1 2010-11-17 00:00:11 <AAA_awright> We could port that
2 2010-11-17 00:00:47 <tcatm> How long will it take?
3 2010-11-17 00:00:56 hippich_ has joined
4 2010-11-17 00:01:00 <tcatm> I can't work fulltime on it.
5 2010-11-17 00:02:05 <noagendamarket> anarchyx could do a music video for your song lol
6 2010-11-17 00:02:52 <UukGoblin> right
7 2010-11-17 00:02:59 <UukGoblin> got all the libs on 10.04 ubuntu
8 2010-11-17 00:03:05 <UukGoblin> and m0mchil's thing gives:
9 2010-11-17 00:03:12 <UukGoblin> Traceback (most recent call last):
10 2010-11-17 00:03:12 <UukGoblin> File "poclbm.py", line 55, in <module>
11 2010-11-17 00:03:12 <UukGoblin> platform = cl.get_platforms()[0]
12 2010-11-17 00:03:12 <UukGoblin> pyopencl.LogicError: clGetPlatformIDs failed: invalid/unknown error code
13 2010-11-17 00:03:24 <UukGoblin> the same thing it did without all the libs
14 2010-11-17 00:07:12 eureka^ has quit (Ping timeout: 250 seconds)
15 2010-11-17 00:08:20 <UukGoblin> diablo's thing gives:
16 2010-11-17 00:08:29 <UukGoblin> Exception in thread "main" java.lang.Exception: No OpenCL platforms found.
17 2010-11-17 00:08:29 <UukGoblin> at com.diablominer.DiabloMiner.DiabloMiner.execute(DiabloMiner.java:163)
18 2010-11-17 00:08:32 <UukGoblin> at com.diablominer.DiabloMiner.DiabloMiner.main(DiabloMiner.java:86)
19 2010-11-17 00:08:54 <UukGoblin> I must be missing some opencl component
20 2010-11-17 00:14:54 eureka^ has joined
21 2010-11-17 00:15:40 theymos has joined
22 2010-11-17 00:16:54 <theymos> Two transactions can apparently have the exact same hash (especially generation transactions). When one of these transactions is referenced in an input, how does Bitcoin decide which one gets spent?
23 2010-11-17 00:24:03 <UukGoblin> aha
24 2010-11-17 00:24:09 <UukGoblin> I'm missing the icd-registration I think
25 2010-11-17 00:35:49 * Kiba is a bit miffed about the lack of feedback
26 2010-11-17 00:35:59 <Kiba> lack of feedback lead to unmotiviated kiba
27 2010-11-17 00:41:18 <MT`AwAy> Kiba: feedback on?
28 2010-11-17 00:41:35 <Kiba> http://www.bitcoin.org/smf/index.php?topic=1557.0
29 2010-11-17 00:43:58 <AAA_awright> Kiba: What about a micropayments system where you can buy items to help you on your "quest"
30 2010-11-17 00:44:00 <AAA_awright> instead
31 2010-11-17 00:46:39 <Kiba> like what?
32 2010-11-17 01:04:09 <altamic> Kiba: I am playing with your toy
33 2010-11-17 01:08:45 <johnyh> who is using Diablos miner? is it good?
34 2010-11-17 01:13:20 altamic has quit (Quit: altamic)
35 2010-11-17 01:15:05 davex__ has joined
36 2010-11-17 01:15:39 <davex__> What does the number in my "3/unconfirmed" status string represent?
37 2010-11-17 01:18:32 <brocktice> johnyh: I've tried it, didn't work well on my 5970s.
38 2010-11-17 01:18:43 <brocktice> But according to him it's because my system is wack
39 2010-11-17 01:19:02 <brocktice> davex__: That means three confirmations have been done, I think?
40 2010-11-17 01:19:10 <brocktice> It's not considered confirmed until it hits six?
41 2010-11-17 01:19:22 <AAA_awright> What block number are we on?
42 2010-11-17 01:19:40 <davex__> ok
43 2010-11-17 01:19:45 <brocktice> 92279 is next I think
44 2010-11-17 01:19:57 <AAA_awright> Kiba: idk, maybe boots of speed or something, grab a list from IdleRPG
45 2010-11-17 01:25:17 <AAA_awright> ~blocks
46 2010-11-17 01:25:23 <AAA_awright> What's the command?
47 2010-11-17 01:25:28 <XxMalouinxX> ;estimate
48 2010-11-17 01:25:32 <theymos> ;;bc,blocks
49 2010-11-17 01:25:34 <gribble> 92281
50 2010-11-17 01:26:02 <AAA_awright> Hmm, there has to be a better way to do this than downloading ten-thousand blocks at a time
51 2010-11-17 01:26:36 <AAA_awright> Has anyone submitted the EFF donation announcement to slashdot yet?
