joined
1368 2011-01-13 13:07:25 altamic has quit (Ping timeout: 255 seconds)
1369 2011-01-13 13:12:20 <UukGoblin> omglolbbq1, it'll pass blocks and transactions around afaik
1370 2011-01-13 13:12:51 <sipa> and verify them as well, i suppose
1371 2011-01-13 13:15:38 WonTu has joined
1372 2011-01-13 13:15:50 <UukGoblin> http://www.arctic.ac/en/p/detail?sArticle=19.%3F <- that cooler pwnz
1373 2011-01-13 13:15:51 WonTu has left ()
1374 2011-01-13 13:15:53 altamic has joined
1375 2011-01-13 13:16:04 <UukGoblin> 45/48 deg C @ 780MHz
1376 2011-01-13 13:20:04 larsivi has quit (Ping timeout: 255 seconds)
1377 2011-01-13 13:22:44 <slush> UukGoblin: And temperature of voltage regulators is?
1378 2011-01-13 13:23:09 <UukGoblin> can't measure, the radiator is attached to them if what my friend said was right
1379 2011-01-13 13:24:07 <slush> I asked to this cooler ArtForzZz yesterday and he told me it have big troubles with regulators by its design...
1380 2011-01-13 13:24:20 <grubles> the stock cooling is fine besides those little parts that over heat and clock down one of the gpus
1381 2011-01-13 13:24:35 <UukGoblin> I'd have to look a bit closer...
1382 2011-01-13 13:25:00 <slush> Personally I don't care about core temperature at all, my big troubles are with regulators
1383 2011-01-13 13:25:32 <slush> at 830MHz I was near 110oC
1384 2011-01-13 13:25:44 <UukGoblin> that's slightly insane
1385 2011-01-13 13:25:51 <UukGoblin> I /am/ worried about core temp too
1386 2011-01-13 13:26:42 <sipa> [13/01/2011 10:34] 000000000002A026CBE68AA91DA9E68C761EC02CB166279C5379B78E3882074E: score 24960.5; accepted
1387 2011-01-13 13:26:45 <sipa> \o/
1388 2011-01-13 13:27:06 <slush> sipa: which miner software?
1389 2011-01-13 13:27:16 <sipa> modified poclbm
1390 2011-01-13 13:33:07 <grubles> modified?
1391 2011-01-13 13:36:35 ArtForzZz is now known as ArtForz
1392 2011-01-13 13:36:40 omglolbbq1 has quit (Ping timeout: 255 seconds)
1393 2011-01-13 13:37:36 <ArtForz> VRM temp with stock cooler at 70%+ fanspeed is fine as long as you dont overvolt
1394 2011-01-13 13:38:26 altamic has quit (Ping timeout: 272 seconds)
1395 2011-01-13 13:38:36 <ArtForz> the problem is the VRMs that sit close to the ass end of the card, they're only cooled by a pretty small ribbed alu strut
1396 2011-01-13 13:38:53 <UukGoblin> oh, /them/ bastards.
1397 2011-01-13 13:39:07 <ArtForz> on the AC xtreme 5970 they use a seperate cooler for those, but that one is even worse than the stock coolers strut
1398 2011-01-13 13:39:43 <ArtForz> most waterblocks get it right
1399 2011-01-13 13:39:53 <ArtForz> = VRM temps well < 80°C
1400 2011-01-13 13:40:22 altamic has joined
1401 2011-01-13 13:40:34 <ArtForz> but yeah, generally you don't want to touch Vcore if you're running on air
1402 2011-01-13 13:40:44 skeledrew1 has quit (Quit: Instantbird 0.3a1pre)
1403 2011-01-13 13:41:00 <ArtForz> OCing from 725 to 850 didnt change VRM temps much in my test box
1404 2011-01-13 13:41:08 skeledrew has joined
1405 2011-01-13 13:41:19 <ArtForz> 6°C or so increase
1406 2011-01-13 13:41:50 tg has quit (Read error: Connection reset by peer)
1407 2011-01-13 13:42:01 <ArtForz> raising Vcore from 1.05 to 1.15 and core @ 900MHz -> VRM temps >100°C @ 70% fan
1408 2011-01-13 13:42:11 skeledrew has quit (Client Quit)
1409 2011-01-13 13:42:29 tg has joined
1410 2011-01-13 13:42:32 <ArtForz> vcore 1.1625 and 930MHz core -> ~110°C @ 100% fan
1411 2011-01-13 13:42:39 skeledrew has joined
1412 2011-01-13 13:43:35 <ArtForz> that was with a single 5970 on a bench with a 3kRPM 120x38mm fan pointed at it
1413 2011-01-13 13:44:24 altamic has quit (Ping timeout: 240 seconds)
1414 2011-01-13 13:44:41 <ArtForz> not to mention increasing Vcore kills power efficiency
1415 2011-01-13 13:45:26 <ArtForz> so if you want cheap and efficient, stick with the stock cooler and Vcore, set fan to 70 or 80% and increase core clock as fas as it'll go
1416 2011-01-13 13:45:36 <ArtForz> if you can't stand the noise, go watercooling
1417 2011-01-13 13:45:45 <grubles> sipa: what do you mean by modified?
1418 2011-01-13 13:45:52 <sipa> grubles: i changed some things myself
1419 2011-01-13 13:46:22 <grubles> care to specify? :[
1420 2011-01-13 13:46:31 <grubles> *:p
1421 2011-01-13 13:47:05 <ArtForz> I'd like to try if a thermalright VRM-R3 or VRM-R4 cooler would fit on a 5970
1422 2011-01-13 13:47:29 altamic has joined
1423 2011-01-13 13:47:30 <ArtForz> it looks like it should fit under a AC xtreme 5970
1424 2011-01-13 13:47:31 <sipa> i added asynchronous fetching (which was implemented by m0mchil himself as well), i made the python part verify the result of the opencl part (so it can show you the full hash and calculate its score), and i undid some changes to the opencl part which decreases performance
1425 2011-01-13 13:48:00 <ArtForz> and iirc the ass-end VRM section is identical between 5870 and 5970
1426 2011-01-13 13:48:41 <UukGoblin> mhm
1427 2011-01-13 13:48:46 <UukGoblin> ArtForz, cool. I mean, hot.
1428 2011-01-13 13:49:08 <ArtForz> btw, all this wouldnt be a problem if AMD werent cheap bastards
1429 2011-01-13 13:49:22 <slush> sipa: hash score was what I spotted in your post, nice feature
1430 2011-01-13 13:49:45 <UukGoblin> hrm
1431 2011-01-13 13:49:50 <ArtForz> if they would've populated all 4 core VRM phases per GPU we wouldnt have to deal with this crap
1432 2011-01-13 13:50:09 <UukGoblin> you know that oxidized thingy that holds the heatsink from the bottom side?
1433 2011-01-13 13:50:16 <ArtForz> backplate?
1434 2011-01-13 13:50:39 <UukGoblin> the difference between an asus and sapphire version is the material they put between it and the pcb
1435 2011-01-13 13:50:53 <UukGoblin> not the backplate, these square-ish thingies that come into it
1436 2011-01-13 13:51:01 <grubles> sipa, very interesting, so you squeezed more hashes/s out then?
1437 2011-01-13 13:51:07 <ArtForz> ahh, you mean the thermal pads
1438 2011-01-13 13:51:18 <UukGoblin> effectively once it gets heated for long enough, it might short circuit the board
1439 2011-01-13 13:51:23 <ArtForz> huh?
1440 2011-01-13 13:51:37 <sipa> grubles: not much :)
1441 2011-01-13 13:51:55 <sipa> and i think Diablo's miner is still a bit faster
1442 2011-01-13 13:51:58 <ArtForz> how would that happen, these things are designed to withstand 120°C 24/7
1443 2011-01-13 13:51:58 <UukGoblin> http://photos.pcpro.co.uk/blogs/wp-content/uploads/2009/11/ATIHD5970-01.jpg
1444 2011-01-13 13:51:58 <grubles> what card are you using?
1445 2011-01-13 13:52:11 <sipa> 4870
1446 2011-01-13 13:52:15 <UukGoblin> these 4-screw thingies that you see 2 of on that pic
1447 2011-01-13 13:52:26 <ArtForz> oh, you mean the X-plates
1448 2011-01-13 13:52:29 <UukGoblin> ArtForz, after a long time, this isolator between them and the pcb
1449 2011-01-13 13:52:52 <UukGoblin> might break
1450 2011-01-13 13:53:04 <ArtForz> I doubt it
1451 2011-01-13 13:53:32 <UukGoblin> basically the only difference between asus and sapphire is that bit of isolator
1452 2011-01-13 13:53:37 <UukGoblin> otherwise the stock card is identical
1453 2011-01-13 13:53:57 <UukGoblin> so I'm thinking that asus made it break a bit later than the cheaper sapphire version
1454 2011-01-13 13:54:02 <ArtForz> it's probably not even a difference between OEMs, mnore likely different dates of manufacture
1455 2011-01-13 13:54:37 <ArtForz> they also changed the "unloaded angle" = preload force of the x-plates at some point in time
1456 2011-01-13 13:55:18 <ArtForz> they also changed the inductors in the Vmem VRM
1457 2011-01-13 13:55:44 altamic has quit (Ping timeout: 240 seconds)
1458 2011-01-13 13:55:44 <ArtForz> = theres at least 3 different revisions of reference design cards
1459 2011-01-13 13:56:13 <brocktice> ArtForz: speaking of which, did you see those bios flash errors in -mining?
1460 2011-01-13 13:56:18 <ArtForz> yep
1461 2011-01-13 13:56:42 <ArtForz> now if I could remember what SSID there meant
1462 2011-01-13 13:56:54 <brocktice> yeah it keeps bringing to mind wifi
1463 2011-01-13 13:57:11 <ArtForz> yeah, but that doesnt make much sense :P
1464 2011-01-13 13:57:17 <brocktice> no, of course not
1465 2011-01-13 13:57:25 <ArtForz> did you back up those bioses?
1466 2011-01-13 13:57:28 <brocktice> oh, you can force flash
1467 2011-01-13 13:57:29 <brocktice> I did
1468 2011-01-13 13:57:35 <ArtForz> yep, just force flash
1469 2011-01-13 13:57:50 <brocktice> heh, although someone did that here with bad luck:
1470 2011-01-13 13:57:55 <brocktice> http://forums.techpowerup.com/showthread.php?t=86588
1471 2011-01-13 13:58:33 <brocktice> damn that ask at 0.51 is screwing up the whole depth chart on mtgox
1472 2011-01-13 13:58:44 <ArtForz> yeah, if it doesnt work you might have to put it in as 2ndary card and reflash back to backup
1473 2011-01-13 13:58:59 * brocktice shrugs, I modded the asus bios and it works fine
1474 2011-01-13 13:59:15 <ArtForz> ahh, yeah, iirc asus has different bios IDs
1475 2011-01-13 13:59:28 <ArtForz> thats what the asus voltage tool looks for :P
1476 2011-01-13 13:59:29 satamusic_ has quit (Quit: Leaving)
1477 2011-01-13 13:59:44 <brocktice> ArtForz: Do you know how to calculate power draw at load from vcore?
1478 2011-01-13 13:59:48 <ArtForz> nope
1479 2011-01-13 13:59:57 <brocktice> hmm
1480 2011-01-13 14:00:04 <ArtForz> you can measure it in windows
1481 2011-01-13 14:00:07 <brocktice> bah
1482 2011-01-13 14:00:12 <ArtForz> those VRMs also have digital current readout
1483 2011-01-13 14:00:22 <brocktice> I think I actually repartitioned my miner's HDD
1484 2011-01-13 14:00:27 <brocktice> but I haven't bothered installing windows yet
1485 2011-01-13 14:00:45 <brocktice> I'd have to make an installer USB stick or hook up a cdrom all ghetto-style
1486 2011-01-13 14:01:02 <ArtForz> yay for USB CD drives
1487 2011-01-13 14:03:34 <ArtForz> anyways, force flashing shouldnt cause a problem here
1488 2011-01-13 14:03:48 <ArtForz> they're all ref cards anyways
1489 2011-01-13 14:04:05 <ArtForz> with stuff like non-ref 5770s you have to be careful
1490 2011-01-13 14:04:32 davout has joined
1491 2011-01-13 14:04:56 <davout> hi
1492 2011-01-13 14:05:49 umonkey has joined
1493 2011-01-13 14:13:00 nevezen has quit (Quit: Lost terminal)
1494 2011-01-13 14:16:26 noagendamarket has quit (Ping timeout: 272 seconds)
1495 2011-01-13 14:16:50 <brocktice> mornin
1496 2011-01-13 14:18:17 <necrodearia> How long has it been since the difficulty has changed?
1497 2011-01-13 14:18:23 <necrodearia> Has it been over two weeks?
1498 2011-01-13 14:18:37 <sipa> no
1499 2011-01-13 14:18:37 <ArtForz> nope
1500 2011-01-13 14:18:39 <ArtForz> ;;bc,stats
1501 2011-01-13 14:18:41 <gribble> Current Blocks: 102464 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 351 blocks | Next Difficulty In About: 2 days, 1 hour, 37 minutes, and 39 seconds | Next Difficulty Estimate: 18169.86466419
1502 2011-01-13 14:18:47 <necrodearia> mm, http://nullvoid.org/bitcoin/difficultiez.php
1503 2011-01-13 14:19:00 <necrodearia> So, not yet, but close
1504 2011-01-13 14:19:15 <ArtForz> erm... not close by quite a way
1505 2011-01-13 14:19:24 <necrodearia> 4 days is 2 weeks
1506 2011-01-13 14:19:34 <sipa> if can only be over two weeks of the next difficulty estimate is lower than the current difficulty
1507 2011-01-13 14:19:49 <necrodearia> 3* days
1508 2011-01-13 14:19:54 <ArtForz> yep
1509 2011-01-13 14:20:03 <ArtForz> current diff started 2011-01-03 05:00 UTC
1510 2011-01-13 14:21:11 <ArtForz> so we're ~11.4% fast so far
1511 2011-01-13 14:21:43 <sipa> current actual speed is quite certainly over 140MH/s now
1512 2011-01-13 14:21:58 <ArtForz> doesnt look like it
1513 2011-01-13 14:22:18 <sipa> why not?
1514 2011-01-13 14:22:53 <ArtForz> last 100 blocks took 455.47 sec/block avg
1515 2011-01-13 14:23:48 <ArtForz> arrr... yeah
1516 2011-01-13 14:23:58 <ArtForz> about 153.8Gh/s
1517 2011-01-13 14:24:21 <brocktice> ;;bc,stats
1518 2011-01-13 14:24:24 <gribble> Current Blocks: 102465 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 350 blocks | Next Difficulty In About: 2 days, 1 hour, 17 minutes, and 30 seconds | Next Difficulty Estimate: 18178.32416989
1519 2011-01-13 14:24:34 <brocktice> ouch, it had stalled for a while
1520 2011-01-13 14:24:54 <ArtForz> well, ot the last 100 blocks were just unusually quick
1521 2011-01-13 14:24:56 <sipa> could even be close to 160GH/s, but that's only looking at very recent data, which can have large statistical variation
1522 2011-01-13 14:25:01 <ArtForz> yep
1523 2011-01-13 14:25:06 <ArtForz> too much noise over 100 blocks
1524 2011-01-13 14:25:39 <ArtForz> we already averaged well >150 for blocks 101519-101663
1525 2011-01-13 14:25:50 <sipa> my estimation algorithm says 151GH/s when using exponential window with tau=1 day
1526 2011-01-13 14:26:03 <sipa> and 140GH/s when using tau=10 days
1527 2011-01-13 14:26:13 <ArtForz> for the 144 block periods after that it was 122, 136, 134, 137, 135
1528 2011-01-13 14:26:28 <sipa> (that's not the average speed during that window though, it's an estimate for the current speed)
1529 2011-01-13 14:26:43 <ArtForz> yeah
1530 2011-01-13 14:26:46 <ArtForz> current 144-block period looks like it'll be >150 again
1531 2011-01-13 14:27:37 <ArtForz> question is whats a good model for hashrate change
1532 2011-01-13 14:28:01 <sipa> i use an exponential increase, where the average rate and the growth rate are parameters
1533 2011-01-13 14:28:28 <sipa> and i estimate those parameters using a maximum likelyhood estimators, applied to recent data with again an exponential window
1534 2011-01-13 14:28:44 <ArtForz> ... that sounds like a 2nd order poly
1535 2011-01-13 14:28:48 <xelister> now diff rise seems rather linear again?
1536 2011-01-13 14:29:14 <ArtForz> well, not linear
1537 2011-01-13 14:29:26 <ArtForz> I'm using a simple exponential model
1538 2011-01-13 14:29:33 <sipa> ArtForz: a*exp(b*t) is my model
1539 2011-01-13 14:29:45 <ArtForz> ahhh
1540 2011-01-13 14:31:12 <sipa> mathematical notation is so much clearer than english :)
1541 2011-01-13 14:31:52 <ArtForz> yep
1542 2011-01-13 14:32:44 <sipa> and my estimated growth rate for tau=1d is 6.4% currently, which is massive, but probably an probabilistic effect
1543 2011-01-13 14:32:55 <sipa> for tau=10d is 1.258%
1544 2011-01-13 14:33:02 <sipa> (daily growth rate, that is)
1545 2011-01-13 14:33:30 <ArtForz> doesnt MLE assume gaussian distribution?
1546 2011-01-13 14:33:53 <sipa> no, it estimates the parameters of a distribution
1547 2011-01-13 14:34:26 <sipa> i derived it myself for a poisson process with lambda=a*exp(b*t)
1548 2011-01-13 14:34:43 <ArtForz> yep, that looks right
1549 2011-01-13 14:35:28 <sipa> i described it briefly here: http://sipa.be/static/bitcoin/estim_exp_math.txt
1550 2011-01-13 14:36:28 <sipa> oh, that doesn't include the derivation for why those N and T variables can be used as MLE
1551 2011-01-13 14:37:18 <ArtForz> yeah, looks good
1552 2011-01-13 14:39:14 <sipa> for a block-window or gaussian window the formulas are a lot more complex
1553 2011-01-13 14:39:42 <ArtForz> I get 1.17%/day with a simple exponential best-fit over last 10 days
1554 2011-01-13 14:44:16 <sipa> close enough
1555 2011-01-13 14:44:45 <ArtForz> yup
1556 2011-01-13 14:44:57 <ArtForz> = a lot slower than what we used to see
1557 2011-01-13 14:45:49 <ArtForz> before diff 12252 or so we had a growth of well > 3%/day
1558 2011-01-13 14:46:34 <sipa> indeed
1559 2011-01-13 14:47:07 <ArtForz> not too surprising really
1560 2011-01-13 14:48:03 <ArtForz> though if prices of >0.35 hold I expect growth to pick up again
1561 2011-01-13 14:49:01 umonkey has quit (Ping timeout: 265 seconds)
1562 2011-01-13 15:00:20 Diablo-D3 has joined
1563 2011-01-13 15:05:53 altamic has joined
1564 2011-01-13 15:11:32 hundfred is now known as user
1565 2011-01-13 15:11:51 user is now known as hundfred
1566 2011-01-13 15:12:42 slush has quit (Quit: Leaving.)
1567 2011-01-13 15:17:56 Zarutian has joined
1568 2011-01-13 15:35:59 RazielZ has quit (Ping timeout: 240 seconds)
1569 2011-01-13 15:36:35 RazielZ has joined
1570 2011-01-13 15:42:17 altamic has quit (Ping timeout: 255 seconds)
1571 2011-01-13 15:58:01 welterde has quit (Read error: Operation timed out)
1572 2011-01-13 15:58:39 altamic has joined
1573 2011-01-13 16:04:52 slush has joined
1574 2011-01-13 16:05:39 duck1123 has quit (Quit: Ex-Chat)
1575 2011-01-13 16:06:21 altamic has quit (Ping timeout: 240 seconds)
1576 2011-01-13 16:11:35 welterde has joined
1577 2011-01-13 16:16:18 reQunix has quit (Remote host closed the connection)
1578 2011-01-13 16:20:11 <niekie> Yay, just reached my first "Confirmed Balance" on Bitcoin Pooled Mining :)
1579 2011-01-13 16:20:25 <niekie> Too bad it's just 0.0014174 BTC ;)
1580 2011-01-13 16:21:11 <lucky> :P
1581 2011-01-13 16:21:15 <lucky> cpu miner?
1582 2011-01-13 16:21:54 <niekie> Yeah :(
1583 2011-01-13 16:22:03 <niekie> 4 systems CPU mining.
1584 2011-01-13 16:22:21 <niekie> Of which 2 servers in some data centre somewhere, and 2 devices closer to me :)
1585 2011-01-13 16:22:39 <niekie> Actually, 2 VPS servers in some data centre somewhere.
1586 2011-01-13 16:22:50 <niekie> And yes, I asked my host, doesn't mind :D
1587 2011-01-13 16:23:20 mtgox has quit (Ping timeout: 265 seconds)
1588 2011-01-13 16:23:56 <davex__> when is the switch to 25 btc/hr?
1589 2011-01-13 16:23:58 <nanotube> niekie: are you paying for power for the "2 devices closer to you" ?
1590 2011-01-13 16:24:00 <davex__> er i mean block
1591 2011-01-13 16:24:05 <nanotube> davex__: block 210000
1592 2011-01-13 16:24:10 <nanotube> ;;bc,blocks
1593 2011-01-13 16:24:11 <gribble> 102481
1594 2011-01-13 16:24:14 <niekie> nanotube: not really :)
1595 2011-01-13 16:24:14 <davex__> ok
1596 2011-01-13 16:24:26 <nanotube> davex__: which is in another 2 years or so :)
1597 2011-01-13 16:24:29 <nanotube> niekie: ah ok :)
1598 2011-01-13 16:24:40 <lucky> i would have a spaceheater on anyway :P
1599 2011-01-13 16:24:58 TheAncientGoat has quit (Ping timeout: 276 seconds)
1600 2011-01-13 16:25:13 * niekie measures current power usage of device 1 closer to me.
1601 2011-01-13 16:25:34 <niekie> 41,2W
1602 2011-01-13 16:26:02 TheAncientGoat has joined
1603 2011-01-13 16:26:20 <ArtForz> lol, cute
1604 2011-01-13 16:29:16 <niekie> (measured with a computer program called "powertop" so not guaranteed to be accurate)
1605 2011-01-13 16:30:51 <niekie> niek@oneironaut:~$ acpi -t
1606 2011-01-13 16:30:51 <niekie> Thermal 0: ok, 69.0 degrees C
1607 2011-01-13 16:30:54 <niekie> Nice space heating :)
1608 2011-01-13 16:33:01 <ArtForz> http://pastebin.com/vGmJyHYf
1609 2011-01-13 16:34:47 <ArtForz> thats not a space heater, THIS is a space heater ;)
1610 2011-01-13 16:35:39 <niekie> Yo dawg, I herd you liek GPUs.