52 2010-11-17 01:28:00 warner has quit (Quit: ERC Version 5.3 (IRC client for Emacs))
53 2010-11-17 01:28:40 <tcatm> which announcement?
54 2010-11-17 01:30:04 <AAA_awright> EFF seems to be accepting Bitcoin donations
55 2010-11-17 01:30:13 <AAA_awright> I wonder what they are planning to do with it
56 2010-11-17 01:30:17 <AAA_awright> Sell it?
57 2010-11-17 01:30:19 <tcatm> yes but they didn't announce it
58 2010-11-17 01:30:36 <AAA_awright> Someone did
59 2010-11-17 01:30:40 <AAA_awright> I wonder why they didn't
60 2010-11-17 01:32:45 <theymos> So apparently transaction and block hashes can be identical to one another without breaking SHA-256. I wonder if Bitcoin is able to handle these edge cases properly.
61 2010-11-17 01:34:33 <tcatm> maybe you can ask on the forum so satoshi might answer
62 2010-11-17 01:35:57 <AAA_awright> I appear to be generating even though I only have 80546 blocks... any idea what's going on?
63 2010-11-17 01:36:08 <AAA_awright> 254k hash/s
64 2010-11-17 01:36:14 <theymos> I think it'd just ignore identical blocks, even if such blocks could be valid next blocks.
65 2010-11-17 01:36:46 <brocktice> Is nobody else seeing a memory leak in bitcoind?
66 2010-11-17 01:37:01 <brocktice> valgrind isn't showing anything yet, I guess it woulld.
67 2010-11-17 01:37:03 <brocktice> *would
68 2010-11-17 01:37:13 <brocktice> but should bitcoin be reading or exceeding 1GB of RAM over time?
69 2010-11-17 01:37:25 <brocktice> reaching, damn my fingers
70 2010-11-17 01:37:42 <tcatm> my bitcoind has been running for over a month without leaking
71 2010-11-17 01:37:54 <brocktice> tcatm: do you use getwork?
72 2010-11-17 01:37:58 <tcatm> no
73 2010-11-17 01:37:58 <brocktice> also which version?
74 2010-11-17 01:38:12 <tcatm> it's based on 311
75 2010-11-17 01:38:28 <brocktice> hm, ok
76 2010-11-17 01:38:29 <tcatm> wrote my own tcp work dispatcher
77 2010-11-17 01:39:42 <doublec> brocktice, I notice memory increasing too
78 2010-11-17 01:40:05 <doublec> brocktice, I periodically restart it
79 2010-11-17 01:40:33 Xunie has quit (Ping timeout: 265 seconds)
80 2010-11-17 01:41:50 <brocktice> doublec: Yeah I now have a cron job on my miner resetting it
81 2010-11-17 01:41:59 <brocktice> doublec: are you using a patched client or normal?
82 2010-11-17 01:42:29 <doublec> brocktice, the normal client
83 2010-11-17 01:42:46 <brocktice> doublec: hm, so it's not the getwork patch then. bitcoind or the graphical client? Which version?
84 2010-11-17 01:42:57 Xunie has joined
85 2010-11-17 01:43:28 <doublec> brocktice, bitcoind. I've noticed it since about 0.3.13
86 2010-11-17 01:43:40 <brocktice> OK, thanks
87 2010-11-17 01:44:38 <brocktice> oh, valgrind did find leaks
88 2010-11-17 01:45:11 <doublec> nice
89 2010-11-17 02:12:49 <johnyh> brocktice: bugs in what, the bitcoind ?
90 2010-11-17 02:14:31 <brocktice> johnyh: actually in the GUI client
91 2010-11-17 02:14:38 <brocktice> it has issues with bitcoind because it forks or something
92 2010-11-17 02:14:41 <brocktice> I'm not valgrind expert
93 2010-11-17 02:14:57 <brocktice> Vodka does bad things to my typing, sorry.
94 2010-11-17 02:15:23 <brocktice> But, on my miner bitcoin slowly uses memory over time, so that I have to reset it.
95 2010-11-17 02:15:59 <brocktice> I tried running it in valgrind for a few minutes and it did find some leakage. doublec was saying he'd noticed it in bitcoind since 0.3.13
96 2010-11-17 02:17:02 <ArtForz> yeah, there seems to be something wrong
97 2010-11-17 02:17:27 <brocktice> ArtForz: do you restart it periodically or what?
98 2010-11-17 02:17:31 <brocktice> That seems to be what works for me.
99 2010-11-17 02:17:33 <ArtForz> yeah
100 2010-11-17 02:17:38 <brocktice> feels dirty though
101 2010-11-17 02:17:42 <ArtForz> I did some TX spamming on TEST net, bitcoind consumed 2.3GB ...
102 2010-11-17 02:18:04 <brocktice> heh
103 2010-11-17 02:18:06 <ArtForz> and nothing THAT crazy, just a few 1000 TX
104 2010-11-17 02:18:14 <brocktice> that reminds me, I'd better get that swapfile working on my miner
105 2010-11-17 02:18:50 <ArtForz> so.. yeah, I think we have a leak somewhere
106 2010-11-17 02:19:17 <ArtForz> my guess is old cached transactions keep hanging around in memory
107 2010-11-17 02:21:24 <brocktice> So, I predict we'll see a pretty big dropoff in new miners.
108 2010-11-17 02:21:33 <brocktice> I looked into buying some more hardware.
109 2010-11-17 02:21:39 <brocktice> The payoff times are getting quite long.
110 2010-11-17 02:21:48 <brocktice> (with upcoming difficulty of 6500 or so)
111 2010-11-17 02:21:59 <ArtForz> yep
112 2010-11-17 02:22:17 <brocktice> Thought I'd wedge in another 5970 and a few more 5770s
113 2010-11-17 02:22:29 <brocktice> But I don't currently think it's worth it. If prices go up for bitcoins that could change.
114 2010-11-17 02:22:30 <ArtForz> and I'm guessing more towards 6800ish
115 2010-11-17 02:23:25 <brocktice> So, I wonder what will happen, guess we'll see.
116 2010-11-17 02:23:50 <ArtForz> unless price skyrockets I'm prolly gonna add 2 more boxes @ 4 5970 and thats it
117 2010-11-17 02:24:18 <brocktice> You said stock 5770s are around 180 Mhash/s right?
118 2010-11-17 02:24:22 <ArtForz> sec
119 2010-11-17 02:24:35 <ArtForz> no, 156ish
120 2010-11-17 02:24:39 <brocktice> ouch
121 2010-11-17 02:24:46 <ArtForz> OCed to 1000 they're 180ish
122 2010-11-17 02:24:51 <brocktice> oh maybe that was the OC'ed value
123 2010-11-17 02:25:03 <ArtForz> errr... no, at 960 they're 180ish
124 2010-11-17 02:25:30 <ArtForz> at 1000 core ~ 188Mh/s using my miner
125 2010-11-17 02:25:46 <brocktice> Yeah that'd be 21 days to pay off at diff 6800
126 2010-11-17 02:25:54 <brocktice> give or take
127 2010-11-17 02:26:05 <brocktice> at $0.2/btc
128 2010-11-17 02:26:22 <ArtForz> and of course diff seems to x1.5 every ~10 days
129 2010-11-17 02:26:48 <brocktice> Like I said, I think the slope will probably decrease soon.
130 2010-11-17 02:27:04 <ArtForz> my guess is, somewhere in the 7k-10k region
131 2010-11-17 02:27:08 <brocktice> unless price rises
132 2010-11-17 02:27:13 <ArtForz> yep
133 2010-11-17 02:27:42 <ArtForz> around there nv mining becomes unprofitable for people paying "normal" power prices
134 2010-11-17 02:29:52 <theymos> ArtForz: Were you the one who made those two transactions with identical hashes (d5d27987d2 in 91812 and 91842)?
135 2010-11-17 02:30:36 <ArtForz> nope
136 2010-11-17 02:30:57 <theymos> Do you know what Bitcoin does when someone spends them?
137 2010-11-17 02:31:11 <ArtForz> good question
138 2010-11-17 02:31:24 <ArtForz> I *think* one of em gets ignored
139 2010-11-17 02:31:41 <theymos> That's my thinking, too.
140 2010-11-17 02:31:49 <brocktice> It's kind of nice that my office doesn't get too cold
141 2010-11-17 02:32:29 <ArtForz> it's WAY too warm for november here
142 2010-11-17 02:32:39 <ArtForz> thats why I'm only running 8/12 5970s
143 2010-11-17 02:33:25 <brocktice> ArtForz: which part of .de?
144 2010-11-17 02:33:50 <ArtForz> bavaria
145 2010-11-17 02:34:10 <brocktice> almost identical to here (Minnesota, US)
146 2010-11-17 02:34:29 <brocktice> ArtForz: Ever heard of a place called Joedlerwirt in Munich?