1611 2011-01-13 16:38:41 <ArtForz> one of the case fans in #4 is kinda fried, thats why it's running a lot hotter
1612 2011-01-13 16:39:25 <EvanR-work> i should get one of those for my apartment
1613 2011-01-13 16:39:29 <EvanR-work> it was friggin cold last night
1614 2011-01-13 16:39:46 <ArtForz> *checks room temp* 24.6°C
1615 2011-01-13 16:40:07 <EvanR-work> and id also be like 20% of bitcoin
1616 2011-01-13 16:40:27 <niekie> ArtForz: you should host a new YouTube show: "Cooking with Bitcoin". Give recipes and advice ;)
1617 2011-01-13 16:40:56 <niekie> I'd so subscribe to your channel.
1618 2011-01-13 16:42:23 <tcatm> yet another bug in RPC :/
1619 2011-01-13 16:43:20 <cosurgi> are we ahead or behined of schedule? (in generating coins)
1620 2011-01-13 16:43:26 <cosurgi> *behind
1621 2011-01-13 16:43:40 altamic has joined
1622 2011-01-13 16:43:52 <ArtForz> ;;bc,stats
1623 2011-01-13 16:43:54 <gribble> Current Blocks: 102483 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 332 blocks | Next Difficulty In About: 1 day, 22 hours, 56 minutes, and 28 seconds | Next Difficulty Estimate: 18200.35579437
1624 2011-01-13 16:44:21 <ArtForz> ahead, by about 11.6%
1625 2011-01-13 16:44:26 <cosurgi> thanks
1626 2011-01-13 16:44:45 <ArtForz> = estnextdiff / curdiff
1627 2011-01-13 16:44:55 <cosurgi> oh, wow. difficulty plans to grow a lot, I see.
1628 2011-01-13 16:45:01 <ArtForz> not really
1629 2011-01-13 16:45:14 <ArtForz> we used to have like 40% growth per 2016 blocks
1630 2011-01-13 16:45:14 <cosurgi> 45 bits now, right?
1631 2011-01-13 16:45:32 <cosurgi> oh. ok.
1632 2011-01-13 16:45:37 <ArtForz> 32 + log2(16307)
1633 2011-01-13 16:46:09 <ArtForz> looks like we'll hit maybe 16% this period
1634 2011-01-13 16:46:28 <ArtForz> which is a lot less than 40%
1635 2011-01-13 16:47:02 <cosurgi> ok, so 45.99 bits goes up to 46.15, errr :) Ok, so it's not much.
1636 2011-01-13 16:47:26 <cosurgi> I don't know why why difficulty is not given in bits, but using this formula.
1637 2011-01-13 16:47:41 <ArtForz> because we dont use bits internally
1638 2011-01-13 16:48:02 <cosurgi> how is that possible? I was thinking that it's all bits focued.
1639 2011-01-13 16:48:09 <ArtForz> nope
1640 2011-01-13 16:48:17 <ArtForz> we use 256 bit hash <= target
1641 2011-01-13 16:48:40 <ArtForz> with target written as 16-23 bits value, 8 bits exponent
1642 2011-01-13 16:48:53 <ArtForz> *binary exonent
1643 2011-01-13 16:48:56 <cosurgi> ok.
1644 2011-01-13 16:48:57 <ArtForz> *binary exponent
1645 2011-01-13 16:49:11 <cosurgi> Makes sense. otherwise there would be only 256 difficulty levels.
1646 2011-01-13 16:49:16 <ArtForz> yep
1647 2011-01-13 16:49:16 <cosurgi> not fine grained enough
1648 2011-01-13 16:49:35 <ArtForz> current target is 0x00000000000404CB000000000000000000000000000000000000000000000000
1649 2011-01-13 16:49:50 <nanotube> ,,bc,hextarget
1650 2011-01-13 16:49:50 <gribble> 00000000000404CB000000000000000000000000000000000000000000000000
1651 2011-01-13 16:50:01 <nanotube> ArtForz: for future ref, to save you copypaste :)
1652 2011-01-13 16:50:44 <ArtForz> aka 0x0404CB << 2**(8*0x18)
1653 2011-01-13 16:50:45 <cosurgi> tell me, why only first 128bits of input are changing? Or the latter 384 bits are changing too, but not too often?
1654 2011-01-13 16:50:53 altamic has quit (Ping timeout: 265 seconds)
1655 2011-01-13 16:50:55 <ArtForz> huh?
1656 2011-01-13 16:51:06 <cosurgi> ok. wait a sec. I'll look at code, and tell exatly what I mean
1657 2011-01-13 16:51:25 samfisher has joined
1658 2011-01-13 16:51:30 <samfisher> hi
1659 2011-01-13 16:51:40 <niekie> Hello samfisher.
1660 2011-01-13 16:52:20 <ArtForz> add 3 to the exponent part, shift it left 24 bits and you get nBits
1661 2011-01-13 16:52:30 <samfisher> I want to start a little website that sells handmade things
1662 2011-01-13 16:52:43 <samfisher> the problem is i lack php/html so i need some sources
1663 2011-01-13 16:53:15 <samfisher> i want to be able to add new content easily and there must be a shopping cart too
1664 2011-01-13 16:53:22 <samfisher> cand you guys guide me please?
1665 2011-01-13 16:54:18 <ArtForz> current nBits = 0x1b0404cb -> target is 0x0404cb << 2**(8*(0x1b - 3))
1666 2011-01-13 16:55:07 <ArtForz> 1.0 difficulty is defined as nBits 0x1d00ffff
1667 2011-01-13 16:55:15 <niekie> samfisher: I would recommend you check in with some current open source solutions like osCommerce for example.
1668 2011-01-13 16:55:53 <niekie> If you want to implement bitcoin transaction support in it coding skills would come in handy though.
1669 2011-01-13 16:57:03 <samfisher> niekie: yea, coding skills saves you from a lot of things
1670 2011-01-13 16:57:10 <samfisher> but as i said, i lack them
1671 2011-01-13 16:57:23 <samfisher> also i beleive you can't learn in a day or so
1672 2011-01-13 16:57:53 <niekie> samfisher: you can learn the basics in a few hours.
1673 2011-01-13 16:58:22 <niekie> Then just keep experimenting, and you'll learn a lot more :)
1674 2011-01-13 16:58:23 tg has quit (Read error: Connection reset by peer)
1675 2011-01-13 16:58:25 <sipa> mtgox at 0.4!
1676 2011-01-13 16:58:35 <samfisher> niekie: can you point me to a tutorial or somethng?
1677 2011-01-13 16:58:47 <ArtForz> wow
1678 2011-01-13 16:58:59 tg has joined
1679 2011-01-13 16:59:27 <niekie> samfisher: which language do you want to learn?
1680 2011-01-13 17:00:38 <cosurgi> ok. Here's a single hash run:
1681 2011-01-13 17:00:38 <cosurgi> TARGET : 00000000000404cb000000000000000000000000000000000000000000000000
1682 2011-01-13 17:00:38 <cosurgi> SHA2-IN: 5be0cd191f83d9ab9b05688c510e527fa54ff53a3c6ef372bb67ae856a09e667 (const)
1683 2011-01-13 17:00:41 <cosurgi> DATA : 00000280000000000000000000000000000000000000000000000000000000000000000000000000000000008000000000000001cb04041b7a2e2f4d094be77b
1684 2011-01-13 17:00:44 <cosurgi> MIDSTAT: 0a45ea0fea15aea361613d3429af008c441d2b000b2ad972179b6fec939762d0
1685 2011-01-13 17:00:47 <cosurgi> HASH__ : 5ccaecc91c72d62feffe22a4f677658a7c325aa6bc41dd59f97b5fa11ce83e35
1686 2011-01-13 17:00:49 <cosurgi> calling: runhash(HASH_1,DATA,MIDSTATE);
1687 2011-01-13 17:00:52 <cosurgi> calling: runhash(HASH ,HASH_1,SHA256_IN);
1688 2011-01-13 17:01:00 <cosurgi> ArtForz: here's my question - about 'DATA' it seems to never change. except for last 128 bits, where 32 of them are nonce.
1689 2011-01-13 17:01:10 <ArtForz> yeah
1690 2011-01-13 17:01:31 <ArtForz> it's backwards
1691 2011-01-13 17:01:31 <cosurgi> is it supposed to be constant?
1692 2011-01-13 17:01:37 <sipa> cosurgi: the block header is 80 bytes
1693 2011-01-13 17:01:43 <ArtForz> it's padding + message length in bits
1694 2011-01-13 17:02:02 <sipa> this is broken down to a block of 64 bytes, and another with 16 bytes remaining + padding + message length as ArtForz said
1695 2011-01-13 17:02:18 <niekie> samfisher: I'd recommend beginning by searching Google for the programming language of your choice + "hello world". Usually those sites will also have other programming examples in that language :)
1696 2011-01-13 17:02:33 <niekie> And with that I'm off for a while, folks! Dinner time, then chores/etc.
1697 2011-01-13 17:02:42 <ArtForz> the padding + size is fixed
1698 2011-01-13 17:03:01 <ArtForz> which is a neat way to optimize out quite a few parts of the first 16 rounds
1699 2011-01-13 17:03:48 <cosurgi> um, still not undersnading. The first 384 bits are really const? those are the padding?
1700 2011-01-13 17:04:04 <cosurgi> DATA : 000002800000000000000000000000000000000000000000000000000000000000000000000000000000000080000000________cb04041b7a2e2f4d094be77b
1701 2011-01-13 17:04:14 <cosurgi> I've put '_' where nonce stays.
1702 2011-01-13 17:04:32 <ArtForz> those first 3 DWRODs are the last 32 bits of merkle root, nTime, nBits
1703 2011-01-13 17:04:34 <cosurgi> The '28' at beginning and this '8' are padding / message length?
1704 2011-01-13 17:04:42 <ArtForz> the 28 is the *end*
1705 2011-01-13 17:05:11 <ArtForz> SHA2 padeding is message 80 00 00 00 00 .... (length of message in bits)
1706 2011-01-13 17:05:24 <ArtForz> 80000000 = start of padding
1707 2011-01-13 17:05:25 <cosurgi> oops. Ok, so it's backwards sorry. But in fact I only wanted to see TARGET in right order, and I used the same order when printing all the other stuff.
1708 2011-01-13 17:05:30 <nolan_d> Not sure if this is the place for client questions, but lots of the commands request an account in them. What is the account? Just some arbitrary string tied to an address?
1709 2011-01-13 17:05:32 <ArtForz> 00000280 = 80 bytes in bits
1710 2011-01-13 17:05:42 <sipa> nolan_d: basically, yes
1711 2011-01-13 17:05:49 freetx has quit (Quit: freetx)
1712 2011-01-13 17:05:50 <sipa> nolan_d: a name for a private key
1713 2011-01-13 17:06:17 <nolan_d> Ah, OK. Is there a format, or does any arbitrary string work? And what is the default account as shown in the GUI?
1714 2011-01-13 17:06:17 <ArtForz> it's not that simple
1715 2011-01-13 17:06:41 <ArtForz> arbitrary string, and you can assign multiple receiving keypairs to one account
1716 2011-01-13 17:06:55 <cosurgi> so, 00000280 and 80000000 are const, and this can be used for optimizing hasing?
1717 2011-01-13 17:07:05 <cosurgi> *hashing
1718 2011-01-13 17:07:07 <nolan_d> Ah, OK. So basically, I could do something like host multiple bitcoin accounts on a single server?
1719 2011-01-13 17:07:11 <ArtForz> yep
1720 2011-01-13 17:07:12 freetx has joined
1721 2011-01-13 17:07:14 <cosurgi> good!
1722 2011-01-13 17:07:14 <sipa> cosurgi: yes, and the zeroes in between too
1723 2011-01-13 17:07:18 <ArtForz> yep
1724 2011-01-13 17:07:35 <ArtForz> W[n] + K[n] with constant W[n] can be rolled into a single constant
1725 2011-01-13 17:08:11 <cosurgi> another question, where is the secondNonce? I mean, the part that server changes when seding getworks to client.
1726 2011-01-13 17:08:12 * sipa hopes the opencl compiles does this
1727 2011-01-13 17:08:13 <ArtForz> same goes for the first bunch of W calculation for rounds 16..31
1728 2011-01-13 17:08:31 <nolan_d> Cool. Has anyone done any work tying the addresses to more easily memorizable/passable strings? I know the point is anonymity in many instances, but it seems like there might be some benefit to, say, tying my own account to some sort of globally unique and memorable ID if I care about ease over anonymity.
1729 2011-01-13 17:08:36 <sipa> cosurgi: it's in the block itself, so only indirectly in the header through the merkle root
1730 2011-01-13 17:08:36 altamic has joined
1731 2011-01-13 17:08:51 <ArtForz> on ATI they do
1732 2011-01-13 17:08:51 <ArtForz> though not directly
1733 2011-01-13 17:08:51 <ArtForz> the IL -> shader ASM compiler is pretty good at extracting stuff like that
1734 2011-01-13 17:09:31 <nolan_d> Maybe something like webfinger might work for that, tying an account address to something that looks like an email address?
1735 2011-01-13 17:09:47 <ArtForz> not perfect, but decent
1736 2011-01-13 17:09:48 <ArtForz> = explicitly spelling out constant caluclation outside of the loop gained ~0.8%
1737 2011-01-13 17:10:09 <tcatm> nolan_d: You can encode your account as an QR code
1738 2011-01-13 17:10:17 <tcatm> s/account/address/
1739 2011-01-13 17:10:37 <cosurgi> sipa: which means that it is scattered over all the bits of MIDSTATE, right?
1740 2011-01-13 17:10:52 <nolan_d> Hmm, and then scan it with some device to do a transfer?
1741 2011-01-13 17:11:05 tylergillies has quit (Ping timeout: 240 seconds)
1742 2011-01-13 17:11:05 <sipa> cosurgi: midstate is the result of the first hashing the first 64 bytes, which contain most of the merkle root
1743 2011-01-13 17:11:13 <sipa> so, yes
1744 2011-01-13 17:11:38 <cosurgi> ok. and server also does hashing on 32 bits?
1745 2011-01-13 17:12:01 <Diablo-D3> s/most/half/
1746 2011-01-13 17:12:23 <Diablo-D3> merkle root is half on the first sha256 block, half on the second
1747 2011-01-13 17:12:32 <cosurgi> I mean - server just increments 32bit secondNonce, which makes all its way down to MIDSTATE ?
1748 2011-01-13 17:12:42 <ArtForz> nope
1749 2011-01-13 17:12:57 <ArtForz> 224 bits of merle root are in first block
1750 2011-01-13 17:12:59 <ArtForz> 32 are in 2nd
1751 2011-01-13 17:13:03 <Diablo-D3> ArtForz: rly?
1752 2011-01-13 17:13:04 <Diablo-D3> meh
1753 2011-01-13 17:13:07 <ArtForz> 32/224 != 1/2
1754 2011-01-13 17:13:47 <ArtForz> unless you use integer arithmetic, then both are 0...
1755 2011-01-13 17:14:23 <cosurgi> I'm asking about this, becasue I wonder if it's possible to extend all this 32bit hashing into 64bits, and include server's secondNonce in whole algorithm. 2^32 times mroee passes, but full possible range covered altogether.
1756 2011-01-13 17:14:23 <nolan_d> Another question, is it possible to have the same bitcoin in multiple places, or only in one? Wondering how it might work to, say, mine them on my desktop but still have a wallet on my phone if that ever becomes reality.
1757 2011-01-13 17:15:06 <ArtForz> cosurgi: whats the point, 32 bits of nonce are plenty
1758 2011-01-13 17:15:12 <nolan_d> Or is the solution some sort of loose federation of web services, with mobile/desktop apps sending API calls to arrange trades/transfer coins in from multiple sources?
1759 2011-01-13 17:15:12 <Diablo-D3> cosurgi: dont forget, you have three things moving
1760 2011-01-13 17:15:21 <Diablo-D3> cosurgi: time, nonce, and merkleroot
1761 2011-01-13 17:15:29 <ArtForz> yep
1762 2011-01-13 17:15:40 <ArtForz> time changes once a second in any non-braindead miner
1763 2011-01-13 17:15:45 <cosurgi> merkleroot are new transaction incoming, right?
1764 2011-01-13 17:15:55 <cosurgi> *transactions
1765 2011-01-13 17:16:03 <Diablo-D3> cosurgi: you can check 2^64 attempts before the block to check changes.
1766 2011-01-13 17:16:09 <Diablo-D3> the whole pool itself isnt even close to this
1767 2011-01-13 17:16:13 <ArtForz> or miners with different bnExtraNonces
1768 2011-01-13 17:16:34 <cosurgi> you mean "cant't" I guess? :)
1769 2011-01-13 17:16:35 <Diablo-D3> cosurgi: a new block found, or a new transaction added, resets he effort
1770 2011-01-13 17:16:41 <ArtForz> so you can have 32 bits of nNonce = 4Gh per second (nTime) per bnExtraNonce = merkleroot
1771 2011-01-13 17:18:13 <ArtForz> which is... plenty
1772 2011-01-13 17:21:26 <cosurgi> umm. Let me try to understand what you just said:... I can hash 32 bits of nonce per second if I have a speed of 4GHs/s ? Umm.. no I don't understand.
1773 2011-01-13 17:21:32 tylergillies has joined
1774 2011-01-13 17:21:33 tylergillies has quit (Changing host)
1775 2011-01-13 17:21:33 tylergillies has joined
1776 2011-01-13 17:22:01 <cosurgi> My current speed is 213229 khash/sec, btw.
1777 2011-01-13 17:22:16 <ArtForz> yep
1778 2011-01-13 17:23:20 <ArtForz> and currently thats pretty mcuh impossible with as signle miner
1779 2011-01-13 17:23:24 <cosurgi> so I'm still lost, at speed 213229 khash/sec I need 20 seconds to scan all 32 bits: 2^32/213254000 = 20.14 sec.
1780 2011-01-13 17:23:31 <ArtForz> yes
1781 2011-01-13 17:24:08 <cosurgi> 2^32/(4*10^9)=1 sec. Ah ok. I get it :)
1782 2011-01-13 17:24:28 <cosurgi> now the second part of your sentence 'per bnExtraNonce = merkleroot' ?
1783 2011-01-13 17:24:38 <cosurgi> merkleroot changes ever one second?
1784 2011-01-13 17:24:43 <ArtForz> no
1785 2011-01-13 17:24:47 <ArtForz> nTime changes every second
1786 2011-01-13 17:24:56 <sipa> merkleroot changes when the transactions themselves change
1787 2011-01-13 17:24:58 <ArtForz> every miner gets it's own merkleroot
1788 2011-01-13 17:25:08 <sipa> or when the coinbase (the second nonce) in it changes
1789 2011-01-13 17:25:18 <ArtForz> actrually It think with getwork every getwork gets its own merkleroot
1790 2011-01-13 17:25:52 <sipa> every getwork gets a random coinbase i think
1791 2011-01-13 17:25:54 <cosurgi> where is nTime? It is hased into `midstate`?
1792 2011-01-13 17:26:02 <cosurgi> *hashed
1793 2011-01-13 17:26:13 <sipa> cosurgi: nTime is the second 32 bits of the seconds 64 bytes of the block header
1794 2011-01-13 17:26:20 <ArtForz> yep
1795 2011-01-13 17:26:35 <ArtForz> it's just a unix timestamp usually
1796 2011-01-13 17:26:52 <sipa> so you can change it without needing to recalculate midstate
1797 2011-01-13 17:27:25 <cosurgi> DATA : 000002800000000000000000000000000000000000000000000000000000000000000000000000000000000080000000_nonce__===sec==7a2e2f4d094be77b
1798 2011-01-13 17:27:31 <ArtForz> yep
1799 2011-01-13 17:27:33 <cosurgi> here? '===sec=== ?
1800 2011-01-13 17:27:33 <ArtForz> no
1801 2011-01-13 17:27:34 <cosurgi> ok.
1802 2011-01-13 17:27:40 <ArtForz> thats nBits
1803 2011-01-13 17:27:55 <Diablo-D3> cosurgi: btw, if it takes you 20 seconds to make it through the nonce, its still going to be another 5 years before you get a gpu that can do it in less than a second
1804 2011-01-13 17:27:57 <ArtForz> 7a2e2f4d = timestamp
1805 2011-01-13 17:28:21 altamic has quit (Ping timeout: 240 seconds)
1806 2011-01-13 17:28:34 <cosurgi> DATA : 000002800000000000000000000000000000000000000000000000000000000000000000000000000000000080000000_nonce__cb04041b===sec==094be77b
1807 2011-01-13 17:28:34 <ArtForz> well, a single miner with 4 5970s is kinda getting glose
1808 2011-01-13 17:28:41 <ArtForz> yep
1809 2011-01-13 17:28:50 <ArtForz> 2.5Gh/s on air
1810 2011-01-13 17:28:59 <cosurgi> ok... but I cannot do hashing on timestamp, or they will notice that it's wrong :)
1811 2011-01-13 17:29:02 <ArtForz> I suspect on water >3Gh/s should be possible
1812 2011-01-13 17:29:08 <ArtForz> dunno
1813 2011-01-13 17:29:10 <cosurgi> how big deviations of timestamp are allowed?
1814 2011-01-13 17:29:29 <ArtForz> my miner happily accepts different timstamps as long as they're in the legal range
1815 2011-01-13 17:30:04 freetx has quit (Ping timeout: 240 seconds)
1816 2011-01-13 17:30:12 <ArtForz> nTime has to be > median timestamp of prev 11 blocks and has to be < network_adjusted_time + 2h
1817 2011-01-13 17:30:47 <cosurgi> which gives me at least 4 extra bits for hashing, doesn't it? :)
1818 2011-01-13 17:31:12 <ArtForz> which nobody needs
1819 2011-01-13 17:31:13 <cosurgi> or, is the timestamp replicated somewhere in merkleroot?
1820 2011-01-13 17:31:34 <cosurgi> yeah, I want to understand this
1821 2011-01-13 17:31:58 <ArtForz> each miner thread gets its own keypair or bnExtranonce = merkleroot anyways
1822 2011-01-13 17:32:13 <cosurgi> ok.
1823 2011-01-13 17:32:20 <ArtForz> so even if each miner thread is doing 4Gh/s, it still doesnt run out of nNonces per nTime
1824 2011-01-13 17:32:38 freetx has joined
1825 2011-01-13 17:33:14 <cosurgi> you mean - because it can also change bnExtranonce= merkleroot?
1826 2011-01-13 17:33:21 <nanotube> ArtForz: how is 'network adjusted time' calculated? you just poll all the nodes you're connected to for their local time?
1827 2011-01-13 17:33:32 <ArtForz> well, for what they think is UTC time
1828 2011-01-13 17:33:41 <nanotube> yea
1829 2011-01-13 17:33:44 <ArtForz> then you take the median of that
1830 2011-01-13 17:33:45 <nanotube> and then median it?
1831 2011-01-13 17:33:47 <cosurgi> Diablo-D3: did you make some new relase of your miner since three weeks?
1832 2011-01-13 17:33:48 <nanotube> ah ok :)
1833 2011-01-13 17:33:50 <nanotube> thanks ArtForz :)
1834 2011-01-13 17:34:16 <Diablo-D3> cosurgi: check git.
1835 2011-01-13 17:34:19 <ArtForz> thats why we have nTime in version() message I think
1836 2011-01-13 17:34:20 <cosurgi> I am thinking about re-doing yout GPU miner but in C++, and adding it as an option to jgrazik's miner.
1837 2011-01-13 17:34:38 <Diablo-D3> I think my last commit was on the 1st but I could be wrong
1838 2011-01-13 17:34:47 <cosurgi> ok :)
1839 2011-01-13 17:34:49 <Diablo-D3> cosurgi: c++ is pointless.
1840 2011-01-13 17:34:52 <Diablo-D3> it wont make it faster.
1841 2011-01-13 17:35:03 <cosurgi> it's learning :)
1842 2011-01-13 17:35:11 <Diablo-D3> oh, and remember what license mine is under
1843 2011-01-13 17:35:14 <nanotube> ArtForz: ah... and how often are the other nodes polled for their timestamp?
1844 2011-01-13 17:35:15 <Diablo-D3> I expect it to be followed.
1845 2011-01-13 17:35:23 <ArtForz> they arent polled
1846 2011-01-13 17:35:24 <nanotube> ArtForz: whenever a block is generated?
1847 2011-01-13 17:35:26 <cosurgi> I mean - I'm learning. Never programmed in OpenCL.
1848 2011-01-13 17:35:41 <ArtForz> version is sent once, on connect
1849 2011-01-13 17:35:48 <nanotube> ah
1850 2011-01-13 17:35:51 <cosurgi> Diablo-D3: oh, sorry, I didn't look in fact. Your license doesn't allow me doing this?
1851 2011-01-13 17:36:22 <Diablo-D3> cosurgi: it means if you intend on publicly releasing it, and giving it to jgrazik, both you and him need to comply with the license
1852 2011-01-13 17:36:33 <cosurgi> OK. that's good.
1853 2011-01-13 17:36:40 <cosurgi> That's what I always do :)
1854 2011-01-13 17:36:49 <nanotube> ArtForz: and then my node stores those values, adjusted for time passed since the information was received, and uses the median as the net-at?
1855 2011-01-13 17:36:57 <ArtForz> yep
1856 2011-01-13 17:36:59 <ArtForz> well, not quite
1857 2011-01-13 17:36:59 freetx has quit (Ping timeout: 240 seconds)
1858 2011-01-13 17:37:00 <nanotube> neato
1859 2011-01-13 17:37:03 <nanotube> mm
1860 2011-01-13 17:37:05 <cosurgi> Will take care to read it, if I will reach this stage.
1861 2011-01-13 17:37:12 <ArtForz> at the time you receive vbersion you store gettime() - nTime
1862 2011-01-13 17:37:27 newsham has joined
1863 2011-01-13 17:37:31 <ArtForz> you keep those for all connected nodes in a list and take the median
1864 2011-01-13 17:37:51 <ArtForz> whenever a node is added, you add a offset to the list
1865 2011-01-13 17:37:52 <nanotube> right, makes sense. cool.
1866 2011-01-13 17:37:55 <ArtForz> I dont think they ever get removed
1867 2011-01-13 17:37:59 <cosurgi> ah, that's GPLv3, I use it everywhere :)
1868 2011-01-13 17:38:55 <cosurgi> $ git pull
1869 2011-01-13 17:38:58 <cosurgi> .../com/diablominer/DiabloMiner/DiabloMiner.java | 295 +++++++++-----------
1870 2011-01-13 17:39:02 <cosurgi> 5 files changed, 140 insertions(+), 160 deletions(-)
1871 2011-01-13 17:39:38 newsham has quit (Client Quit)
1872 2011-01-13 17:39:48 darrob has quit (Ping timeout: 240 seconds)
1873 2011-01-13 17:39:55 <Diablo-D3> cosurgi: the patch format has never been too clever.
1874 2011-01-13 17:42:02 davout has quit (Quit: i <3 pork (http://dev.ojnk.net))
1875 2011-01-13 17:44:05 <cosurgi> hmm... about #define sharound, that part: ((a&b)|(c&(a|b)), it can be replaced with (a & b) ^ (c & (a ^ b)), almost the same, but isn't xor faster?
1876 2011-01-13 17:44:16 <cosurgi> that's 'Maj'.
1877 2011-01-13 17:44:49 <ArtForz> no
1878 2011-01-13 17:44:51 <cosurgi> At least I know that xor conserves energy, and leads to less heating of GPU, because no electrons are lost: the same number of input and output '1' bits/
1879 2011-01-13 17:45:07 <ArtForz> it's exactly as fast as ((a & b) | (c & (a | b)))
1880 2011-01-13 17:45:15 <cosurgi> ok :)
1881 2011-01-13 17:46:00 <ArtForz> though if you have special boolean opcodes you can optimize some things a bit
1882 2011-01-13 17:46:16 <cosurgi> which parts?
1883 2011-01-13 17:46:21 <nolan_d> Is there a registry of bitcoin projects/libraries anywhere? Seeing what is being/has been done might help me get a better handle on some of this.
1884 2011-01-13 17:46:30 <ArtForz> like for example if you have NAND, NOR, AND-NOT or similar silliness
1885 2011-01-13 17:46:58 <cosurgi> oh, right. That's processor dependent. I've no idea if ATI has this.
1886 2011-01-13 17:47:23 <ArtForz> nope
1887 2011-01-13 17:47:30 <ArtForz> but ATI has BFI_INT
1888 2011-01-13 17:48:02 <cosurgi> what is it?
1889 2011-01-13 17:48:19 <nanotube> nolan_d: no 'registry'... but there's a not-necessarily-complete list of bitcoin clients https://en.bitcoin.it/wiki/Category:Clients
1890 2011-01-13 17:48:33 <nolan_d> Cool, thanks.
1891 2011-01-13 17:48:38 <ArtForz> (a&b)|((!a)&c)
1892 2011-01-13 17:48:47 <ArtForz> err... ~a
1893 2011-01-13 17:48:56 <ArtForz> (a&b)|((~a)&c)
1894 2011-01-13 17:49:32 <ArtForz> and thats... Ch() in a single opcode
1895 2011-01-13 17:49:34 <cosurgi> I've seen this one!
1896 2011-01-13 17:49:44 <cosurgi> right, it's this one :)
1897 2011-01-13 17:49:58 <ArtForz> you can also use it do do Maj() in 3 opcodes instead of 4
1898 2011-01-13 17:50:16 <cosurgi> why Diablo isn't doing this?
1899 2011-01-13 17:50:25 <ArtForz> because you can't get to it from OpenCL
1900 2011-01-13 17:50:33 <cosurgi> uhh
1901 2011-01-13 17:50:41 <ArtForz> you can't get to it from IL either
1902 2011-01-13 17:50:51 <cosurgi> so, how to get to it?
1903 2011-01-13 17:51:02 <ArtForz> binary patching of shader ASM
1904 2011-01-13 17:51:12 <cosurgi> have you done that?
1905 2011-01-13 17:51:23 <ArtForz> mrb_ has
1906 2011-01-13 17:51:41 <cosurgi> what was the % of improvement?
1907 2011-01-13 17:51:54 <ArtForz> at least he did for whitepixel
1908 2011-01-13 17:52:01 <ArtForz> should gain 4% or so iirc
1909 2011-01-13 17:52:14 nevezen has joined
1910 2011-01-13 17:52:15 <cosurgi> nice
1911 2011-01-13 17:53:11 grondilu has joined
1912 2011-01-13 17:53:53 <cosurgi> Diablo-D3: can you help me here a bit? I was using your .zipped release so far, I want to run the git version, but it seems that I need to do some compiling first.
1913 2011-01-13 17:54:17 <cosurgi> I get this:
1914 2011-01-13 17:54:26 <cosurgi> $ ./DiabloMiner-Linux.sh --user=bitcoin --pass=bitcoinbasswordException in thread "main" java.lang.NoClassDefFoundError: com/diablominer/DiabloMiner/DiabloMiner
1915 2011-01-13 17:54:30 <cosurgi> Caused by: java.lang.ClassNotFoundException: com.diablominer.DiabloMiner.DiabloMiner
1916 2011-01-13 17:54:33 <cosurgi> etc..
1917 2011-01-13 17:54:53 <ArtForz> my current inner loop is 820 VLIW clauses, with BFI_INT it's 779... so 5%
1918 2011-01-13 17:55:29 <cosurgi> Exception in thread "main" java.lang.NoClassDefFoundError: com/diablominer/DiabloMiner/DiabloMiner
1919 2011-01-13 17:55:32 <cosurgi> Caused by: java.lang.ClassNotFoundException: com.diablominer.DiabloMiner.DiabloMiner at java.net.URLClassLoader$1.run(URLClassLoader.java:202) at java.security.AccessController.doPrivileged(Native Method)
1920 2011-01-13 17:56:25 <cosurgi> ArtForz: so you patched this too. Is there some help somewhere about how to use mrb_'s patch?
1921 2011-01-13 17:56:37 <ArtForz> whitepixel is FOSS
1922 2011-01-13 17:56:47 <cosurgi> ok.
1923 2011-01-13 17:56:55 <ArtForz> I use a slightly different method though
1924 2011-01-13 17:57:36 <ArtForz> btw, without BFI_INT I'm getting 556.5 Mh/s on a stock 5970
1925 2011-01-13 17:57:39 <cosurgi> what do you do?
1926 2011-01-13 17:58:07 <ArtForz> I use a opcode I dont use anywhere and just replace that one with BFI_INT
1927 2011-01-13 17:58:39 <cosurgi> and 584.3 Mh/s with BFI_INT?
1928 2011-01-13 17:58:50 <ArtForz> somewhere around there, on a stock card
1929 2011-01-13 17:59:10 <cosurgi> hmm... I would get 223
1930 2011-01-13 17:59:16 <ArtForz> ~660Mh/s @ 820
1931 2011-01-13 17:59:38 <cosurgi> @ 820 means overclocked?
1932 2011-01-13 17:59:45 <ArtForz> yep
1933 2011-01-13 17:59:51 <cosurgi> ok. :)
1934 2011-01-13 17:59:51 <ArtForz> 820MHz core, stock is 725
1935 2011-01-13 18:00:02 <cosurgi> how do you overclock it?
1936 2011-01-13 18:00:06 <ArtForz> aticonfig
1937 2011-01-13 18:00:08 <ArtForz> my 5970s average OC is ~825
1938 2011-01-13 18:00:14 <cosurgi> can I do this from command line?
1939 2011-01-13 18:00:17 <ArtForz> yep
1940 2011-01-13 18:00:31 <cosurgi> heheh. what's the command? :-))
1941 2011-01-13 18:00:47 <ArtForz> check out --od-get-clocks and --od-set-clocks
1942 2011-01-13 18:01:22 <ArtForz> aticonfig --pplib-cmd "set fanspeed 0 X" is also kinda useful where X is speed in %
1943 2011-01-13 18:01:41 <ArtForz> pplib-cmd is weird though, it doesnt honor --adapter but instead uses DISPLAY
1944 2011-01-13 18:02:00 <cosurgi> r$ aticonfig --adapter=0 --od-getclocks
1945 2011-01-13 18:02:01 <cosurgi> Adapter 0 - ATI Radeon HD 5800 Series
1946 2011-01-13 18:02:01 <cosurgi> Core (MHz) Memory (MHz)
1947 2011-01-13 18:02:01 <cosurgi> Current Clocks : 157 300
1948 2011-01-13 18:02:01 <cosurgi> Current Peak : 825 1050
1949 2011-01-13 18:02:03 <cosurgi> Configurable Peak Range : [600-1000] [1050-1450]
1950 2011-01-13 18:02:05 <cosurgi> GPU load : 0%
1951 2011-01-13 18:02:10 <cosurgi> I just stopped mining, to run aticonfig :)
1952 2011-01-13 18:02:17 popey has left ()
1953 2011-01-13 18:02:21 <ArtForz> you also have --od-gettemp iirc
1954 2011-01-13 18:02:24 <ArtForz> or so
1955 2011-01-13 18:02:29 <ArtForz> short hand is --odgt
1956 2011-01-13 18:02:33 <cosurgi> can I fiddle with aticonfing while hashing - is it safe?
1957 2011-01-13 18:02:37 <ArtForz> yep
1958 2011-01-13 18:03:14 <ArtForz> aticonfig --adapter=all --odgt
1959 2011-01-13 18:03:14 <ArtForz> Adapter 0 - ATI Radeon HD 5700 Series
1960 2011-01-13 18:03:14 <ArtForz> Sensor 0: Temperature - 79.00 C
1961 2011-01-13 18:03:14 <ArtForz> Adapter 1 - ATI Radeon HD 5900 Series
1962 2011-01-13 18:03:14 <ArtForz> Sensor 0: Temperature - 72.00 C
1963 2011-01-13 18:03:16 <ArtForz> Adapter 2 - ATI Radeon HD 5900 Series
1964 2011-01-13 18:03:18 <ArtForz> Sensor 0: Temperature - 70.50 C
1965 2011-01-13 18:03:20 <ArtForz> Adapter 3 - ATI Radeon HD 5700 Series
1966 2011-01-13 18:03:22 <ArtForz> Sensor 0: Temperature - 85.00 C
1967 2011-01-13 18:03:24 <ArtForz> Adapter 4 - ATI Radeon HD 5900 Series
1968 2011-01-13 18:03:26 <ArtForz> Sensor 0: Temperature - 58.00 C
1969 2011-01-13 18:03:28 <ArtForz> Adapter 5 - ATI Radeon HD 5900 Series
1970 2011-01-13 18:03:30 <ArtForz> Sensor 0: Temperature - 56.00 C
1971 2011-01-13 18:04:05 <cosurgi> wow, aticonfig has plenty of options
1972 2011-01-13 18:04:10 AAA_awright has joined
1973 2011-01-13 18:04:36 <cosurgi> aticonfig --pplib-cmd "set fanspeed 0 100" is what I want, I guess?
1974 2011-01-13 18:04:41 freetx has joined
1975 2011-01-13 18:04:43 <nevezen> that's not all in one computer, is it?
1976 2011-01-13 18:04:49 <ArtForz> that is
1977 2011-01-13 18:04:56 <nevezen> wow
1978 2011-01-13 18:05:05 <cosurgi> how many PCIex16 slots?
1979 2011-01-13 18:05:07 <nevezen> so 5x pcie slots?
1980 2011-01-13 18:05:07 <ArtForz> 4
1981 2011-01-13 18:05:10 <ArtForz> 2 5970s and 2 5770s
1982 2011-01-13 18:05:18 <ArtForz> MSI 790FX-GD70
1983 2011-01-13 18:05:32 <ArtForz> without risers
1984 2011-01-13 18:05:38 <ArtForz> the 5770s are running passively cooled
1985 2011-01-13 18:05:55 <ArtForz> and they're short enough to not restrict the intake fan on the 5970s
1986 2011-01-13 18:06:26 <cosurgi> hmm.. what clock would you recommend for me?
1987 2011-01-13 18:06:26 <ArtForz> and yeah, the one sandwiched between the 5970s is running kinda hot... *shrug*
1988 2011-01-13 18:06:30 <ArtForz> no clue
1989 2011-01-13 18:06:52 <cosurgi> should I scale memory speed linearly with core clock?
1990 2011-01-13 18:07:01 <ArtForz> I keep mem speed at stock
1991 2011-01-13 18:07:06 <cosurgi> ok.
1992 2011-01-13 18:07:08 <ArtForz> increase core clock in 10MHz steps until the card crashes, back down 20Mhz
1993 2011-01-13 18:07:32 <ArtForz> you usually have to hard reset the box if a card crashes
1994 2011-01-13 18:07:43 <cosurgi> I see.
1995 2011-01-13 18:08:10 <ArtForz> oh, and 100% fan speed on 5850/70/5970 is... impressive
1996 2011-01-13 18:08:25 <ArtForz> aka "DAMN THATS LOUD"
1997 2011-01-13 18:08:39 <cosurgi> heh :) I didn't try yet.
1998 2011-01-13 18:08:47 <ArtForz> iirc using auto-speed the fan never goes over 60%
1999 2011-01-13 18:08:49 <cosurgi> let me first set fan to 100% then try to overclock :)
2000 2011-01-13 18:09:03 <cosurgi> can I set fan back to 'auto' ?
2001 2011-01-13 18:09:06 <ArtForz> nope
2002 2011-01-13 18:09:13 <ArtForz> you have to reboot for that :/
2003 2011-01-13 18:09:21 <cosurgi> err..
2004 2011-01-13 18:09:33 <ArtForz> there is a command that supposed to do it, buit instead it sets fan speed to 0% (whoops)
2005 2011-01-13 18:09:48 grondilu has quit (Quit: leaving)
2006 2011-01-13 18:10:07 <nevezen> are they all for bitcoin usage? :)
2007 2011-01-13 18:10:12 <ArtForz> yep
2008 2011-01-13 18:10:31 <ArtForz> I run 24 5970s, 5 5770s, a 6870 and a 6970
2009 2011-01-13 18:10:32 <nevezen> how much MH/s/card?
2010 2011-01-13 18:10:43 <ArtForz> total is 16.5Gh/s
2011 2011-01-13 18:10:59 <ArtForz> using ~9kW
2012 2011-01-13 18:11:09 * nevezen shakes head
2013 2011-01-13 18:11:15 <nevezen> that's too much
2014 2011-01-13 18:11:21 <grubles> but can it run crysis............zing
2015 2011-01-13 18:11:52 <cosurgi> how do I check current fanspeed?
2016 2011-01-13 18:12:07 <ArtForz> aticonfig --pplib-cmd 'get fanspeed 0'
2017 2011-01-13 18:12:12 Zarutian has left ()
2018 2011-01-13 18:12:16 <cosurgi> Result: Fan Speed: 60%
2019 2011-01-13 18:12:17 <cosurgi> ok.
2020 2011-01-13 18:12:37 <cosurgi> wow, that's damn loud :)
2021 2011-01-13 18:12:43 <cosurgi> I mean, 100%
2022 2011-01-13 18:12:50 <ArtForz> yep
2023 2011-01-13 18:13:02 <ArtForz> 80% is not quite as annoying and nearly as good as 100%
2024 2011-01-13 18:13:11 <nevezen> is the return good though?
2025 2011-01-13 18:13:19 <ArtForz> unless you remove the rear bracket, then 100% actually does help a bit
2026 2011-01-13 18:13:23 <ArtForz> lets see
2027 2011-01-13 18:13:47 <ArtForz> total hadrware cost fso far is ~$22k
2028 2011-01-13 18:14:17 <grubles> so how many coins is that per day then?
2029 2011-01-13 18:14:23 <ArtForz> income from selling bitcoins so far... $31k
2030 2011-01-13 18:14:57 <ArtForz> about 1kBTc at current difficulty
2031 2011-01-13 18:15:18 <cosurgi> do you use only mgtox for selling? What channels do you use for selling?
2032 2011-01-13 18:15:47 <ArtForz> bcm, mtgox, bc, private otc trades
2033 2011-01-13 18:16:02 <xelister> modern phones suck cock
2034 2011-01-13 18:16:08 <xelister> Old phone:
2035 2011-01-13 18:16:11 <xelister> - make fucking phone calls
2036 2011-01-13 18:16:15 <cosurgi> do you use paypal for transfers?
2037 2011-01-13 18:16:15 <xelister> "Modern" phone:
2038 2011-01-13 18:16:18 <ArtForz> nope
2039 2011-01-13 18:16:20 <xelister> - stupid java games
2040 2011-01-13 18:16:25 <xelister> - blowjobs - yes
2041 2011-01-13 18:16:28 <cosurgi> what do you use?
2042 2011-01-13 18:16:28 <ArtForz> I fucking hate PayPal
2043 2011-01-13 18:16:30 <xelister> - multimedia shit - yes
2044 2011-01-13 18:16:33 <ArtForz> usually LR$
2045 2011-01-13 18:16:38 <xelister> - option to end a conversation - nope
2046 2011-01-13 18:16:47 <ArtForz> also LREUR and Pecunix
2047 2011-01-13 18:16:52 <xelister> - opiton to not hang randomly after each 20-30 convesations - nope
2048 2011-01-13 18:16:54 <cosurgi> ok. PayPal did something bad to you, wrt to bitcoins?
2049 2011-01-13 18:17:05 <ArtForz> no, years before that
2050 2011-01-13 18:17:21 <luke-jr> cosurgi: PayPal is notoriously evil
2051 2011-01-13 18:17:21 <cosurgi> ok. I understand, everyone hates them.
2052 2011-01-13 18:17:26 <ArtForz> froze my account with ~3kEUR in it
2053 2011-01-13 18:17:45 <cosurgi> I ask, because - I have not enough options for doing international transfers.
2054 2011-01-13 18:17:56 <ArtForz> well, actually they froze my account and still allowed incoming payments, fuckers
2055 2011-01-13 18:18:00 <xelister> ArtForz: own'd
2056 2011-01-13 18:18:03 <xelister> why they froze it
2057 2011-01-13 18:18:08 <luke-jr> ArtForz: I doubt you could make that $9k profit if you started the same today
2058 2011-01-13 18:18:13 <ArtForz> "suspected money laundering"
2059 2011-01-13 18:18:20 <xelister> PEDORRIST!
2060 2011-01-13 18:18:49 <luke-jr> anyhow, looks like I'm getting one of those new Sandy Bridge CPU+GPUs
2061 2011-01-13 18:18:49 <xelister> how about a joint bounty to assasinate Flash, Silverlight and Paypal CEOs
2062 2011-01-13 18:19:03 <ArtForz> sandy bridge CPUs are fucking PWNage
2063 2011-01-13 18:19:08 <ArtForz> GPUs... not so much
2064 2011-01-13 18:19:10 <cosurgi> wow. 0.4 mtgoxUSD
2065 2011-01-13 18:19:11 <xelister> * and Ati software team
2066 2011-01-13 18:19:14 <nevezen> wow, that's big of them to accuse you of
2067 2011-01-13 18:19:34 <ArtForz> the fun part is you pretty much have no recourse
2068 2011-01-13 18:20:04 <nevezen> what caused them to think that way?
2069 2011-01-13 18:20:24 <ArtForz> probably me receiving about $50k from the US over the course of ~6 months
2070 2011-01-13 18:20:24 <luke-jr> ArtForz: I'm waiting for a review of the GPU before I checkout
2071 2011-01-13 18:20:27 <cosurgi> wow, I already earned enough BTC to repay this GPU that I bought a month ago. This will come true when I sell those BTC
2072 2011-01-13 18:20:33 <luke-jr> ArtForz: supposedly the drivers were just released stable
2073 2011-01-13 18:20:40 <ArtForz> well, I *did* get my money back after a threatening letter from my laywer
2074 2011-01-13 18:21:04 <xelister> ArtForz: how much time effort?
2075 2011-01-13 18:21:13 <ArtForz> took about 4 months
2076 2011-01-13 18:21:56 <ArtForz> fucking paranoid not-a-bank
2077 2011-01-13 18:22:08 sipa1024 has joined
2078 2011-01-13 18:22:51 <cosurgi> overclocked to 900 MHz, got 232122 khash/sec now. still not crashing.
2079 2011-01-13 18:23:01 <nevezen> he threatened you? what?
2080 2011-01-13 18:23:11 <ArtForz> no
2081 2011-01-13 18:23:18 <luke-jr> ArtForz: SB CPU worth mining on?
2082 2011-01-13 18:23:21 <cosurgi> no, ArtForz's lawyer contacted paypal, and he was serious :)
2083 2011-01-13 18:23:26 <ArtForz> yep
2084 2011-01-13 18:23:39 <luke-jr> single or quad threads better?
2085 2011-01-13 18:24:05 <ArtForz> pretty much "look, if you suspect something illegal, report it to the cops, the feds and the IRS, still doesnt give you the right to freeze funds indefinetly"
2086 2011-01-13 18:25:08 <luke-jr> ArtForz: but that's how they make their money
2087 2011-01-13 18:25:08 <ArtForz> and yes, that money was actually from a nearly legit business and reported as income
2088 2011-01-13 18:25:14 <ArtForz> exactly
2089 2011-01-13 18:25:18 <luke-jr> nearly legit? :P
2090 2011-01-13 18:25:25 <ArtForz> yep
2091 2011-01-13 18:25:28 <sipa1024> loi 'nearly legit'
2092 2011-01-13 18:25:40 <luke-jr> you make it sound as if true legit is rare
2093 2011-01-13 18:25:45 <luke-jr> and illegit the norm
2094 2011-01-13 18:26:02 <ArtForz> well, gold farming isn't exactly cut-and-dry legit
2095 2011-01-13 18:26:37 sipa has quit (Disconnected by services)
2096 2011-01-13 18:26:43 sipa1024 is now known as sipa
2097 2011-01-13 18:27:14 <cosurgi> Diablo-D3: I symlinked 'target' directory from your zipped release into git's directory. And now miner starts without complaining about missing classes. But am I running now a new version, or an old version of your miner?
2098 2011-01-13 18:27:29 <sipa> well you're creating virtual goods, and selling them for an agreed-upon price on the internet
2099 2011-01-13 18:27:34 <ArtForz> yep
2100 2011-01-13 18:27:41 <ArtForz> it's not illegal, but against the TOS
2101 2011-01-13 18:27:44 <cosurgi> I am not sure what is compiled .java and if .cl is compiled automatically everytime, or beforehand.
2102 2011-01-13 18:27:53 <sipa> everytime
2103 2011-01-13 18:28:59 <ArtForz> my bots also didnt use/modify/... the game client in any way, so DMCA violation ala Blizzard is also out
2104 2011-01-13 18:29:26 <ArtForz> clientless bots are so much more fun to write
2105 2011-01-13 18:30:00 <ArtForz> though breaking GameGuard was even more fun (popular anti-cheat/hack used by korean MMOs)
2106 2011-01-13 18:30:02 <luke-jr> PayPal TOS forbids selling information?
2107 2011-01-13 18:30:17 <ArtForz> PayPal TOs forbids getting paid for coding?
2108 2011-01-13 18:30:25 <ArtForz> because thats what it was
2109 2011-01-13 18:30:40 <ArtForz> I wrote the bots and sold em to goldfarmers
2110 2011-01-13 18:30:41 <luke-jr> ArtForz: DMCA doesn't forbid modifying stuff
2111 2011-01-13 18:31:10 <ArtForz> well, see blizzard vs. x over wowglider
2112 2011-01-13 18:31:51 <luke-jr> oh, against game TOS?
2113 2011-01-13 18:31:52 <ArtForz> somehow modifying client code in memory is creating a derivative work, and distributing the tools to do that is illegal under the DMCA?!?
2114 2011-01-13 18:31:58 <ArtForz> yep
2115 2011-01-13 18:32:15 <ArtForz> which of course still "only" makes it a civil matter
2116 2011-01-13 18:32:27 <luke-jr> well, that only matters if you use it :p
2117 2011-01-13 18:32:30 <luke-jr> not if you sell it
2118 2011-01-13 18:32:33 <ArtForz> yep
2119 2011-01-13 18:32:43 <nevezen> finally recieved bitcoin portions from the pool
2120 2011-01-13 18:32:48 <ArtForz> so in my book what I did was perfectly legal ;)
2121 2011-01-13 18:32:54 <cosurgi> ArtForz: are there other useful multi PCIex16 mobos?
2122 2011-01-13 18:32:58 <ArtForz> yes
2123 2011-01-13 18:33:03 <cosurgi> this one is out of stock, here
2124 2011-01-13 18:33:27 <ArtForz> but actually cheapest option is to go with cheap C2D based boards with 2* x16 PCIe
2125 2011-01-13 18:33:48 <cosurgi> what is C2D ?
2126 2011-01-13 18:33:54 <nevezen> so essentially, generating bitcoins has become profitable for you?
2127 2011-01-13 18:33:59 <ArtForz> Core2Duo
2128 2011-01-13 18:34:02 <cosurgi> k.
2129 2011-01-13 18:34:06 <ArtForz> just throw on a celery and the smallest ram you can find
2130 2011-01-13 18:34:15 <ArtForz> yep
2131 2011-01-13 18:34:17 <luke-jr> why not 4*?
2132 2011-01-13 18:34:44 <nevezen> is it enough to cover electricity consumption too? :)
2133 2011-01-13 18:34:54 <ArtForz> because the cheapest 4-slot board+cpu+ram is more expensive than 2 cheap 2-slot boards + 2 cpu + 2 ram
2134 2011-01-13 18:34:58 <ArtForz> yep
2135 2011-01-13 18:35:02 <ArtForz> by a few 100%
2136 2011-01-13 18:35:18 <ArtForz> and in winter I pay 0 for electricity
2137 2011-01-13 18:35:30 <ArtForz> cogeneration is fun
2138 2011-01-13 18:35:36 <nevezen> and that mini farm keeps you warm too
2139 2011-01-13 18:35:39 <ArtForz> yep
2140 2011-01-13 18:35:53 <nevezen> how do you deal with the noise though?
2141 2011-01-13 18:36:00 <ArtForz> you learn to ignore it
2142 2011-01-13 18:36:09 <ArtForz> though now that it's above freezing I'll have to move a few boxes to grid power... which sucks
2143 2011-01-13 18:36:18 <nevezen> not in your basement, out of sight out of mind?
2144 2011-01-13 18:36:22 <ArtForz> nope
2145 2011-01-13 18:36:41 <luke-jr> ArtForz: some people can't ignore it
2146 2011-01-13 18:36:44 <ArtForz> ground floor, at the bottom of the staircase so the heat nicely dissipates over 1st and 2nd floor
2147 2011-01-13 18:37:01 <luke-jr> ArtForz: my roommate in college insisted he couldn't sleep with it, and kept killing the power -.-
2148 2011-01-13 18:37:05 <ArtForz> I pay ~ $0.3/kWh here for grid power
2149 2011-01-13 18:37:57 <ArtForz> which is the main reason I have a micro cogeneration plant
2150 2011-01-13 18:38:06 <sipa> what's cogeneration?
2151 2011-01-13 18:38:10 jgarzik has quit (Changing host)
2152 2011-01-13 18:38:10 jgarzik has joined
2153 2011-01-13 18:38:22 <ArtForz> using heat+electricity from a generator
2154 2011-01-13 18:38:47 <cosurgi> and what is a generator?
2155 2011-01-13 18:38:59 <ArtForz> basically replacing central heating oil furnace with a diesel genset with heat exchangers
2156 2011-01-13 18:39:03 <luke-jr> fail
2157 2011-01-13 18:39:21 <ArtForz> a generator usually generates electricity
2158 2011-01-13 18:39:25 <nevezen> more enviromentally friendly? hehe
2159 2011-01-13 18:39:30 <ArtForz> yep
2160 2011-01-13 18:39:33 <luke-jr> cosurgi: bitcoin mining, but for electricity :P
2161 2011-01-13 18:39:48 <ArtForz> it's about 95% efficient
2162 2011-01-13 18:39:54 <cosurgi> and people say that bitcoins are more expensive than electricity
2163 2011-01-13 18:40:06 <ArtForz> remember, you're USING the "waste" heat
2164 2011-01-13 18:40:17 <luke-jr> ArtForz: not here
2165 2011-01-13 18:40:19 <nevezen> what about in the summer?
2166 2011-01-13 18:40:28 <ArtForz> yeah, thats the problem
2167 2011-01-13 18:40:28 <luke-jr> ArtForz: our generator is too far from the house :P
2168 2011-01-13 18:40:36 grubles has quit (Ping timeout: 240 seconds)
2169 2011-01-13 18:40:39 <luke-jr> (and it rarely gets too cold)
2170 2011-01-13 18:40:41 <ArtForz> in summer I'm forced to either run on grid power or shit down miners
2171 2011-01-13 18:40:45 <ArtForz> *shuit
2172 2011-01-13 18:40:47 <ArtForz> LOL
2173 2011-01-13 18:40:59 <cosurgi> buy a freezer :))
2174 2011-01-13 18:41:03 <cosurgi> a large one :)))
2175 2011-01-13 18:41:10 <ArtForz> but yeah, thats another reason why I want to increase efficiency
2176 2011-01-13 18:41:20 <nevezen> think if you did that, you'd have to worry about condensation
2177 2011-01-13 18:41:41 <sipa> ArtForz: with BTC prices going up, still no plans to invest in further gpu's?
2178 2011-01-13 18:42:02 <kiba`> GPU is just a way to buy bitcoin with dollars
2179 2011-01-13 18:42:06 <sipa> since the benefit of asics is lower when the btc/power price is high
2180 2011-01-13 18:42:26 <ArtForz> hmmm... not really
2181 2011-01-13 18:42:43 <ArtForz> for further ASIC runs I can pretty much match 5970 for $/Mh
2182 2011-01-13 18:42:47 <kiba`> my bad..homestead
2183 2011-01-13 18:42:51 <Diablo-D3> [01:23:16] <cosurgi> Diablo-D3: I symlinked 'target' directory from your zipped release into git's directory. And now miner starts without complaining about missing classes. But am I running now a new version, or an old version of your miner?
2184 2011-01-13 18:42:58 <Diablo-D3> cosurgi: dont do that, that does nothing
2185 2011-01-13 18:43:05 <cosurgi> hmm.. I am particularly unlucky it seems. It's 5 days and still no new block. At 213 MHs (just overclocked to 236 MHs)
2186 2011-01-13 18:43:09 <cosurgi> Diablo-D3: glad you are back.
2187 2011-01-13 18:43:10 <Diablo-D3> cosurgi: undo all of that, and run mvn package
2188 2011-01-13 18:43:16 <cosurgi> Diablo-D3: so what should I do?
2189 2011-01-13 18:43:22 <Diablo-D3> cosurgi: git doesnt contain binaries, only the source
2190 2011-01-13 18:43:28 <lucky> cosurgi, you should join a pool :P
2191 2011-01-13 18:43:59 <sipa> funny how people (me included) always say join "a" pool
2192 2011-01-13 18:44:04 <sipa> when there is only one :)
2193 2011-01-13 18:44:11 <Diablo-D3> didnt there used to be a second?
2194 2011-01-13 18:44:11 <cosurgi> Diablo-D3: what is mvn package? is it in debian repository, just waiting to install?
2195 2011-01-13 18:44:30 <sipa> puddinpop's doesn't exist anymore i think
2196 2011-01-13 18:44:34 <Diablo-D3> cosurgi: apt-get install maven2
2197 2011-01-13 18:44:34 <lucky> sipa, indeed ;P
2198 2011-01-13 18:44:54 <nevezen> there's only one? slush's?
2199 2011-01-13 18:44:59 <cosurgi> ok.. downloading..
2200 2011-01-13 18:45:00 <sipa> yes
2201 2011-01-13 18:45:10 <lucky> what's funny is how much money slush probably makes off it. host one dinky server, i bet most people opt to give him a 2% donation. 2% of easily 10% of generated blocks these days :P
2202 2011-01-13 18:45:40 <cosurgi> $ git status
2203 2011-01-13 18:45:40 <cosurgi> # On branch master
2204 2011-01-13 18:45:40 <cosurgi> nothing to commit (working directory clean)
2205 2011-01-13 18:45:46 <lucky> a bitcoin an hour, roughly.
2206 2011-01-13 18:45:53 <ArtForz> I still dont quite get what the point of putting GPUs in a pool is though
2207 2011-01-13 18:46:08 <lucky> ArtForz, roughly the same return, but more constant :p
2208 2011-01-13 18:46:10 <nevezen> I think you get results
2209 2011-01-13 18:46:25 <cosurgi> Diablo-D3: ok, maven2 installed, what now?
2210 2011-01-13 18:46:27 <nevezen> may not be one btc but small increments in less amount of time
2211 2011-01-13 18:46:44 <cosurgi> just run `mvn` in trunk root directory?
2212 2011-01-13 18:46:53 <Diablo-D3> cosurgi: mvn package
2213 2011-01-13 18:46:54 <sipa> 0.2% of daily produces BTC: 15 BTC/day
2214 2011-01-13 18:47:07 <cosurgi> ok. does something.
2215 2011-01-13 18:47:14 <cosurgi> Downloading lots of stuff.
2216 2011-01-13 18:47:15 <sipa> should be more than enough to pay his server bill
2217 2011-01-13 18:47:16 <Diablo-D3> cosurgi: the first time it downloads a ton of shit
2218 2011-01-13 18:47:23 <sipa> *if* actually many people do donate
2219 2011-01-13 18:47:26 <lucky> sipa, and increasing as the pool's sahre of generation increases ;p
2220 2011-01-13 18:47:31 <lucky> sipa, yeah, if.
2221 2011-01-13 18:47:40 <sipa> lucky: the network's speed also increases
2222 2011-01-13 18:47:46 <kiba`> I donate
2223 2011-01-13 18:47:58 <lucky> sipa kiba... a dyslexic would go crazy :o
2224 2011-01-13 18:48:08 <nevezen> where are the latest amount of blocks stored?
2225 2011-01-13 18:48:14 <sipa> ;;bc,stats
2226 2011-01-13 18:48:15 <ArtForz> yea, and long-term less than what you would've made on your own
2227 2011-01-13 18:48:17 <gribble> Current Blocks: 102500 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 315 blocks | Next Difficulty In About: 1 day, 20 hours, 32 minutes, and 15 seconds | Next Difficulty Estimate: 18240.24054734
2228 2011-01-13 18:48:19 <lucky> Current Bitcoin block#: 102500
2229 2011-01-13 18:48:28 <sipa> ArtForz: if you don't donate it's the same :)
2230 2011-01-13 18:48:32 <ArtForz> not really
2231 2011-01-13 18:48:34 <nevezen> I mean when you're downloading them..
2232 2011-01-13 18:48:37 <sipa> on average
2233 2011-01-13 18:48:53 <ArtForz> pool has higher latency
2234 2011-01-13 18:49:01 <sipa> slightly
2235 2011-01-13 18:49:02 <sipa> yes
2236 2011-01-13 18:49:03 samfisher has quit (Read error: Connection reset by peer)
2237 2011-01-13 18:49:08 <lucky> unless you have a *major* GPU investment, in the short-medium term (a time period of say, less than a year) it's a gamble to a large degree
2238 2011-01-13 18:49:16 <lucky> you might generate three blocks in 3 days, or spent 3 weeks.
2239 2011-01-13 18:49:24 <luke-jr> ArtForz: that's why I run N+1 threads âº
2240 2011-01-13 18:49:37 <lucky> over time it evens out, but yeah...
2241 2011-01-13 18:49:52 <kiba`> people wants bitcoin now, rather than later
2242 2011-01-13 18:49:56 <ArtForz> still doesnt change the fact that you'll be mining on stale data about getwork_interval/1200
2243 2011-01-13 18:49:57 <lucky> aye.
2244 2011-01-13 18:50:22 <lucky> i wonder what the economic ramifications of the amount coined dropping to 25 BTC will be
2245 2011-01-13 18:50:25 <kiba`> others are not as patient as you, ArtForz
2246 2011-01-13 18:50:27 <ArtForz> which is about 0.4% for 5s interval
2247 2011-01-13 18:50:53 <kiba`> beside, GPU mining will get unprofitable
2248 2011-01-13 18:50:54 <sipa> ArtForz: ok, but the same things holds when mining locally
2249 2011-01-13 18:50:56 <kiba`> sooner or later
2250 2011-01-13 18:50:58 <ArtForz> not to mention downtime when the pool server goes down/is unreachable/...
2251 2011-01-13 18:51:10 <sipa> of course, latency is definitely higher
2252 2011-01-13 18:51:17 <ArtForz> yeah, except locally I run a push-based "getwork"
2253 2011-01-13 18:51:34 <ArtForz> with worst case latency of ~300ms
2254 2011-01-13 18:51:38 <cosurgi> is that list up to date? http://pastebin.com/AvymGnMJ
2255 2011-01-13 18:51:39 <sipa> that's pretty nice
2256 2011-01-13 18:51:41 <cosurgi> speeds of GPUs
2257 2011-01-13 18:52:00 <cosurgi> Diablo-D3: [ERROR] BUILD FAILURE
2258 2011-01-13 18:52:03 <cosurgi> Unable to locate the Javac Compiler in: /usr/lib/jvm/java-6-openjdk/jre/../lib/tools.jar
2259 2011-01-13 18:52:08 <jgarzik> push makes far more sense than pull/poll, for getwork networks
2260 2011-01-13 18:52:09 <cosurgi> ok. I see
2261 2011-01-13 18:52:12 <cosurgi> Please ensure you are using JDK 1.4 or above and
2262 2011-01-13 18:52:12 <ArtForz> if my line goes down it doesnt matter if I mine locally or in a pool, I wont get any
2263 2011-01-13 18:52:31 <ArtForz> except with mining locally I dont have to worry about a server going down/unreachable/...
2264 2011-01-13 18:52:38 <sipa> yeah; i've been suggesting slush to implement a blocking getwork that waits until nez work is available
2265 2011-01-13 18:52:42 <Diablo-D3> cosurgi: you dont have a jdk.
2266 2011-01-13 18:52:50 <sipa> new
2267 2011-01-13 18:52:50 <Diablo-D3> cosurgi: install the matching jdk for your jre
2268 2011-01-13 18:53:05 <niekie> Grah.
2269 2011-01-13 18:53:13 <niekie> Link2VoIP blocked international calling on my account.
2270 2011-01-13 18:53:16 <ArtForz> *shrug*
2271 2011-01-13 18:53:25 <Diablo-D3> cosurgi: it'd be like apt-get install openjdk-6-jdk
2272 2011-01-13 18:53:25 <jgarzik> wait? no. keep working on current block until a new one appears.
2273 2011-01-13 18:53:41 <jgarzik> then 'push' gets you working on new work
2274 2011-01-13 18:53:53 <ArtForz> yep
2275 2011-01-13 18:54:10 <ArtForz> and you can emulate that if you have one thread getting notified about a new block arriving
2276 2011-01-13 18:54:16 <sipa> jgarzik: sure, the miner would ask for work from the server in one thread, but keep calculating wih his existing data
2277 2011-01-13 18:54:34 <sipa> and that ask blocks until new work is available
2278 2011-01-13 18:54:35 <ArtForz> basically if your "watch-for-next-block" getwork gets a response, kick all your miner threadsd to issue a getwork RIGHT NOW
2279 2011-01-13 18:54:35 <cosurgi> [INFO] BUILD SUCCESSFUL
2280 2011-01-13 18:54:37 <cosurgi> :)))
2281 2011-01-13 18:54:46 <Diablo-D3> cosurgi: now just run it like normal.
2282 2011-01-13 18:54:54 <cosurgi> umm.
2283 2011-01-13 18:54:59 <cosurgi> Exception in thread "main" java.lang.UnsatisfiedLinkError: /home/bitcoin/diablo-miner/trunk/DiabloMiner/target/libs/natives/linux/liblwjgl.so: /usr/lib/jvm/java-6-openjdk/jre/lib/i386/libjawt.so: symbol awt_Unlock, version SUNWprivate_1.1 not defined in file libmawt.so with link time reference
2284 2011-01-13 18:55:05 <jgarzik> sipa: no
2285 2011-01-13 18:55:09 <jgarzik> sipa: miner should never ask for blocks
2286 2011-01-13 18:55:11 <Diablo-D3> cosurgi: you forgot to export your DISPLAY variable.
2287 2011-01-13 18:55:16 <jgarzik> sipa: miner should sit around, waiting to receive blocks
2288 2011-01-13 18:55:32 <ArtForz> which is exactly what my miner does
2289 2011-01-13 18:55:35 <cosurgi> oh, right, sorry :) That was another screen
2290 2011-01-13 18:55:36 <sipa> jgarzik: s/ask blocks/call getwork()/
2291 2011-01-13 18:55:44 <jgarzik> well, where "waiting" is defined as "keep crunching current block" of course :)
2292 2011-01-13 18:55:56 <sipa> yes, i think we mean the same thing
2293 2011-01-13 18:56:02 <sipa> but with different names
2294 2011-01-13 18:56:07 <Diablo-D3> btw, mine falls through, I hope, with getwork fail
2295 2011-01-13 18:56:08 <cosurgi> yay! Now I have git trunk DiabloMiner running!
2296 2011-01-13 18:56:13 <Diablo-D3> cosurgi: you did anyhow.
2297 2011-01-13 18:56:18 <Diablo-D3> cosurgi: zip is always updated to match newest git.
2298 2011-01-13 18:56:36 <cosurgi> yeah. but 'git pull' is much more convenient :)
2299 2011-01-13 18:56:37 <jgarzik> sipa: no; ideal miner /never/ calls getwork. ideal miner creates TCP connection to bitcoind, and then waits for TCP traffic from bitcoind giving it new work.
2300 2011-01-13 18:56:41 <ArtForz> miner connects to node, node sends "start hashing this block header" and "update ntime", miner responds with nonces tried for last "work interval" and asynchronously sends a "found one with full block header" for H==0 nonces
2301 2011-01-13 18:56:49 <Diablo-D3> cosurgi: git pull && mvn package you mean
2302 2011-01-13 18:57:04 <cosurgi> yes, must remember that :)
2303 2011-01-13 18:57:06 <jgarzik> 'getwork' is fundamentally a polling (pull) solution. it will never be a push solution.
2304 2011-01-13 18:57:08 <sipa> jgarzik: sure, but you could implement that as a getwork that blocks (as in, no answer is sent) as long as no data is available
2305 2011-01-13 18:57:21 <ArtForz> yep
2306 2011-01-13 18:57:33 <sipa> i don't want to have the pool/client make tcp connections to the miner
2307 2011-01-13 18:57:41 <sipa> they should still go from miner to client
2308 2011-01-13 18:57:43 <ArtForz> huh?
2309 2011-01-13 18:57:53 <ArtForz> read the "miner connects to node" part again
2310 2011-01-13 18:57:59 <jgarzik> indeed
2311 2011-01-13 18:58:24 <ArtForz> node runs a TCP server, miner connect()s to it
2312 2011-01-13 18:58:51 <sipa> yes
2313 2011-01-13 18:58:59 <sipa> we're all saying the same thing :)
2314 2011-01-13 18:59:33 <ArtForz> if a miner doesnt respond with a "did x nonces since last update-something" within 3 seconds of getting sent a "work on new ntime/block header", it's presumed dead and disconnected
2315 2011-01-13 18:59:55 <sipa> what i'm trying to say is that you can wrap it all in the existing getwork() protocol
2316 2011-01-13 19:00:02 <luke-jr> jgarzik: hey, are you the low-level protocol layer rewrite guy?
2317 2011-01-13 19:00:10 <ArtForz> same way, if a miner doesnt get a "update ntime or "update block header" in 3 seconds, it thinks the node is dead, disconnects and tries reconnecting after 5-10 seconds
2318 2011-01-13 19:00:40 <ArtForz> yes, but piling kludges on top of each other doesnt sound like good protocol design
2319 2011-01-13 19:00:52 <sipa> true
2320 2011-01-13 19:01:51 <jgarzik> luke-jr: maybe but probably not
2321 2011-01-13 19:01:54 altamic has joined
2322 2011-01-13 19:01:59 <luke-jr> hrm
2323 2011-01-13 19:02:21 <jgarzik> luke-jr: I certainly have plenty of ideas about improving the protocol :)
2324 2011-01-13 19:02:22 <cosurgi> Diablo-D3: thanks a lot :)
2325 2011-01-13 19:02:25 <luke-jr> IMO, the new bitcoind should be reduced to *only* p2p maintenance stuff :P
2326 2011-01-13 19:02:39 <luke-jr> let clients use a library for creating transactions, etc
2327 2011-01-13 19:03:02 <jgarzik> a libbitcoin would be nice
2328 2011-01-13 19:03:43 <sipa> ArtForz: what kind of cases do you use for the 5970's ?
2329 2011-01-13 19:03:48 <luke-jr> this way, you could also run your bitcoind on a server you don't trust, and have your client connect to it for networking
2330 2011-01-13 19:04:23 <ArtForz> test fit of 1st prototype case: http://bayimg.com/KAAeaaAdp
2331 2011-01-13 19:04:48 <ArtForz> main stack of 5970s: http://bayimg.com/eABDfaAdd
2332 2011-01-13 19:04:54 <luke-jr> basically, a bitcoin peer with no interest in the network itself (no wallet, etc)
2333 2011-01-13 19:06:49 <jgarzik> luke-jr: with a proper libbitcoin, one should be able to implement a fat client (existing bitcoin) or a modular, segmented, well-separated client
2334 2011-01-13 19:07:10 * jgarzik likes the idea of experimenting with process boundaries
2335 2011-01-13 19:08:33 <luke-jr> jgarzik: that just makes a fat libbitcoin :P
2336 2011-01-13 19:08:46 <cosurgi> not exactly
2337 2011-01-13 19:08:54 <jgarzik> luke-jr: not necessarily
2338 2011-01-13 19:09:14 <jgarzik> luke-jr: you don't load 100% of glibc, just the bits you need
2339 2011-01-13 19:09:21 <luke-jr> oh?
2340 2011-01-13 19:09:35 <cosurgi> it would have to be split into different parts, the architecture of modules. Who cares that it's all in the same directory or library. The point is that it will be easier to replace just a single small module with something else
2341 2011-01-13 19:09:40 <jgarzik> luke-jr: presuming you are smart and put separate functionality in separate .o modules
2342 2011-01-13 19:09:41 <luke-jr> in any case, you definitely /install/ 100% of glibc
2343 2011-01-13 19:10:25 <luke-jr> jgarzik: .o boundaries vanish at link time afaik
2344 2011-01-13 19:10:31 <jgarzik> luke-jr: incorrect
2345 2011-01-13 19:11:22 <luke-jr> maybe system-dependent? :P
2346 2011-01-13 19:11:23 <jgarzik> luke-jr: the linker lays out a static file, but module boundaries are generally page aligned, and other details
2347 2011-01-13 19:11:53 <jgarzik> luke-jr: thus, separate .o modules implies separate pages in shared lib, which implies you only page in the necessary code
2348 2011-01-13 19:12:03 <luke-jr> libbitcoinp2p, libbitcointx, libbitcoinblock, libbitcoinscript, libbitcoinwork ?
2349 2011-01-13 19:12:30 <luke-jr> jgarzik: that assumes a smart OS, :p
2350 2011-01-13 19:12:34 <jgarzik> luke-jr: no that's just a pain for everyone. link against libbitcoin, and just use only the APIs you need. everything else works.
2351 2011-01-13 19:12:51 <jgarzik> luke-jr: that covers Linux and Windows and other modern Unices
2352 2011-01-13 19:13:13 TheAncientGoat has quit (Ping timeout: 276 seconds)
2353 2011-01-13 19:13:27 <luke-jr> how about a no-code libbitcoin that just depends on all of the above?
2354 2011-01-13 19:14:08 <jgarzik> luke-jr: that just slows down the executable load for everybody
2355 2011-01-13 19:14:14 <luke-jr> sigh
2356 2011-01-13 19:14:41 <ArtForz> you need most parts of that everywhere anyways
2357 2011-01-13 19:14:45 <jgarzik> luke-jr: your OS's shared lib loader loads 1 lib faster than N libs
2358 2011-01-13 19:14:50 <jgarzik> luke-jr: avoid pointless separation
2359 2011-01-13 19:15:06 <ArtForz> yep
2360 2011-01-13 19:15:07 <luke-jr> jgarzik: it's the UNIX way!
2361 2011-01-13 19:15:12 <ArtForz> no it's not
2362 2011-01-13 19:15:16 <luke-jr> one small module that does its job well
2363 2011-01-13 19:15:57 <ArtForz> yes, but job != part of a job
2364 2011-01-13 19:16:22 <luke-jr> ArtForz: there's no reason for the p2p daemon to use libbitcointx, libbitcoinblock, libbitcoinwork, etc
2365 2011-01-13 19:16:27 <jgarzik> you are also thinking about a _program_ not a lib. there will be plenty of elements shared across all bitcoin programs.
2366 2011-01-13 19:16:29 <luke-jr> nor for anything else to use libbitcoinp2p2
2367 2011-01-13 19:16:31 <luke-jr> p2p*
2368 2011-01-13 19:16:36 <ArtForz> so you want to pass on TX and blocks without checking if they're valid?
2369 2011-01-13 19:16:39 <jgarzik> you still need the core data structures
2370 2011-01-13 19:16:46 <luke-jr> ArtForz: sure, why not?
2371 2011-01-13 19:16:51 <jgarzik> otherwise it's just blind messaging, without every checking sums
2372 2011-01-13 19:17:24 <ArtForz> forget it, he's just trolling
2373 2011-01-13 19:17:36 <luke-jr> the whole "valid" nonsense is just a pain anyway
2374 2011-01-13 19:17:41 <jgarzik> ArtForz: I gave up at precisely same point :)
2375 2011-01-13 19:17:47 <luke-jr> let the block generators decide if it's valid or not
2376 2011-01-13 19:17:58 <kiba`> jgarzik: you do know that mtgox doesn't do paypal USD right?
2377 2011-01-13 19:18:08 <kiba`> your bitcoinwatch.com site is getting incresingly outdated
2378 2011-01-13 19:18:19 <jgarzik> kiba`: mtgox API still returns paypalUSD
2379 2011-01-13 19:18:30 <jgarzik> kiba`: I'll change the text though
2380 2011-01-13 19:18:32 <ArtForz> yep
2381 2011-01-13 19:18:41 <ArtForz> s/PayPal/LibertyReserve/ , there ;)
2382 2011-01-13 19:18:49 <jgarzik> yep
2383 2011-01-13 19:18:57 <kiba`> are you going to add bitcoin central?
2384 2011-01-13 19:19:09 <jgarzik> I've been hesitated adding the newer markets and currencies, because they have _very_ low volume
2385 2011-01-13 19:19:24 <jgarzik> I don't want to list a market that no one cares about
2386 2011-01-13 19:19:35 <jgarzik> if it's popular, I'll happily add it
2387 2011-01-13 19:19:55 <ArtForz> well, how is it supposed to get popular if no one knows about it? ;)
2388 2011-01-13 19:20:04 <jgarzik> agreed it is a catch-22
2389 2011-01-13 19:20:12 <ArtForz> I sold about 100BTC on BC so far
2390 2011-01-13 19:20:37 <kiba`> I thought BC is going to take off
2391 2011-01-13 19:20:40 <cosurgi> so make the list of all markets on a separate page, called "all markets, even those not used, yet"
2392 2011-01-13 19:20:47 <kiba`> but I guess BC also have a hard time competing with MtGox
2393 2011-01-13 19:20:47 <cosurgi> and catch-22 solved :)
2394 2011-01-13 19:20:58 <jgarzik> for instance, I constantly think about dropping btcex, because the RUB trade hits zero on many days. Ditto some of BCM's lesser used features.
2395 2011-01-13 19:21:00 molecular has quit (Ping timeout: 240 seconds)
2396 2011-01-13 19:21:07 <ArtForz> well, mtgox didnt do LREUR last time I checked
2397 2011-01-13 19:21:09 <jgarzik> MoneyBookers was intentionally excluded for similar reasons.
2398 2011-01-13 19:21:26 molecular has joined
2399 2011-01-13 19:21:29 <ArtForz> and BC doesnt have the 0.6% + 1% fees
2400 2011-01-13 19:21:33 <jgarzik> ArtForz: mtgox does some sort of euro deposits thing... I guess he credits you in USD for an EUR deposit?
2401 2011-01-13 19:21:33 <kiba`> it's kinda odd that the Russian have their own forum
2402 2011-01-13 19:21:44 <kiba`> but their exchange have very little volume
2403 2011-01-13 19:22:12 <ArtForz> jgarzik: iirc, yes
2404 2011-01-13 19:22:34 <jgarzik> maybe action has moved from RUB to WMR
2405 2011-01-13 19:23:13 <kiba`> webmoney rubes?
2406 2011-01-13 19:23:23 <jgarzik> cosurgi: I'm leaning against "make another page" for bitcoinwatch. The next update will link each market to bitcoincharts.com.
2407 2011-01-13 19:23:33 <jgarzik> bitcoinwatch aims to be a "single page status of bitcoin economy" site
2408 2011-01-13 19:23:43 <jgarzik> suggestions for improvement along those lines are welcome
2409 2011-01-13 19:23:51 <jgarzik> including links to external sites such as bitcoincharts.com
2410 2011-01-13 19:23:57 <kiba`> jgarzik: well
2411 2011-01-13 19:24:01 <kiba`> they do have some volume
2412 2011-01-13 19:24:17 <kiba`> 50.2 for RUB/ BTC
2413 2011-01-13 19:24:33 <kiba`> for YAD/BTC, it's 200
2414 2011-01-13 19:24:51 <kiba`> these are in the last 24 hours
2415 2011-01-13 19:24:55 mtgox has joined
2416 2011-01-13 19:25:55 <kiba`> people still don't care much about euro apperantly
2417 2011-01-13 19:26:32 <luke-jr> no markets allow euro?
2418 2011-01-13 19:26:50 <kiba`> bitcoin central does trade in euro
2419 2011-01-13 19:27:30 <luke-jr> LR is useless :p
2420 2011-01-13 19:27:42 <ArtForz> yeah, except they kinda dont mention how you're supposed to add/withdraw euros...
2421 2011-01-13 19:29:14 <kiba`> did ya mention the problem to davout?
2422 2011-01-13 19:29:25 <ArtForz> nope, not yet
2423 2011-01-13 19:31:52 <kiba`> all you ever do is minez
2424 2011-01-13 19:32:01 <ArtForz> ... pretty much
2425 2011-01-13 19:33:32 <jgarzik> ??
2426 2011-01-13 19:33:39 <jgarzik> I traded LREUR quite easily
2427 2011-01-13 19:33:44 <jgarzik> works the same as LRUSD on BC
2428 2011-01-13 19:33:45 <ArtForz> yes
2429 2011-01-13 19:33:52 <cosurgi> https://bitcoin-central.net/ ?
2430 2011-01-13 19:33:55 <jgarzik> yes
2431 2011-01-13 19:34:13 <cosurgi> does it support VISA transfers? how it works, I mean?
2432 2011-01-13 19:34:17 <ArtForz> notice they also have plain ole EUR in balance and order book
2433 2011-01-13 19:34:33 <jgarzik> cosurgi: just BTC<->LRUSD BTC<->LREUR right now, AFAIK
2434 2011-01-13 19:34:45 <jgarzik> ArtForz: yes, he does manual EUR wires, too
2435 2011-01-13 19:34:55 <ArtForz> well, he should maybe say so... somewhere
2436 2011-01-13 19:35:07 <jgarzik> ArtForz: said so on the forums under the BC threads repeatedly
2437 2011-01-13 19:35:11 <cosurgi> what is LREUR and how do I get it transferred to my reall bacnk account?
2438 2011-01-13 19:35:15 <luke-jr> erm
2439 2011-01-13 19:35:25 <luke-jr> I just bought 100 BTC at 0.40 PPUSD ea
2440 2011-01-13 19:35:29 <luke-jr> on BitCoin Market
2441 2011-01-13 19:35:29 <cosurgi> *real bank account
2442 2011-01-13 19:35:30 <jgarzik> cosurgi: long answer ;-)
2443 2011-01-13 19:35:31 <luke-jr> how do I pay?
2444 2011-01-13 19:36:06 <jgarzik> cosurgi: the key problem with LRUSD and LREUR is the responsibility is yours, to find USD->LRUSD or EUR->LREUR exchanger that works for you.
2445 2011-01-13 19:36:10 <cosurgi> how do I get paid? :) (though I didn't sell any yet :)
2446 2011-01-13 19:36:21 <cosurgi> ok.
2447 2011-01-13 19:36:31 <cosurgi> whare should I start looking?
2448 2011-01-13 19:36:39 <cosurgi> EUR
2449 2011-01-13 19:36:40 <jgarzik> cosurgi: Liberty Reserve's website, which links to exchangers
2450 2011-01-13 19:36:51 <ArtForz> LRs "list of exchangers" is a good place to start
2451 2011-01-13 19:36:53 <jgarzik> cosurgi: you need a Liberty Reserve account, which can balances in USD or EUR
2452 2011-01-13 19:37:11 <jgarzik> the wiki needs a Liberty Reserve FAQ
2453 2011-01-13 19:37:13 <ArtForz> ecardone has pretty low fees sometimes
2454 2011-01-13 19:37:15 <cosurgi> is it like a bank? not-a-bank like paypal, or what?
2455 2011-01-13 19:37:27 <ArtForz> LR is a private currency
2456 2011-01-13 19:37:37 <ArtForz> well, actually it used to be 3 private currencies
2457 2011-01-13 19:37:38 <jgarzik> cosurgi: LRUSD is a digital, privacy currency locked to value of USD. ditto for EUR.
2458 2011-01-13 19:37:39 <x6763> liberty reserve's website sucks at explaining what liberty reserve is
2459 2011-01-13 19:37:49 <ArtForz> LRUSD, LREUR and LRGAU
2460 2011-01-13 19:37:58 <ArtForz> but LRGAU wasd discontinued iirc
2461 2011-01-13 19:38:04 <sipa> ArtForz: you use a 1kW power supply for 2 5970's?
2462 2011-01-13 19:38:11 <ArtForz> yep
2463 2011-01-13 19:38:11 <jgarzik> ArtForz: it was discontinued. I wonder why?
2464 2011-01-13 19:38:22 <sipa> wouldn't 850W be enough?
2465 2011-01-13 19:38:30 <ArtForz> I guess so
2466 2011-01-13 19:38:31 <jgarzik> probably physical gold storage is a PITA, and costs extra money that USD/EUR does not.
2467 2011-01-13 19:39:38 <ArtForz> I designed with some headroom in case I want to overvolt
2468 2011-01-13 19:39:52 <ArtForz> a 5970 @ 1.1625V is ~400W
2469 2011-01-13 19:40:13 <kiba`> there are more people who have liberty reserve
2470 2011-01-13 19:40:20 <kiba`> than there are people who knows about bitcoin
2471 2011-01-13 19:40:26 <luke-jr> please help â¹
2472 2011-01-13 19:40:31 <ArtForz> errr... I'd guess so
2473 2011-01-13 19:40:33 <luke-jr> I don't want seller to say I didn't pay
2474 2011-01-13 19:40:38 <luke-jr> how do I payâ :/
2475 2011-01-13 19:40:41 <cosurgi> jgarzik: this seems complicated :)
2476 2011-01-13 19:40:48 <kiba`> luke-jr: what did you do?
2477 2011-01-13 19:40:55 <jgarzik> cosurgi: it is, unfortunately :(
2478 2011-01-13 19:40:57 <luke-jr> kiba`: bought 100 BTC on Bitcoin market
2479 2011-01-13 19:41:01 <ArtForz> so anyways, I've been mainly using ecardone for LR[USD/EUR]<->EUR
2480 2011-01-13 19:41:07 <luke-jr> kiba`: with PayPal USD
2481 2011-01-13 19:41:09 <kiba`> oh
2482 2011-01-13 19:41:14 <cosurgi> I want some way to get money. From people.
2483 2011-01-13 19:41:15 <luke-jr> but I don't see any way to pay it
2484 2011-01-13 19:41:19 <kiba`> I don't know how to use the bitcoin market
2485 2011-01-13 19:41:30 <ArtForz> if you can wait a bit, they sometimes have really low fees and a decent exchange rate
2486 2011-01-13 19:41:43 <jgarzik> cosurgi: the fundamental thing is that LRUSD, LREUR and bitcoins are __hard currency__ that cannot be charged back. Paypal and credit cards are soft currency, where your exchange can be reversed.
2487 2011-01-13 19:42:02 <jgarzik> cosurgi: therefore, it is slightly more difficult to get hard currency, if all you have is soft currency.
2488 2011-01-13 19:42:08 <ArtForz> for example last week they "charged" -0.4% for LREUR->EUR bank wire
2489 2011-01-13 19:42:15 <luke-jr> jgarzik: significantly*
2490 2011-01-13 19:42:34 <jgarzik> cosurgi: http://BitcoinBuy.com/ lists some BTC exchangers that might be local
2491 2011-01-13 19:43:02 <ArtForz> yes, as in, you get a 1004EUR wire for 1000LREUR
2492 2011-01-13 19:43:03 larsivi has joined
2493 2011-01-13 19:43:11 <cosurgi> Oh, I see slush here.
2494 2011-01-13 19:43:45 <ArtForz> their fees wildly fluctuate from day to day
2495 2011-01-13 19:44:26 <jgarzik> cosurgi: If you only want <= 20 BTC, http://coinpal.ndrix.com/ lets you purchase BTC with PayPal
2496 2011-01-13 19:45:06 <cosurgi> jgarzik: I want to sell BTC :)
2497 2011-01-13 19:45:24 <EvanR-work> then youre going to need to wire a lot of money to mtgox ;)
2498 2011-01-13 19:45:25 <jgarzik> cosurgi: oh! that's a whole lot easier.
2499 2011-01-13 19:45:44 <cosurgi> so, how do I do that?
2500 2011-01-13 19:45:45 <jgarzik> cosurgi: see what ArtForz just said.
2501 2011-01-13 19:45:57 <nevezen> do people trade game currency with bitcoins?
2502 2011-01-13 19:46:08 <EvanR-work> linden dollars, wow gold, poker stars
2503 2011-01-13 19:46:12 <slush> cosurgi: yes?
2504 2011-01-13 19:46:22 <nevezen> what about eve online?
2505 2011-01-13 19:46:25 <jgarzik> cosurgi: (1) go to mtgox or BC, sell your BTC at market value, then transfer LREUR or LRUSD to your liberty reserve. (2) exchange LRUSD/LREUR for "real" money with an exchanger such as ecardone
2506 2011-01-13 19:46:26 <cosurgi> umm.. which one? :), this "20:37 < ArtForz> so anyways, I've been mainly using ecardone for LR[USD/EUR]<->EUR"
2507 2011-01-13 19:46:29 <cosurgi> ?
2508 2011-01-13 19:46:34 <EvanR-work> nevezen: dunno, ask in #bitcoin-otc
2509 2011-01-13 19:47:08 <cosurgi> slush: are you buying BTC too? Or just selling them?
2510 2011-01-13 19:47:17 <ArtForz> nice thing about ecardone, they dont require proof of identity for "small" sums
2511 2011-01-13 19:47:39 mtgox has quit (Ping timeout: 240 seconds)
2512 2011-01-13 19:47:42 <cosurgi> jgarzik: so I need to make an account on liberty reserve. ok. People who buy, are able to send money to my LR account?
2513 2011-01-13 19:47:53 <ArtForz> I havent yet figured out what a non-small amount is, but it has to be > $5k
2514 2011-01-13 19:48:11 <jgarzik> cosurgi: when you withdraw from mtgox or bitcoin central, the market will send money to your LR account.
2515 2011-01-13 19:48:41 <jgarzik> cosurgi: then, you send money FROM your LR account TO the exchanger who will give you a euro wire transfer etc.
2516 2011-01-13 19:48:51 <ArtForz> yep
2517 2011-01-13 19:48:54 <slush> cosurgi: currently I'm waiting what happen on market :)
2518 2011-01-13 19:48:56 <cosurgi> does ecardone work in Europe?
2519 2011-01-13 19:49:00 <ArtForz> yes
2520 2011-01-13 19:49:06 <ArtForz> they work mainly in eurpe iirc
2521 2011-01-13 19:49:13 <cosurgi> good.
2522 2011-01-13 19:49:14 <jgarzik> cosurgi: if you are selling a large amount, mtgox or bitcoin central may be willing to simply wire you directly.
2523 2011-01-13 19:49:21 <slush> cosurgi: And offering only few bitcoins for CZK, for czech newcomers
2524 2011-01-13 19:49:27 <jgarzik> cosurgi: say, over 10000 BTC
2525 2011-01-13 19:49:36 RazielZ has quit ()
2526 2011-01-13 19:49:48 <ArtForz> another small note: LR charges min(1%, 2.99) for *receiving* LRanything
2527 2011-01-13 19:50:01 <cosurgi> very nice!
2528 2011-01-13 19:50:13 <ArtForz> so if you want to minimize LR fees, move money in larger chunks
2529 2011-01-13 19:50:24 <jgarzik> slush: is there a CZK currency exchange for BTC?
2530 2011-01-13 19:50:35 <ArtForz> mtgox additionall charges 1% for LR payout and 0.6% on trades
2531 2011-01-13 19:50:46 <ArtForz> afaict BC ... doesn't
2532 2011-01-13 19:50:53 <jgarzik> ArtForz: correct
2533 2011-01-13 19:51:24 <cosurgi> jgarzik: thanks for help, I'll be investigating that more.
2534 2011-01-13 19:51:30 <jgarzik> interesting. bitcoin central has EUR and LREUR separate. nice.
2535 2011-01-13 19:51:52 <jgarzik> I suggested to davout that he support trading between non-BTC currencies, such as PGAU/LREUR
2536 2011-01-13 19:52:09 <slush> jgarzik: I don't think so. There are only few bitcoiners in CZ
2537 2011-01-13 19:52:50 <ArtForz> yeah, that'd be kinda cool
2538 2011-01-13 19:52:51 <cosurgi> why somebody has put 400BTC for 100 EUR each on bitcoin central :-o
2539 2011-01-13 19:53:54 <ArtForz> black swan bet? typo trolling?
2540 2011-01-13 19:54:06 <jgarzik> If davout adds that feature, his site will explode in popularity IMO
2541 2011-01-13 19:54:15 <cosurgi> holy shit, suddenly when watching those tables I felt like plying elite2, but this time for real :)
2542 2011-01-13 19:54:29 <ArtForz> just in case soemone enters a buy for 245/BTC instead of .245 :P
2543 2011-01-13 19:54:34 * jgarzik has always wanted a single website for trading all the "hard, digital" currencies. Programmatically, it is simple.
2544 2011-01-13 19:54:52 <cosurgi> leggaly it is not, I guess :)
2545 2011-01-13 19:55:13 <ArtForz> probably
2546 2011-01-13 19:55:32 <jgarzik> if you can trade BTC/LREUR, you can trade PGAU, GDP-{USD,EUR,GBP}, c-gold.
2547 2011-01-13 19:55:37 <jgarzik> for all currency pairs
2548 2011-01-13 19:55:50 <ArtForz> guess you'd have to incorporate in assfuckistan to get around currency exchange laws
2549 2011-01-13 19:56:14 <tcatm> Any idea how I could ask the RPC whether there are new transactions since last RPC?
2550 2011-01-13 19:56:22 <ArtForz> there are exchangers doing any<->any for most digital currencies
2551 2011-01-13 19:56:26 <jgarzik> ArtForz: presumably that applies with 1 currency pair too
2552 2011-01-13 19:56:32 <cosurgi> there's a small island that declared independency years ago, just forgot its name.
2553 2011-01-13 19:56:36 <jgarzik> ArtForz: automated-deposit exchangers?
2554 2011-01-13 19:56:49 <jgarzik> most exchangers I find are slow and human-based
2555 2011-01-13 19:56:54 <jgarzik> and $$$ in fees
2556 2011-01-13 19:56:57 <x6763> has anyone here bought bitcoins from Bitcoin Morpheous?
2557 2011-01-13 19:57:03 <ArtForz> fees are horrible, but there's https://ecardone.com/exchange/exchange-liberty-reserve-to.php
2558 2011-01-13 19:57:09 <cosurgi> ok guys, nice talking. I need to do some stuff for tomorrow ;)
2559 2011-01-13 19:57:19 <jgarzik> x6763: I've done MTG-USD trade with him
2560 2011-01-13 19:57:49 AAA_awright has quit (Ping timeout: 272 seconds)
2561 2011-01-13 19:57:53 <ArtForz> they do at least 100% automated LR$<->GDP and LR$<->Pecunix
2562 2011-01-13 19:57:56 <x6763> jgarzik: no problems i'm assuming?
2563 2011-01-13 19:58:16 <jgarzik> x6763: nope
2564 2011-01-13 19:58:52 AAA_awright has joined
2565 2011-01-13 19:59:25 <x6763> jgarzik: good, thanks
2566 2011-01-13 19:59:28 <kiba`> This thinking leads to centralization of power and capital, which are exactly what frameworks like bitcoin are designed to dismantle.
2567 2011-01-13 19:59:33 <kiba`> what nosense!
2568 2011-01-13 19:59:37 <x6763> anyone else here done any trades or bought bitcoins through Bitcoin Morpheous?
2569 2011-01-13 19:59:38 <ArtForz> what?
2570 2011-01-13 19:59:38 <kiba`> s/nosense/nonsense
2571 2011-01-13 19:59:56 <ArtForz> kiba: who are you replying to?
2572 2011-01-13 20:00:20 <kiba`> a guy named gene
2573 2011-01-13 20:00:22 <kiba`> he must be new
2574 2011-01-13 20:00:41 <kiba`> registered December 14, 2010
2575 2011-01-13 20:01:10 <cosurgi> kiba`: he's not here.
2576 2011-01-13 20:01:27 <kiba`> the forum, cosurgi
2577 2011-01-13 20:01:42 <kiba`> http://www.bitcoin.org/smf/index.php?topic=2763.msg37739#msg37739
2578 2011-01-13 20:01:50 <kiba`> this is my real reply
2579 2011-01-13 20:01:55 <kiba`> to the nonsense he sprout
2580 2011-01-13 20:03:15 andrew12 has joined
2581 2011-01-13 20:03:50 <andrew12> anyone got a link to the block chain? (like a zip)
2582 2011-01-13 20:03:59 * andrew12 is at school again. =p
2583 2011-01-13 20:07:39 <kiba`> my school is closed yet again tommorow
2584 2011-01-13 20:09:22 <nevezen> snow/
2585 2011-01-13 20:09:31 <kiba`> hai
2586 2011-01-13 20:13:18 <ArtForz> oi!
2587 2011-01-13 20:13:31 <ArtForz> 1k volume @ 0.4
2588 2011-01-13 20:15:13 <kiba`> everything is going up!
2589 2011-01-13 20:15:32 <LobsterMan> http://i.imgur.com/ADgkC.jpg
2590 2011-01-13 20:16:02 <luke-jr> ArtForz: really?
2591 2011-01-13 20:16:05 <luke-jr> did I do that?
2592 2011-01-13 20:16:37 <ArtForz> unless you bought about 1kBTC on mtgox in the last hour... no
2593 2011-01-13 20:17:18 <EvanR-work> lol
2594 2011-01-13 20:17:46 <LobsterMan> it could be people trading with themselves
2595 2011-01-13 20:17:56 <LobsterMan> using multiple accounts to try to make it look like the price is higher than it should be
2596 2011-01-13 20:18:02 <ArtForz> I'm pretty sure they're not
2597 2011-01-13 20:18:03 <LobsterMan> either way, i'm not complaining :]
2598 2011-01-13 20:18:32 <luke-jr> ArtForz: 100 @ 0.40 on BTC Market
2599 2011-01-13 20:18:38 <ArtForz> yep
2600 2011-01-13 20:18:49 <ArtForz> but that was PPUSD
2601 2011-01-13 20:18:56 <luke-jr> so?
2602 2011-01-13 20:19:11 <kiba`> how come paypal dollars and LRUSD match up evenly?
2603 2011-01-13 20:19:12 <ArtForz> PPUSD is usually always a bit higher than LRUSD
2604 2011-01-13 20:19:46 <luke-jr> ~/.bitcoin is what I need to backup right?
2605 2011-01-13 20:19:57 <sipa> just wallet.dat
2606 2011-01-13 20:20:02 <ArtForz> really just ~/.bitcoin/wallet.dat
2607 2011-01-13 20:20:16 <ArtForz> or whatever filename you specify in a backupwallet RPC call
2608 2011-01-13 20:20:51 lyspooner has joined
2609 2011-01-13 20:21:07 <nevezen> what are the rest of the files in there for?
2610 2011-01-13 20:21:13 <ArtForz> various stuff
2611 2011-01-13 20:21:21 <kiba`> blocks data
2612 2011-01-13 20:21:34 <ArtForz> addr.dat contains node addresses, blkXXXX.dat is the block data, blkindex is tx and block index
2613 2011-01-13 20:21:56 <necrodearia> http://nullvoid.org/bitcoin/ seems to be functional again as it was in the past. I am not sure if it is still useful, but as of a few hours ago, a few have asked about the site. Let me know if it seems inaccurate again.
2614 2011-01-13 20:22:15 <nevezen> does it still bootstrap using irc?
2615 2011-01-13 20:22:26 <ArtForz> yep
2616 2011-01-13 20:22:49 <ArtForz> first tries IRC, iirc fallback to hardcoded node list after 60 sec
2617 2011-01-13 20:23:11 <ArtForz> and you can always manually -addnode a known node IP
2618 2011-01-13 20:23:17 <luke-jr> how secure are GPG private keys? eg, if I put my key on my server, will it be safe provided a strong passphrase?
2619 2011-01-13 20:23:40 <cosurgi> hmm.. I dream about some king of BTC bank, that issues non-credit cards, that I can use commonplace to buy stuff. I would send BTCs to my account in that bank, and then go shopping to pay for food.
2620 2011-01-13 20:23:47 <sipa> if the passphrase is safe, GPG is rock solid
2621 2011-01-13 20:23:48 <nevezen> does it use DHT?
2622 2011-01-13 20:23:57 <ArtForz> no
2623 2011-01-13 20:24:15 <ArtForz> it still wouldnt fix the bootstrap problem
2624 2011-01-13 20:24:43 <nevezen> so it's not very decentralized..
2625 2011-01-13 20:24:49 <ArtForz> huh?
2626 2011-01-13 20:25:00 <cosurgi> ArtForz: you mean - this kind of bank wouldn't boostrap?
2627 2011-01-13 20:25:03 <sipa> bootstrapping is not decentralized, no, you need some point to start
2628 2011-01-13 20:25:12 <nevezen> right
2629 2011-01-13 20:25:12 <luke-jr> GPG have an option to say "never publish this key, public or private"?
2630 2011-01-13 20:25:13 <ArtForz> doesnt make it centralized though
2631 2011-01-13 20:25:19 <sipa> luke-jr: ?
2632 2011-01-13 20:25:30 <luke-jr> sipa: I'm making a key solely for backups
2633 2011-01-13 20:25:40 <luke-jr> don't want it to get to the outside world in any form really
2634 2011-01-13 20:25:41 <nevezen> not saying it is artforz.. but it seems like a point of failure
2635 2011-01-13 20:25:42 <ArtForz> you just have to get a list of public nodes from somewhere
2636 2011-01-13 20:25:52 <sipa> luke-jr: just don't send it anywhere?
2637 2011-01-13 20:26:02 <luke-jr> sipa: won't GPG automatically upload it sometimes?
2638 2011-01-13 20:26:05 <sipa> no
2639 2011-01-13 20:26:10 <luke-jr> O.o
2640 2011-01-13 20:26:16 <sipa> there is gpg --send-keys command
2641 2011-01-13 20:26:21 <EvanR-work> globally centralized privacy guard!
2642 2011-01-13 20:26:22 <luke-jr> it's taking a long time to generate the key XD
2643 2011-01-13 20:26:39 <EvanR-work> luke-jr: look at porn for a while, you probably ran out of entropy
2644 2011-01-13 20:26:45 <ArtForz> I can take a fresh install, run -noirc and -addnode a node accepting incoming connections and it bootstraps just fine
2645 2011-01-13 20:26:49 <EvanR-work> network traffic adds entropy
2646 2011-01-13 20:27:13 <luke-jr> EvanR-work: never
2647 2011-01-13 20:29:40 <ArtForz> I don't see how that's centralized
2648 2011-01-13 20:30:05 <EvanR-work> if gpg automatically uploads your stuff... where does it upload to
2649 2011-01-13 20:30:19 <luke-jr> EvanR-work: the keyserver?
2650 2011-01-13 20:30:21 <nevezen> it needs a place to start automatically; bootstrap
2651 2011-01-13 20:30:30 <EvanR-work> the keyserver?
2652 2011-01-13 20:30:35 <nevezen> but if all the hard coded nodes are unconnectable..
2653 2011-01-13 20:30:58 <luke-jr> # GnuPG can send and receive keys to and from a keyserver. These
2654 2011-01-13 20:31:04 <luke-jr> ~/.gnupg/gpg.conf
2655 2011-01-13 20:31:16 <EvanR-work> ok so more than one
2656 2011-01-13 20:31:28 <luke-jr> I think the default is just one
2657 2011-01-13 20:31:41 <ArtForz> ... then someone has to replace the hardcoded list
2658 2011-01-13 20:33:01 <luke-jr> w00t, now I have 2,7î§04 TBC
2659 2011-01-13 20:33:08 <Diablo-D3> what was that?
2660 2011-01-13 20:33:11 <Diablo-D3> I couldnt quite hear you
2661 2011-01-13 20:33:32 <EvanR-work> luke-jr: xterm and unifont dont show î§, is there a better symbol for this
2662 2011-01-13 20:33:44 <ArtForz> yum, unicode soup!
2663 2011-01-13 20:34:13 <luke-jr> de-bong, ra-mill, po-mill, go
2664 2011-01-13 20:34:23 <sipa> luke-jr: you'd still need to choose which keyserver it would send to
2665 2011-01-13 20:34:26 <luke-jr> EvanR-work: not really.
2666 2011-01-13 20:34:28 <sipa> in the config file
2667 2011-01-13 20:34:56 <luke-jr> EvanR-work: it's equivalent to hexadecimal 'e', but looks like a upside-down, mirrored 2
2668 2011-01-13 20:35:21 <EvanR-work> why not use e
2669 2011-01-13 20:35:28 <luke-jr> http://luke.dashjr.org/education/tonal/glyphs/tonal-table.png
2670 2011-01-13 20:35:36 <necrodearia> As the value of 0.01 bitcoin nears us$0.01 will there be any updates to allow minimum amount of 0.001 to be sent/received?
2671 2011-01-13 20:35:39 <luke-jr> because 'e' is hexadecimal, not tonal
2672 2011-01-13 20:35:53 <EvanR-work> eh
2673 2011-01-13 20:35:57 <luke-jr> necrodearia: most of the network would need to upgrade
2674 2011-01-13 20:36:02 <ArtForz> because using standard notation is boring
2675 2011-01-13 20:36:16 <luke-jr> ArtForz: that *is* the (tonal) standard
2676 2011-01-13 20:36:18 <cosurgi> ArtForz: what power supply would you recommend for dual ATI 5850 (151 W) + cheapest sempron ?
2677 2011-01-13 20:36:27 <cosurgi> 500 W ?
2678 2011-01-13 20:36:41 <ArtForz> I'd probably go for 550 or 600W
2679 2011-01-13 20:36:46 <cosurgi> ok.
2680 2011-01-13 20:36:49 <ArtForz> most PSUs seem to be most efficient around 50% load
2681 2011-01-13 20:37:09 <luke-jr> EvanR-work: my preferred fixed-width font is http://luke.dashjr.org/education/tonal/glyphs/fonts/Console/Luke.ConsoleMedium-16.pcf
2682 2011-01-13 20:37:25 <luke-jr> which I added a bunch of missing Unicode too (including smileys!)
2683 2011-01-13 20:37:26 <niekie> luke-jr/EvanR-work: can I quote that entropy/porn quote? (names omitted if you desire :))
2684 2011-01-13 20:37:33 <ArtForz> and in my experience it's easier to reuse/resell slightly bigger PSUs
2685 2011-01-13 20:37:37 <midnightmagic> i just wanted to apologise for pushing up the difficulty in testnet. i'd calculated wrongly about how long the difficulty would remain high after i left. :-(
2686 2011-01-13 20:37:48 <luke-jr> http://luke.dashjr.org/education/tonal/glyphs/fonts/Luxi/LuxiTonal.ttf contains only the Tonal glyphs sized for Luxi Mono
2687 2011-01-13 20:37:49 <EvanR-work> niekie: doesnt matter what we think, this channel is publically logged
2688 2011-01-13 20:38:03 <niekie> Oh right, it is.
2689 2011-01-13 20:38:10 <niekie> Oh well, it's still nice to ask :)
2690 2011-01-13 20:38:13 <EvanR-work> heh
2691 2011-01-13 20:38:48 <EvanR-work> midnightmagic: you may have inadvertantly created a market for them ;)
2692 2011-01-13 20:39:07 <midnightmagic> i was just screwing around trying to understand the invalid/stale errors.
2693 2011-01-13 20:39:11 <luke-jr> lol
2694 2011-01-13 20:39:34 <midnightmagic> EvanR-work: don't trust it, I had well over 50% of the cpu power and was dominating testnet for about a day. i could've done anything..
2695 2011-01-13 20:39:38 <luke-jr> speaking of which, would it work to just create a new network?
2696 2011-01-13 20:39:42 <luke-jr> would existing peers cooperate?
2697 2011-01-13 20:39:48 <EvanR-work> midnightmagic: heh
2698 2011-01-13 20:39:49 <ArtForz> difficulty 11.6 isn't exactly high
2699 2011-01-13 20:39:56 <midnightmagic> ? mine reads 23.
2700 2011-01-13 20:40:24 <ArtForz> BBE says 11.67
2701 2011-01-13 20:40:45 <ArtForz> iirc testnet uses a different value for 1.0 difficulty
2702 2011-01-13 20:41:00 <ArtForz> BBE shows normal-net equivalent difficulty
2703 2011-01-13 20:41:24 <midnightmagic> [bc@khan ~]$ bitcoind getinfo {"version" : 31902, "balance" : lots.00000000, "blocks" : 25497, "difficulty" : 23.35656123, "testnet" : true, "keypoololdest" : 1294845564, "paytxfee" : 0.00000000,}
2704 2011-01-13 20:41:50 <midnightmagic> ah i get it.
2705 2011-01-13 20:42:16 <ArtForz> testnet 1.0 diff = main net 0.5 diff
2706 2011-01-13 20:42:33 <midnightmagic> wise.
2707 2011-01-13 20:42:47 <ArtForz> lets see...
2708 2011-01-13 20:42:47 <midnightmagic> speaking of wisdom, where is satoshi anyway?
2709 2011-01-13 20:42:53 <ArtForz> ;;bc,calc 10000 11.6
2710 2011-01-13 20:42:54 <gribble> Error: invalid syntax (<string>, line 1)
2711 2011-01-13 20:43:02 <ArtForz> whoops, what was the syntax again
2712 2011-01-13 20:43:23 <ArtForz> right
2713 2011-01-13 20:43:24 <ArtForz> ;;bc,calcd 10000 11.6
2714 2011-01-13 20:43:24 <gribble> The average time to generate a block at 10000 Khps, given the supplied difficulty of 11.6, is 1 hour, 23 minutes, and 2 seconds
2715 2011-01-13 20:44:04 <midnightmagic> yeah, so the typical low-end video card can't meaningfully pull back a block within a few minutes like they could before.
2716 2011-01-13 20:44:21 <ArtForz> ;;bc,calcd 600000 11.6
2717 2011-01-13 20:44:21 <gribble> The average time to generate a block at 600000 Khps, given the supplied difficulty of 11.6, is 1 minute and 23 seconds
2718 2011-01-13 20:44:29 <midnightmagic> and cpu test people are screwed.
2719 2011-01-13 20:44:32 <ArtForz> thats a single 5970 ...
2720 2011-01-13 20:44:39 <EvanR-work> midnightmagic: hes hanging out with emmanuel goldstein ;)
2721 2011-01-13 20:45:01 <midnightmagic> :-)
2722 2011-01-13 20:45:27 <nevezen> so an nvidia ion puts out about the same as my dual core processor
2723 2011-01-13 20:45:32 <midnightmagic> so i basically destroyed immediate-gratification testnet temporarily, and for that i apologise.
2724 2011-01-13 20:45:36 <nevezen> ~2MH/s
2725 2011-01-13 20:45:49 <ArtForz> I think we alread yhad higher testnet difficulties
2726 2011-01-13 20:46:04 <midnightmagic> it was sitting at 4.x (so, I guess 2.x) before i got hold of it.
2727 2011-01-13 20:46:27 <ArtForz> check ~ block 16500
2728 2011-01-13 20:46:45 <ArtForz> humm, no, only 5.3
2729 2011-01-13 20:47:13 <midnightmagic> i started before then
2730 2011-01-13 20:47:20 <midnightmagic> no i'm a stiinking liar.
2731 2011-01-13 20:47:21 <ArtForz> I doubt it
2732 2011-01-13 20:47:28 <x6763> luke-jr: as far as i know, the only difference between the main network and the test network is the magic bytes at the beginning of the messages in the bitcoin protocol...creating a new network would mean that the clients would have to recognize a new set of magic bytes, which would require a change in the bitcoin software, meaning that existing clients would not support a new network
2733 2011-01-13 20:47:52 <ArtForz> because that was me testing a few miner tweaks with a bunch of 5770s :P
2734 2011-01-13 20:47:54 <luke-jr> x6763: so the clients all check to see if they recognize the magic bytes? -.-
2735 2011-01-13 20:48:03 <luke-jr> why is this p2p network so draconian? XD
2736 2011-01-13 20:48:13 <EvanR-work> its democracy
2737 2011-01-13 20:48:16 <midnightmagic> ArtForz: :) well it was 4.x when I started in with my little desktop 9600gt to see what it would do. honest. :-)
2738 2011-01-13 20:48:17 <luke-jr> just pass everything as long as it isn't fraud
2739 2011-01-13 20:48:30 <luke-jr> EvanR-work: democracy would mean it was configurable
2740 2011-01-13 20:48:33 <EvanR-work> it is
2741 2011-01-13 20:48:34 <ArtForz> yea
2742 2011-01-13 20:48:42 <luke-jr> modifying the source isn't configuring
2743 2011-01-13 20:48:42 <EvanR-work> open source
2744 2011-01-13 20:48:45 <EvanR-work> yes it is
2745 2011-01-13 20:48:48 <luke-jr> no
2746 2011-01-13 20:48:56 <EvanR-work> distribute your own client, the exact opposite of draconian
2747 2011-01-13 20:49:01 <x6763> luke-jr: yes, they check those bytes in order to differentiate between the main network and the test network
2748 2011-01-13 20:49:09 <luke-jr> also, the source is ugly spaghetti
2749 2011-01-13 20:49:09 <EvanR-work> you can even add a configuration gui to ease your angst
2750 2011-01-13 20:49:12 <midnightmagic> see, here: http://blockexplorer.com/testnet/b/21730
2751 2011-01-13 20:49:17 <ArtForz> yea
2752 2011-01-13 20:49:27 <ArtForz> *shrug*
2753 2011-01-13 20:49:56 <midnightmagic> anyway, just saying sorry. kinda left the cards in testnet longer, by accident.
2754 2011-01-13 20:50:09 <midnightmagic> fell asleep. etc.
2755 2011-01-13 20:50:09 <ArtForz> really, nobody cares
2756 2011-01-13 20:50:20 <midnightmagic> people were scolding me.
2757 2011-01-13 20:50:30 <midnightmagic> so someone does.
2758 2011-01-13 20:50:43 <ArtForz> I should throw my old 4870 on there just for fun
2759 2011-01-13 20:51:12 <ArtForz> ;;bc,calcd 98000 11.6
2760 2011-01-13 20:51:12 <gribble> The average time to generate a block at 98000 Khps, given the supplied difficulty of 11.6, is 8 minutes and 28 seconds
2761 2011-01-13 20:52:08 <x6763> luke-jr: unless there are other special things the official client does for the test network, i would think it would be relatively simple to modify the client to use other sets of magic bytes to create new networks, or even make it configurable
2762 2011-01-13 20:52:27 <EvanR-work> luke-jr: disassembly output of a microsoft binary is more ugly
2763 2011-01-13 20:52:50 <ArtForz> the client isnt structured too well, but I've seen FAR worse code
2764 2011-01-13 20:53:05 <x6763> luke-jr: but it's a pain trying to understand most anything in the official client code
2765 2011-01-13 20:53:37 <ArtForz> dunno, I understand the interesting parts just fine
2766 2011-01-13 20:53:55 <ArtForz> what I hate most is java ravioli code
2767 2011-01-13 20:54:05 <EvanR-work> lol
2768 2011-01-13 20:54:31 <ArtForz> = dig through X layers of abstraction just to find one routine that actually *does* something
2769 2011-01-13 20:54:34 <x6763> most of what i've learned about the bitcoin protocol has come from the forums, irc, and packet sniffing
2770 2011-01-13 20:54:45 <x6763> and the protocol page in the wikis
2771 2011-01-13 20:55:06 <ArtForz> well, the network protocol really is pretty simple
2772 2011-01-13 20:55:32 <ArtForz> the script sig engine... not so much
2773 2011-01-13 20:56:06 <EvanR-work> yeah whats that crap all about
2774 2011-01-13 20:56:15 charlesnw has joined
2775 2011-01-13 20:56:23 <ArtForz> it's a simple bytecode based stack machine
2776 2011-01-13 20:56:35 <EvanR-work> isnt the idea 'i know, lets add a scripting language into it!' automatically wrong
2777 2011-01-13 20:56:35 <x6763> the script engine isn't as bad as i was expecting it to be, but there's about a dozen different opcodes that i don't understand yet (conditionals, crypto, and pseudo-words)
2778 2011-01-13 20:56:44 <ArtForz> it would've been useful before isStandard()
2779 2011-01-13 20:57:27 <x6763> i just had to spend a couple minutes on the Forth wikipedia page to understand most of what the script engine was doing
2780 2011-01-13 20:57:45 TD_ has joined
2781 2011-01-13 20:57:58 <ArtForz> yeah, it's really not that bad
2782 2011-01-13 20:58:06 <ArtForz> just a bit unintuitive
2783 2011-01-13 20:58:44 <ArtForz> though the script mangling done for sig creation/validation is a bit of a mindfuck
2784 2011-01-13 20:59:41 <x6763> i have no idea what this means for OP_IF: <expression> if [statements] [else [statements]] endif
2785 2011-01-13 21:00:35 <ArtForz> errr... it's a if
2786 2011-01-13 21:00:35 TD has quit (Ping timeout: 276 seconds)
2787 2011-01-13 21:00:36 TD_ is now known as TD
2788 2011-01-13 21:00:54 <x6763> yeah, it's if, but where are these statements supposed to be that i'm supposed to execute?
2789 2011-01-13 21:01:22 <x6763> how big are the statements, etc?
2790 2011-01-13 21:01:40 <ArtForz> between OP_IF and OP_ENDIF
2791 2011-01-13 21:01:50 <ArtForz> or OP_IF / OP_ELSE / OP_ENDIF
2792 2011-01-13 21:02:33 <ArtForz> OP_IF = pop value, if it's TRUE execute the block up to matching OP_ELSE or OP_ENDIF
2793 2011-01-13 21:02:53 <ArtForz> if it's false, skip that block and execute from matching OP_ELSE to OP_ENDIF
2794 2011-01-13 21:03:14 <ArtForz> if false and theres no matching OP_ELSE, just skip to OP_ENDIF
2795 2011-01-13 21:04:04 <x6763> alright, makes sense, thanks
2796 2011-01-13 21:04:45 <ArtForz> iirc it uses recursion to handle these
2797 2011-01-13 21:04:54 charlesnw has quit (Quit: ZNC - http://znc.sourceforge.net)
2798 2011-01-13 21:05:25 charlesnw has joined
2799 2011-01-13 21:05:52 <ArtForz> well, or recursion mangled into a list + loop
2800 2011-01-13 21:08:15 Zarutian has joined
2801 2011-01-13 21:10:51 <x6763> OP_CHECKSIG, etc, is a little more confusing...it sounds like it's dealing with other parts of the transaction, too, and not just the script...is that right?
2802 2011-01-13 21:10:58 <ArtForz> yep
2803 2011-01-13 21:12:05 altamic has quit (Ping timeout: 255 seconds)
2804 2011-01-13 21:13:56 <x6763> hmmm...the hash would be something like this? hash(outputs + inputs + script[separator:])
2805 2011-01-13 21:14:47 <ArtForz> sounda bout right, it's really a bit convoluted
2806 2011-01-13 21:17:50 <ArtForz> it creates a new tx containing a subset of inputs/outputs and is hashing that
2807 2011-01-13 21:18:25 <ArtForz> aka SignatureHash in script.cpp
2808 2011-01-13 21:19:35 <x6763> alright, i'll have to see if i can make some sense out of that code...i also just remembered that i still need to get the ECC stuff figured out, too
2809 2011-01-13 21:20:58 <luke-jr> [15:49:07] <x6763> luke-jr: but it's a pain trying to understand most anything in the official client code
2810 2011-01-13 21:21:10 <luke-jr> exactly proof that open source != configurable, even for programmers :P
2811 2011-01-13 21:21:17 <x6763> lol
2812 2011-01-13 21:21:47 <luke-jr> EvanR-work: the scripting stuff looks useful potentially
2813 2011-01-13 21:21:58 <kiba`> some people's existence is incomperhenable to me
2814 2011-01-13 21:22:53 <ArtForz> well, seeing how many people partially reimplemented the official client, it can't be THAT hard to understand
2815 2011-01-13 21:23:31 <x6763> satoshi's code seems to be very unreadable to me, but that may just be because i'm not a c++ programmer, or haven't actually programmed professionally or gotten a degree or anything
2816 2011-01-13 21:24:11 <x6763> i've gotten a lot of the client reimplemented, but probably less than 1% is from looking at satoshi's code
2817 2011-01-13 21:24:50 <ArtForz> I'm pretty sure most of it is
2818 2011-01-13 21:25:13 <kiba`> Wikipedia clearly say cypherpunks have libertarian ideal
2819 2011-01-13 21:25:49 <ArtForz> probably most of it indirectly
2820 2011-01-13 21:25:50 <kiba`> but gene said they were anarchists with strong socialist undercurrent
2821 2011-01-13 21:26:02 <ArtForz> the bitcoin network spec didnt create itself out of thin air
2822 2011-01-13 21:26:04 <x6763> ArtForz: yes, indirectly
2823 2011-01-13 21:26:11 <kiba`> At least in the US, the term has been reclaimed by far right-wing elements and is associated with the myth of the rugged individual that lives on the frontier and can survive all alone.
2824 2011-01-13 21:26:14 <kiba`> that what he said
2825 2011-01-13 21:26:16 <kiba`> WTF
2826 2011-01-13 21:26:26 <kiba`> his map of reality is clearly skewed
2827 2011-01-13 21:27:27 <ArtForz> kiba: YHBT, HAND?
2828 2011-01-13 21:27:40 <kiba`> ?
2829 2011-01-13 21:27:55 <ArtForz> You Have Been Trolled, Have A Nice Day
2830 2011-01-13 21:29:16 <kiba`> the difference between trolling and actually believing in it is very thin, ArtForz
2831 2011-01-13 21:29:23 <ArtForz> yep
2832 2011-01-13 21:29:34 <ArtForz> outcome is really the same though
2833 2011-01-13 21:30:48 <kiba`> ya know
2834 2011-01-13 21:31:02 <kiba`> I suspect most opponents of libertarianism don't really understand what libertarianism stand for
2835 2011-01-13 21:31:06 <ArtForz> but yeah, libertarianism is becoming a kind of a problematic word
2836 2011-01-13 21:32:13 <ArtForz> it's about as bad as hacker or bricked nowadays
2837 2011-01-13 21:32:29 <ArtForz> hint: if you can un-brick something, it's not bricked
2838 2011-01-13 21:32:55 <kiba`> the word hacker is kinda getting mainstream
2839 2011-01-13 21:33:09 <ArtForz> yeah, mainly thanks to HAD and friends
2840 2011-01-13 21:39:04 andrew12 has quit (Quit: Page closed)
2841 2011-01-13 21:40:38 <midnightmagic> i love reading satoshi's code. i find it beautiful.
2842 2011-01-13 21:44:28 lyspooner has quit (Quit: ChatZilla 0.9.86 [Firefox 3.6.6/20100625231939])
2843 2011-01-13 21:47:22 <kiba`> and then BAM!
2844 2011-01-13 21:47:32 <kiba`> we droped to .3176 USD
2845 2011-01-13 21:47:55 <midnightmagic> price climbs too high, people get nervous.
2846 2011-01-13 21:47:56 <slush> > nanotube: well, ask is still .4 :P heh
2847 2011-01-13 21:50:30 alystair has joined
2848 2011-01-13 21:52:43 <kiba`> the government doesn't control the speed
2849 2011-01-13 21:52:51 <kiba`> so the market can be really freakishly fast
2850 2011-01-13 21:52:56 <tcatm> Is there a nogui fork of bitcoind?
2851 2011-01-13 21:52:58 <x6763> thanks for all your help ArtForz!
2852 2011-01-13 21:53:15 <kiba`> unregulated tradebots
2853 2011-01-13 21:53:29 <charlesnw> howdy folks. new to bitcoin. sgornick turned me on to it.
2854 2011-01-13 21:53:47 <charlesnw> well i discovered it in june 2009, but didn't pay much attention to it. lol.
2855 2011-01-13 21:53:52 <charlesnw> wish i had started mining back then
2856 2011-01-13 21:54:01 ThomasV has joined
2857 2011-01-13 21:56:09 <nanotube> hey charlesnw
2858 2011-01-13 21:56:44 <charlesnw> hey nanotube
2859 2011-01-13 21:56:45 <kiba`> charlesnw: better start now
2860 2011-01-13 21:56:54 <charlesnw> yep. i've got my client going :)
2861 2011-01-13 21:56:54 <kiba`> accumlating bitcoin
2862 2011-01-13 21:56:57 <kiba`> so you can get rich
2863 2011-01-13 21:57:00 <charlesnw> :)
2864 2011-01-13 21:57:12 <kiba`> pool mining is the best option
2865 2011-01-13 21:57:24 <kiba`> if you want to get bitcoin with some reasonableness
2866 2011-01-13 21:57:33 <charlesnw> i have some gpu resources around. just asked about ps3 (as i have two) but i guess those are too slow. got a couple nvidia cards. no good ati cards though.
2867 2011-01-13 21:57:43 <charlesnw> yes i would love to contribute to a pool
2868 2011-01-13 21:58:12 <charlesnw> and those ps3 are running linux. i've been using them for wpa/md5 hashing/cracking
2869 2011-01-13 21:58:20 <charlesnw> and they are pretty good at that
2870 2011-01-13 21:58:39 <ArtForz> not really
2871 2011-01-13 21:58:41 <charlesnw> i was thinking i would sell gpu time for bitcoins. rent out a small farm with all the software pre configured
2872 2011-01-13 21:58:59 <ArtForz> http://blog.zorinaq.com/?e=42
2873 2011-01-13 21:59:03 <charlesnw> as that seems to be a decent barrier to entry. getting the sdk and what not setup
2874 2011-01-13 21:59:06 <charlesnw> ArtForz, yeah i have seen that
2875 2011-01-13 21:59:17 <charlesnw> and the guy lives about 20 minutes from me
2876 2011-01-13 21:59:21 <ArtForz> 33.1GPW/s for md5 on 4 5970s
2877 2011-01-13 21:59:36 <charlesnw> and works for a company across from the offices of my former company :)
2878 2011-01-13 21:59:42 <charlesnw> small world
2879 2011-01-13 22:00:10 <charlesnw> i plan to pickup an ATI card at some point
2880 2011-01-13 22:00:34 <ArtForz> for crypto work ATI GPGPU just fucking rules
2881 2011-01-13 22:00:41 <charlesnw> it seems to
2882 2011-01-13 22:01:15 <ArtForz> well, actually for any decently vectorizable small problem
2883 2011-01-13 22:01:32 <ArtForz> nvidia looks slightly less crap for noncoherent branching
2884 2011-01-13 22:01:45 <kiba`> charlesnw: so you're looking to enter the mining business?
2885 2011-01-13 22:01:46 <ArtForz> because they have smaller SIMDs
2886 2011-01-13 22:02:20 alhazred has joined
2887 2011-01-13 22:02:24 <ArtForz> but if you dont have that, ATI is ~4-5x faster
2888 2011-01-13 22:02:26 <charlesnw> kiba`, i suppose so. i'm curious anyway. :) i figure i should place a small bet
2889 2011-01-13 22:02:40 <charlesnw> and i've been looking for something cool to do with my gpu capacity
2890 2011-01-13 22:03:10 <charlesnw> and the ps3/nvidia stuff is what i have on hand. and about 12 cores or so spread across a few systems.
2891 2011-01-13 22:03:52 noagendamarket has joined
2892 2011-01-13 22:04:14 <ArtForz> yeah, I wonder if it's really worthwile though
2893 2011-01-13 22:04:21 noagendamarket has quit (Changing host)
2894 2011-01-13 22:04:21 noagendamarket has joined
2895 2011-01-13 22:04:38 <ArtForz> imo nvidia and PS3s are better suited to more "branchy" problems
2896 2011-01-13 22:05:11 <midnightmagic> spus?
2897 2011-01-13 22:05:17 <ArtForz> yep
2898 2011-01-13 22:05:48 <charlesnw> ArtForz, sounds bout right
2899 2011-01-13 22:06:07 <ArtForz> 128 bit SIMD units do a lot better on branchy code than the 1024 or 2048 bit wide SIMDs of ATIs
2900 2011-01-13 22:09:35 <luke-jr> charlesnw: mining isn't very profitable anymore IMO
2901 2011-01-13 22:09:54 <luke-jr> ArtForz made a profit because he's an early adopter w/ the best hw :P
2902 2011-01-13 22:09:59 <ArtForz> it seems a lot of people disagree, or difficulty wouldnt keep going up ;)
2903 2011-01-13 22:10:04 <sgornick> luke-jr: maybe rethinking that at $0.40 perhaps?
2904 2011-01-13 22:10:10 <luke-jr> ArtForz: how many of them break even?
2905 2011-01-13 22:10:18 <ArtForz> who wkows
2906 2011-01-13 22:10:22 <noagendamarket> the price just went back to .35
2907 2011-01-13 22:10:29 <ArtForz> but it sure as hell looks like we're still gaining hashrate
2908 2011-01-13 22:10:37 <luke-jr> sgornick: there's a fixed amount of "free" BTC; the more competition, the less you get unless you're on top
2909 2011-01-13 22:10:46 <charlesnw> luke-jr, you may be right. but it's still an interesting project
2910 2011-01-13 22:10:48 <luke-jr> ArtForz: of course, people don't usually give up
2911 2011-01-13 22:10:58 <lfm> art but we still try to discourage more people from joining!grin
2912 2011-01-13 22:11:00 <luke-jr> they cant exactly return the hw
2913 2011-01-13 22:11:08 <ArtForz> well, it'd be pretty stupid to mine at a loss
2914 2011-01-13 22:11:08 <luke-jr> "wasn't profitable"
2915 2011-01-13 22:11:11 <charlesnw> luke-jr, but there are other uses for the hardware
2916 2011-01-13 22:11:26 <luke-jr> ArtForz: the big loss is up-front
2917 2011-01-13 22:11:29 <charlesnw> it doesn't hurt to have several decent processing clusters around
2918 2011-01-13 22:11:53 <ArtForz> so get into the sl3 unlocking business :P
2919 2011-01-13 22:12:03 <charlesnw> yeah i saw that
2920 2011-01-13 22:12:04 <charlesnw> :)
2921 2011-01-13 22:12:13 <ArtForz> not to mention you can currently resell 5970s without taking that big of a hit
2922 2011-01-13 22:12:32 <charlesnw> prob even make a profit
2923 2011-01-13 22:12:34 <charlesnw> if they are scarce
2924 2011-01-13 22:12:46 <ArtForz> not quite, but close
2925 2011-01-13 22:12:54 <ArtForz> but I wouldnt bet for it to stay that way after 6990 gets released
2926 2011-01-13 22:13:34 <midnightmagic> any specs for 6990 anywhere?
2927 2011-01-13 22:13:40 <ArtForz> nope
2928 2011-01-13 22:13:43 <midnightmagic> bah
2929 2011-01-13 22:13:56 <ArtForz> AMD is a bit paranoid in that area :/
2930 2011-01-13 22:14:04 <ArtForz> good guess is "dual cayman, downclocked"
2931 2011-01-13 22:14:18 <midnightmagic> i thought i saw fewer elements than a 5970 on some slide snapshot.
2932 2011-01-13 22:14:29 <midnightmagic> but i guess that doesn't mean much.
2933 2011-01-13 22:14:36 <ArtForz> 6970 has fewer elements than a 5870
2934 2011-01-13 22:15:14 <ArtForz> they'll be probably using that new power limiting feature introduced with 6950/70 to keep the card < 300W while still allowing decent clocks for "normal" 3D use
2935 2011-01-13 22:15:48 <midnightmagic> i infer from that that a 6970 is faster.
2936 2011-01-13 22:16:00 <ArtForz> for 3D, yes
2937 2011-01-13 22:16:04 <ArtForz> for mining, about the same
2938 2011-01-13 22:16:22 <ArtForz> combine that with better crossfire scaling of 6970 compared to 5870, and you have potential for > +30% 3D perf vs 5970
2939 2011-01-13 22:16:48 <midnightmagic> i'm more interested in computing, even outside of hobby mining.
2940 2011-01-13 22:17:16 <midnightmagic> for a long, long time i was searching for a cellBE grid-on-a-card but man, seriously dude those things are expensive outside a PS3.
2941 2011-01-13 22:17:17 <ArtForz> well, 69xx series is interesting for computing, especially if you do DP stuff
2942 2011-01-13 22:17:36 <midnightmagic> everytime you say 'DP' I giggle a little.
2943 2011-01-13 22:19:27 <ArtForz> 5870 = 272 DP GFLOPS peak, 6970 = 338 DP GFLOPs peak
2944 2011-01-13 22:19:39 <midnightmagic> i love the idea of cheap computing, i'm totally stoked about the possibilities for graph search and sorting problems that my miners will (hopefully) give me post-"worth it" stage.
2945 2011-01-13 22:19:47 <ArtForz> = 24% faster
2946 2011-01-13 22:20:28 <midnightmagic> is that double-counting the muladd op that you were going on about a while ago still?
2947 2011-01-13 22:20:29 <luke-jr> if you're buying an AMD GPU and running blobs *anyway*, then mining *might* be worth it
2948 2011-01-13 22:20:33 <ArtForz> nope
2949 2011-01-13 22:20:48 <luke-jr> but I'm pretty sure buying a GPU just for mining will never break even this late in the game
2950 2011-01-13 22:21:12 <midnightmagic> luke-jr: not so.
2951 2011-01-13 22:21:16 <ArtForz> a tesla M2050 gets ~257.6 real DP FLOPS peak
2952 2011-01-13 22:21:24 <luke-jr> midnightmagic: why?
2953 2011-01-13 22:21:39 <ArtForz> or C2050 or whatever
2954 2011-01-13 22:21:44 <midnightmagic> luke-jr: a cluster of them can still do it easily.
2955 2011-01-13 22:21:55 <ArtForz> not really
2956 2011-01-13 22:22:03 <lfm> depends on power costs a lot
2957 2011-01-13 22:22:04 <luke-jr> midnightmagic: a cluster can easily pay for itself?
2958 2011-01-13 22:22:14 <ArtForz> if one GPU doesnt pay for itself, 10 won't do it either
2959 2011-01-13 22:22:38 <luke-jr> how many BTC are handed out each day?
2960 2011-01-13 22:22:45 <ArtForz> 7200
2961 2011-01-13 22:22:50 <ArtForz> well, 7200 nominal
2962 2011-01-13 22:23:00 <ArtForz> currently about 12% more
2963 2011-01-13 22:23:05 <luke-jr> O.o
2964 2011-01-13 22:23:10 <luke-jr> so $3,600 USD/day
2965 2011-01-13 22:23:19 <luke-jr> not bad actually, but you'd need to get a significant portion
2966 2011-01-13 22:23:24 <ArtForz> huh? since when was 1btc 0.5USD ?
2967 2011-01-13 22:23:33 <luke-jr> ArtForz: since that's the price I'm selling at :P
2968 2011-01-13 22:23:43 <ArtForz> well, good luck with that
2969 2011-01-13 22:23:49 <midnightmagic> LOL are you the one selling that huge 20,000 chunk on mtgox?
2970 2011-01-13 22:23:50 <luke-jr> give it a week or two
2971 2011-01-13 22:23:56 <luke-jr> midnightmagic: I only have 107 total
2972 2011-01-13 22:24:19 <luke-jr> and 100 of that I just bought with USD
2973 2011-01-13 22:24:44 <ArtForz> thats the weird part, by adding more GPUs, you actually lower your return per GPU
2974 2011-01-13 22:25:08 <ArtForz> you grow the total network size, you're effectively competing with yourself
2975 2011-01-13 22:25:11 <luke-jr> if it wasn't for the blob, I'd get a GPU just to make *part* of it back
2976 2011-01-13 22:25:20 <midnightmagic> it surprises me that 5970 gpu mining isn't considered viable by you guys anymore.
2977 2011-01-13 22:25:43 <midnightmagic> but you proportionally pull a larger slice of produced bitcoins.
2978 2011-01-13 22:25:52 <ArtForz> yes
2979 2011-01-13 22:25:52 <luke-jr> I think you can probably make more back in BTC, than the price will deprecate over time
2980 2011-01-13 22:25:57 <ArtForz> but it becomes less and less more
2981 2011-01-13 22:26:34 <midnightmagic> true. but not so much that it's not worth it. the return happens more quickly, which means getting ahead of the growth curve allows you to maintain your share.
2982 2011-01-13 22:27:17 <midnightmagic> even after you add your 100 chips (assuming they all work) the total share is still enough to well pay-off further gpu buys in my cluster.
2983 2011-01-13 22:27:29 * luke-jr notes BTC has grown in value by 10x in 8 months
2984 2011-01-13 22:27:36 <luke-jr> (gregorian months)
2985 2011-01-13 22:27:46 <hacim> luke-jr: when you can generate a block every 1.5 days at the current difficulty with a 5970?
2986 2011-01-13 22:27:50 <midnightmagic> just takes a little longer.
2987 2011-01-13 22:28:01 <ArtForz> well, if the chips work at 200MHz I should be able to match 5970 on hash/$ for the next batches
2988 2011-01-13 22:28:08 <luke-jr> hacim: ?
2989 2011-01-13 22:28:20 <midnightmagic> you brought it down to half cost?
2990 2011-01-13 22:28:28 <hacim> luke-jr: that seems to indicate buying a GPU is profitable and you will break-even in ~2months
2991 2011-01-13 22:28:32 <ArtForz> first batch is this expensive because of setup costs
2992 2011-01-13 22:28:54 <luke-jr> hacim: if you maintain that for 2 months
2993 2011-01-13 22:29:05 <midnightmagic> very nice. are you willing to sell some? :-)
2994 2011-01-13 22:29:10 <hacim> luke-jr: i dont understand that point
2995 2011-01-13 22:29:28 <ArtForz> next batches will be 40-50% for cost per chip, depending on size of single order
2996 2011-01-13 22:29:29 <midnightmagic> hacim: he's trying to say that your 5970 will generate fewer as difficulty increases.
2997 2011-01-13 22:29:31 <luke-jr> hacim: the difficulty rises, more people try to compete
2998 2011-01-13 22:29:39 <hacim> luke-jr: if your point is that you will never break even if you go to costa rica and sleep in a hammock and drink beer all day, then yeah of course
2999 2011-01-13 22:30:00 chuck251 has joined
3000 2011-01-13 22:30:11 * luke-jr ponders if he can throw the GPU into a VM and let the VM run the blob driver and miner
3001 2011-01-13 22:30:25 <midnightmagic> luke-jr: not for xen, as i understand it.
3002 2011-01-13 22:30:26 <hacim> are you saying that if I were to buy a 5970 right now, and fire it up, I would not break even in 2 months?
3003 2011-01-13 22:30:31 <luke-jr> midnightmagic: Xen sucks
3004 2011-01-13 22:31:06 <midnightmagic> hacim: it's a difficulty growth curve (or line). the more difficulty scales up, the fewer coins you mine. you should be able to plug it into an equation based on past difficulty increases over time.
3005 2011-01-13 22:31:12 <midnightmagic> luke-jr: not for NetBSD.
3006 2011-01-13 22:31:16 <ArtForz> signs point to no
3007 2011-01-13 22:31:45 <hacim> so this is just a guess based on the growth curve, not actual data
3008 2011-01-13 22:31:45 <luke-jr> midnightmagic: also need to consider market price
3009 2011-01-13 22:32:16 <midnightmagic> all you can do is guess based on available information.
3010 2011-01-13 22:32:20 andrew12 has joined
3011 2011-01-13 22:32:23 <ArtForz> which also might go... pretty much anywhere
3012 2011-01-13 22:32:25 <luke-jr> ArtForz: what signs?
3013 2011-01-13 22:32:52 <ArtForz> historical growth data
3014 2011-01-13 22:33:12 <hacim> so has anyone actually done those calculations?
3015 2011-01-13 22:33:15 <ArtForz> it's VERY unlikely that we wont see continued exponential growth for the next few months
3016 2011-01-13 22:33:21 <hacim> i wouldn't be surprised if there is an irc bot that can do it
3017 2011-01-13 22:33:32 Prof_BiG_BanG has quit (Ping timeout: 260 seconds)
3018 2011-01-13 22:33:45 <lfm> hacim they are just projections (guesses) not calulations
3019 2011-01-13 22:33:45 <midnightmagic> art probably has. i won't tell. :)
3020 2011-01-13 22:34:03 <hacim> we should be able to determine the current rate of growth and then extrapolate based on that to determine when the breakeven point is
3021 2011-01-13 22:34:08 <luke-jr> ArtForz: I mean, KVM with a video card
3022 2011-01-13 22:34:16 <ArtForz> yep
3023 2011-01-13 22:34:23 <luke-jr> ArtForz: you don't think it will work?
3024 2011-01-13 22:34:34 <midnightmagic> ArtForz: casual people will drop out and weaken the currency. bleh
3025 2011-01-13 22:34:36 <ArtForz> luke-jr: should work
3026 2011-01-13 22:34:51 <luke-jr> ArtForz: so no blobs on my real desktop, just inside the VM?
3027 2011-01-13 22:34:55 <hacim> where is nanotube when its time for some math!
3028 2011-01-13 22:34:59 <luke-jr> and just use my IGP for real video
3029 2011-01-13 22:35:09 <ArtForz> well, last time I checked xen was FOSS
3030 2011-01-13 22:35:12 <luke-jr> wish I had a junk card or something here to test
3031 2011-01-13 22:35:14 <midnightmagic> xensource isn't.
3032 2011-01-13 22:35:18 <luke-jr> ArtForz: KVM, not Xen :P
3033 2011-01-13 22:35:33 <ArtForz> no clue bout KVM
3034 2011-01-13 22:35:38 <luke-jr> Xen can't do it I bet and midnightmagic says
3035 2011-01-13 22:35:52 <ArtForz> not as primary VGA
3036 2011-01-13 22:36:04 <ArtForz> but we dont exactly want to use the card for display output
3037 2011-01-13 22:36:12 <luke-jr> http://bitbud.com/2008/11/12/a-virtualization-first-from-kvm/
3038 2011-01-13 22:36:16 <luke-jr> that was back in 2008
3039 2011-01-13 22:36:34 <ArtForz> yeah, guess it should work
3040 2011-01-13 22:37:20 <luke-jr> all the complex stuff is in the GPU anyway, right?
3041 2011-01-13 22:37:29 <ArtForz> yep
3042 2011-01-13 22:37:46 <ArtForz> the trick usually is to boot on the virtual VGA, then fire up xorg using the passthrough card and propertiary driver
3043 2011-01-13 22:38:32 <luke-jr> I might be willing to do that
3044 2011-01-13 22:39:05 <ArtForz> I dont see how a propertiary blob in a VM is worse than a propertiary blob on a dedicated box, but whatever
3045 2011-01-13 22:39:51 chuck has quit (Quit: leaving)
3046 2011-01-13 22:40:08 <luke-jr> ArtForz: VM can't screw with my system
3047 2011-01-13 22:40:28 <luke-jr> maybe Xen could, but not KVM
3048 2011-01-13 22:40:30 <ArtForz> I wouldnt want to try mining on a box I actually use. who. wants. to. work. on. a. box. as. responsive. as. this. ?
3049 2011-01-13 22:40:40 <luke-jr> â¦
3050 2011-01-13 22:40:52 <luke-jr> why would it affect responsive latency?
3051 2011-01-13 22:40:52 BitterSweet has joined
3052 2011-01-13 22:41:02 <luke-jr> the CPU side can be niced down
3053 2011-01-13 22:41:09 <ArtForz> it pretty much makes X unusable
3054 2011-01-13 22:41:26 <ArtForz> windows take several seconds to redraw
3055 2011-01-13 22:41:39 <midnightmagic> i use a mining machine for browsing, works okay with dwm.
3056 2011-01-13 22:41:45 <ArtForz> my guess is the GPU is completely ignoring X drawing while it's busy in GPGPU land
3057 2011-01-13 22:41:55 <ArtForz> okay, I also run the equivalent of -f 4
3058 2011-01-13 22:42:08 <luke-jr> ArtForz: I'm still talking using IGP for my host system
3059 2011-01-13 22:42:15 <luke-jr> while the AMD GPU does nothing but mine
3060 2011-01-13 22:42:31 <midnightmagic> only window moving is slow but I'm going to test a dwm patch this evening to stop trying to do smooth window moves.
3061 2011-01-13 22:43:54 andrew12 has quit (Changing host)
3062 2011-01-13 22:43:54 andrew12 has joined
3063 2011-01-13 22:45:19 Tritonio_GR has joined
3064 2011-01-13 22:45:19 Tritonio has joined
3065 2011-01-13 22:45:19 Tritonio_GR has quit (Client Quit)
3066 2011-01-13 22:45:49 <lfm> some people will put up with more than others
3067 2011-01-13 22:46:32 <midnightmagic> youtube and full-screen x264 video works.
3068 2011-01-13 22:46:35 <chuck251> On the calculator at http://www.alloscomp.com/bitcoin/calculator.php , I am not clear what the 50% time means? Why would 50% not be the same as Average?
3069 2011-01-13 22:47:02 <ArtForz> statistics
3070 2011-01-13 22:48:17 Tritonio has quit (Client Quit)
3071 2011-01-13 22:48:37 Tritonio has joined
3072 2011-01-13 22:48:41 <cosurgi> that's why my 3 days, are 5 already :)
3073 2011-01-13 22:48:48 <cosurgi> 5 days and no block :)
3074 2011-01-13 22:49:18 <lfm> thats called luck
3075 2011-01-13 22:54:04 <slush> cosurgi: ,,pool ? :-)
3076 2011-01-13 22:54:05 <gribble> No fancy GPU farm, and don't want to wait for months for a block gen? Join the mining pool! http://mining.bitcoin.cz/
3077 2011-01-13 22:54:35 <ArtForz> yup, pools is a quick way to check if your GPU works as expected
3078 2011-01-13 22:56:04 <midnightmagic> testnet is too.
3079 2011-01-13 22:56:30 <ArtForz> well, testnet isnt quite as quick
3080 2011-01-13 22:56:44 <midnightmagic> pool is quicker? awesome!!
3081 2011-01-13 22:57:11 <ArtForz> pool is 1.0 diff
3082 2011-01-13 22:59:36 altamic has joined
3083 2011-01-13 23:00:17 Tritonio has quit (Quit: Leaving)
3084 2011-01-13 23:01:03 newsham has joined
3085 2011-01-13 23:06:32 Prof_BiG_BanG has joined
3086 2011-01-13 23:06:32 <davex__> does anyone run a cartellized pool? Or was that idea ever disproven to work?
3087 2011-01-13 23:07:01 <chuck251> I was in slush's pool, then I saw I'd found 4 blocks (when average seemed like it should be 1 or 2), so feeling lucky I started to mine solo, and now I am 16 hours over the expected average :)
3088 2011-01-13 23:07:27 <kiba`> LOL
3089 2011-01-13 23:07:48 <kiba`> probablity man, probablity
3090 2011-01-13 23:09:06 <donpdonp> its a curious setup to have half the world's 5970's furiously racing towards the same hash. there can only be one winner. sounds like some kind of Mad Max Thunder Dome setup
3091 2011-01-13 23:09:37 <nolan_d> What Linux library do I need to use the OpenCL-based miner? Getting an error that the OpenCL lib can't be found but "apt-cache search opencl" doesn't show anything relevant.
3092 2011-01-13 23:09:44 <donpdonp> I laughed at the pool before, but it seems to have changed the rules of the game if you get a little btc without finding the winning block
3093 2011-01-13 23:10:00 <arcatan> chuck251: 50 % means the median. let's say there are four people mining and three of them find a block in one year and the fourth one takes 9 years to find one. now the average is (1+1+1+9)/4=3 years, but the median is only 1 year
3094 2011-01-13 23:10:16 <ArtForz> iirc you need the ATI stream SDK, preferably v2.1
3095 2011-01-13 23:10:19 <arcatan> because 50 % of the miners found a block in one year or less
3096 2011-01-13 23:10:22 <lfm> donpdonp everyone works on their own hashes, not the same one
3097 2011-01-13 23:10:31 <Diablo-D3> donpdonp: you need to install the one for your driver.
3098 2011-01-13 23:10:41 <ArtForz> well, it's still kinda a race
3099 2011-01-13 23:10:44 <nolan_d> Hmm, OK, wasn't sure if that was prepackaged for Ubuntu. I'll head to Google.
3100 2011-01-13 23:11:38 larsivi has quit (Remote host closed the connection)
3101 2011-01-13 23:11:39 <ArtForz> errr.. ATI or nvidia?
3102 2011-01-13 23:11:40 <donpdonp> lfm: ORLY? isnt there just 'the next block' to find the winning nonce for and hence the correct hash value?
3103 2011-01-13 23:11:53 <ArtForz> nope
3104 2011-01-13 23:12:13 <ArtForz> theres about 2**224 valid next blocks
3105 2011-01-13 23:12:33 <alhazred> the next block includes your transaction to pay yourself for generating it, which changes the merkle root
3106 2011-01-13 23:12:33 <lfm> donpdonp there are zillions of possible next blocks
3107 2011-01-13 23:12:45 <ArtForz> yep, about 2**224 / difficulty
3108 2011-01-13 23:13:00 <donpdonp> ok well that must be a big hole in my understanding. the block is a 'chain' and hashing the next block needs the hash from the previous block. thats why it seems one can only work on the 'next' block
3109 2011-01-13 23:13:11 <donpdonp> s/is a chain/ is part of a chain/
3110 2011-01-13 23:13:12 <nolan_d> lspci identifies it as a Radeon 3100.
3111 2011-01-13 23:13:29 <ArtForz> errr... I dont think OpenCL works on 3-series cards
3112 2011-01-13 23:13:38 mtgox has joined
3113 2011-01-13 23:13:58 <nolan_d> All righty then, back to the CPU-based miner.
3114 2011-01-13 23:14:08 <lfm> donpdonp yes but you are trying to find "a" next block, not "the" next block
3115 2011-01-13 23:14:13 <ArtForz> yep
3116 2011-01-13 23:14:15 <ArtForz> ATI 4xxx, 5xxx, and 6xxx series, and 4xxx is already pretty much too slow and inefficient
3117 2011-01-13 23:14:53 <ArtForz> soemnone shoudl write a Brook+ miner for people stuck on 2xxx and 3xxx ATI cards
3118 2011-01-13 23:15:46 <newsham> nolan: opencl libs are installed when you install nvidia driver (if thats the kind of card you have)
3119 2011-01-13 23:15:58 <ArtForz> he ahs a ATI 3xxx
3120 2011-01-13 23:16:11 <nolan_d> Another question, I've read the page on pooled mining. It compares approaches, but is there one that is preferred/used more heavily? A comparison is one thing, but actual use is another. :)
3121 2011-01-13 23:16:15 <ArtForz> but 3xxx isnt supported by ATIs OCL runtime
3122 2011-01-13 23:16:31 <ArtForz> most people seem to be using slushs pool
3123 2011-01-13 23:16:40 <ArtForz> probably because it'S the only active pool :P
3124 2011-01-13 23:16:49 <newsham> thats an advantage
3125 2011-01-13 23:16:49 <nanotube> ;;bc,poolstats
3126 2011-01-13 23:16:50 <gribble> {"hashes_ps": 15957794804, "shares": 1802, "active_workers": 306, "round_duration": "0:08:05", "round_started": "2011-01-13 23:04:45", "getwork_ps": 100}
3127 2011-01-13 23:16:54 <alhazred> why is the transaction hash of a block stored as a merkle tree hash rather than just a straight hash over all the transactions?
3128 2011-01-13 23:17:00 <nolan_d> Fair reason. :) That's why I asked, thanks.
3129 2011-01-13 23:17:03 <nanotube> ArtForz: oh nobody is running a puddinpop-style pool anymore?
3130 2011-01-13 23:17:17 <ArtForz> iirc puddinpop shit down his pool
3131 2011-01-13 23:17:21 <ArtForz> *shut
3132 2011-01-13 23:17:47 <nanotube> alhazred: makes it easier to mine. also makes it possible to remove spent tx in the future.
3133 2011-01-13 23:17:57 <lfm> alhazred it is so you can drop a sub-tree and keep the rest dynsamiclly
3134 2011-01-13 23:18:06 <ArtForz> yep, probably the "makes it possible to forget about spent TX" part
3135 2011-01-13 23:18:19 <alhazred> won't dropping a sub tree change the hash?
3136 2011-01-13 23:18:30 <ArtForz> not if you keep the top hash of the branch
3137 2011-01-13 23:18:37 <kiba`> yay
3138 2011-01-13 23:18:43 <kiba`> I sold my first writing
3139 2011-01-13 23:18:49 <lfm> wtg
3140 2011-01-13 23:19:02 <donpdonp> lfm: the next block is the set of transactions 'up next' to incorporate into a block, right? perhaps if there is a large number of tx's to be incorporated into the chain, those are divided into different potential blocks?
3141 2011-01-13 23:19:14 <newsham> you can prune the tree
3142 2011-01-13 23:19:36 <lfm> alhazzred you keep the same hashes when you drop parts of the tree
3143 2011-01-13 23:19:50 Tritonio has joined
3144 2011-01-13 23:19:50 <tcatm> Does CRITICAL_BLOCK(cs_mapWallet) always execute the block?
3145 2011-01-13 23:19:52 <alhazred> ah right i get it
3146 2011-01-13 23:20:16 <alhazred> does the client prune branches currently?
3147 2011-01-13 23:20:28 <lfm> note yet
3148 2011-01-13 23:20:33 <lfm> not yet
3149 2011-01-13 23:21:12 <ArtForz> no need yet really
3150 2011-01-13 23:21:44 <ArtForz> total size of block chain is 62.4MB
3151 2011-01-13 23:23:40 <alhazred> you need the entire block chain to check the balance of a public key, am i rite?
3152 2011-01-13 23:23:52 Tritonio has quit (Client Quit)
3153 2011-01-13 23:24:17 <lfm> tcatm yes I think so, it just can be blocked at the beginning if another thrtead is in the same block
3154 2011-01-13 23:24:20 <alhazred> i mean, not just the headers
3155 2011-01-13 23:25:21 <alhazred> just wondering how a thin client that only gets the block headers is going to work if you don't have the transactions.
3156 2011-01-13 23:25:45 <tcatm> lfm: mhm g++ complains: warning: control reaches end of non-void function
3157 2011-01-13 23:26:05 <lfm> alhazred actually you dont need the block headers, you only need the outstanding txn
3158 2011-01-13 23:26:41 Zarutian has quit (Quit: Zarutian)
3159 2011-01-13 23:27:05 <lfm> tcatm missing return val;
3160 2011-01-13 23:27:48 <tcatm> well, it has a return inside the critical block that is always reached when the block is executed
3161 2011-01-13 23:28:07 <tcatm> (rpc.cpp getaccountaddress())
3162 2011-01-13 23:28:18 <lfm> tcatm there should not be a return in the middle of a critical section!
3163 2011-01-13 23:29:02 <lfm> tcatm you should always leave the critical section at the bottom
3164 2011-01-13 23:29:46 <lfm> tcatm you may need to use a flag variable or something
3165 2011-01-13 23:30:22 <tcatm> than that might be the bug
3166 2011-01-13 23:31:31 <ArtForz> yep
3167 2011-01-13 23:31:40 <tcatm> https://github.com/bitcoin/bitcoin/issues/#issue/28 related to this? :)
3168 2011-01-13 23:31:44 <ArtForz> never return from inside a critical section
3169 2011-01-13 23:34:54 BitterSweet has quit (Quit: Leaving.)
3170 2011-01-13 23:35:27 altamic has quit (Ping timeout: 276 seconds)
3171 2011-01-13 23:36:11 <tcatm> grml. bitcoin coding style requires 4 spaces instead of tabs.
3172 2011-01-13 23:37:12 <ArtForz> what a waste of bits
3173 2011-01-13 23:37:47 TD has quit (Quit: TD)
3174 2011-01-13 23:38:33 <ArtForz> 4 spaces consume 75% more bits than a tab, think of the poor bits!
3175 2011-01-13 23:38:45 <ArtForz> or 300% more
3176 2011-01-13 23:38:51 <ArtForz> sounds more impressive
3177 2011-01-13 23:39:45 <Diablo-D3> 2
3178 2011-01-13 23:39:48 <Diablo-D3> 2 fucking spaces
3179 2011-01-13 23:39:50 <Diablo-D3> not 1
3180 2011-01-13 23:39:51 <Diablo-D3> not 3
3181 2011-01-13 23:39:52 <Diablo-D3> not 4
3182 2011-01-13 23:39:52 <Diablo-D3> 2
3183 2011-01-13 23:39:53 fabianhjr has joined
3184 2011-01-13 23:40:00 <fabianhjr> Hi kiba` :)
3185 2011-01-13 23:40:05 <kiba`> hi
3186 2011-01-13 23:40:17 <ArtForz> but I want to watch a space orgy
3187 2011-01-13 23:40:21 <kiba`> when will you publish your magaizne?
3188 2011-01-13 23:40:32 <kiba`> ArtForz: an orgy in a space opera?
3189 2011-01-13 23:40:40 <kiba`> or an orgy on the international space station?
3190 2011-01-13 23:41:24 <ArtForz> "2 fucking spaces" is boring, there need to be at least 10 spaces and a bunch of tabs and newlines for good measure
3191 2011-01-13 23:41:25 <lfm> or an orgy of ETs
3192 2011-01-13 23:41:47 <fabianhjr> kiba`: it got delayed. No article submissions. I will probably just release a simple magazine first and then do the community driven idea.
3193 2011-01-13 23:42:00 <Diablo-D3> [06:36:17] <ArtForz> but I want to watch a space orgy
3194 2011-01-13 23:42:14 <lfm> yes all code should be triple spaced so your teacher can put notes between the lines of the printout
3195 2011-01-13 23:42:15 <Diablo-D3> all we need is that green chick from star trek and that blue chick from farscape
3196 2011-01-13 23:42:17 <Diablo-D3> actually
3197 2011-01-13 23:42:24 <Diablo-D3> lets just go with ALL the chicks from farscape
3198 2011-01-13 23:42:36 <lfm> and 7 of 9
3199 2011-01-13 23:43:57 Slix` has joined
3200 2011-01-13 23:44:45 <Diablo-D3> meh
3201 2011-01-13 23:44:49 <Diablo-D3> 7 of 9 is ugly, to be honest
3202 2011-01-13 23:44:57 <Diablo-D3> sure, her borg implants are nice, but bleh
3203 2011-01-13 23:45:13 <Diablo-D3> and I prefer them natural, anyhow
3204 2011-01-13 23:48:46 fabianhjr has quit (Read error: Connection reset by peer)
3205 2011-01-13 23:49:09 fabianhjr has joined
3206 2011-01-13 23:49:18 <fabianhjr> Stupid Telmex. D:<
3207 2011-01-13 23:51:42 <nevezen> is the bitcoin client suppose to create a new address every time it starts up?
3208 2011-01-13 23:52:05 <nanotube> not every time it starts, no.
3209 2011-01-13 23:52:39 <nevezen> what's the point for it creating new addresses?
3210 2011-01-13 23:52:40 <nanotube> if there haven't been new tx, no new addresses should be automagically created
3211 2011-01-13 23:52:49 <nanotube> extra anonymity
3212 2011-01-13 23:53:21 <nevezen> so you can use them as your own recieving addresses?
3213 2011-01-13 23:55:21 <fabianhjr> kiba`: which format is the iinterview? txt?
3214 2011-01-13 23:56:37 <sipa> ;;bc,stats
3215 2011-01-13 23:56:40 <gribble> Current Blocks: 102543 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 272 blocks | Next Difficulty In About: 1 day, 14 hours, 27 minutes, and 28 seconds | Next Difficulty Estimate: 18324.44507020
3216 2011-01-13 23:58:58 <nanotube> nevezen: yes any address in your receiving addresses list can be used to receive.