147 2010-11-17 02:35:30 <ArtForz> nope
148 2010-11-17 02:35:52 <brocktice> Oh, it's a lot of fun, might check it out if you're there looking for a good place to drink and listen to yodeling.
149 2010-11-17 02:36:21 <brocktice> Right off Marienplatz
150 2010-11-17 02:37:13 <brocktice> anyway, I'm out, g'night all
151 2010-11-17 02:47:11 bryogenic has left ()
152 2010-11-17 02:48:06 <AAA_awright> How do two identical hashes form?
153 2010-11-17 02:48:16 <AAA_awright> Did we crack SHA1?
154 2010-11-17 02:48:29 <ArtForz> it's 2 identical coinbase TX
155 2010-11-17 02:48:41 <theymos> They reuse their key, and happen to get the same ExtraNonce.
156 2010-11-17 02:48:55 <AAA_awright> It doesn't contain the previous block hash/
157 2010-11-17 02:48:58 <AAA_awright> ?
158 2010-11-17 02:49:05 <theymos> Transactions don't.
159 2010-11-17 02:49:16 <ArtForz> coinbase? no
160 2010-11-17 02:49:33 <AAA_awright> Oh, transactions... hmm
161 2010-11-17 02:50:10 <AAA_awright> Yeah back to my comment, we need a million-bitcoin prize for someone who can prove or disprove the possibility that we can keep all the same functionality without a transaction log
162 2010-11-17 02:51:00 <theymos> It's impossible to prove that something doesn't exist (ie no double-spends) without having a complete log of all unspent transaction.
163 2010-11-17 02:51:09 <AAA_awright> Has that been proven?
164 2010-11-17 02:51:26 <AAA_awright> Either make it impossible or mathematically hard
165 2010-11-17 02:51:33 <AAA_awright> Bitcoin is hard
166 2010-11-17 02:51:53 <AAA_awright> What I'm thinking is a transaction that changes a bit array, in such a way that over-spending doesn't cleanly apply the change, and is rejected
167 2010-11-17 02:52:36 <AAA_awright> It sort of makes intuitive sense and I would think it's possible
168 2010-11-17 02:55:11 <davex__> AAA_awright, Isn't that what the paper bitcoin was based on showed? (i haven't actually read it)
169 2010-11-17 02:55:40 <AAA_awright> The whitepaper only shows how Bitcoin works, and that it is mathematically hard (that we know of)
170 2010-11-17 02:57:04 <ArtForz> how do you spose clients sync their world view?
171 2010-11-17 02:57:09 AAA_awright_ has joined
172 2010-11-17 03:00:02 <ArtForz> bitcoin uses the block chain to determine which of several TX using the same input is the "winner"
173 2010-11-17 03:01:35 AAA_awright has quit (Ping timeout: 276 seconds)
174 2010-11-17 03:04:48 AAA_awright_ is now known as AAA_awright
175 2010-11-17 03:07:40 Diablo-D3 has joined
176 2010-11-17 03:13:57 <davex__> can bitcoin use an alternative wallet to ~/.bitcoin/wallet.dat?
177 2010-11-17 03:14:26 <ArtForz> it can use a alternative datadir
178 2010-11-17 03:27:25 xelister has joined
179 2010-11-17 03:30:29 <Diablo-D3> you can also get special limited edition all cow leather ones
180 2010-11-17 03:33:11 <MT`AwAy> haha I'm so rich, got over 800 coins [on testnet]
181 2010-11-17 03:33:15 <MT`AwAy> t_t
182 2010-11-17 03:33:30 <ArtForz> 121600
183 2010-11-17 03:34:01 <ArtForz> and another 150k or so in another wallet
184 2010-11-17 03:34:15 * xelister steals ArtForz's wallet
185 2010-11-17 03:34:15 <MT`AwAy> ArtForz: on live?
186 2010-11-17 03:34:22 <ArtForz> TEST :P
187 2010-11-17 03:34:29 <MT`AwAy> pfew
188 2010-11-17 03:34:51 * MT`AwAy can imagine how someone could sink the economy by selling 300k coins for really cheap on mtgox
189 2010-11-17 03:35:01 <ArtForz> on live I only mined 200k total
190 2010-11-17 03:35:15 <MT`AwAy> well, most of the coins I make on test I send them to the faucet
191 2010-11-17 03:35:27 <xelister> MT`AwAy: what?
192 2010-11-17 03:35:35 <MT`AwAy> once I finish testing
193 2010-11-17 03:35:36 <xelister> coins made on testnet are usable anywhere on live?
194 2010-11-17 03:35:42 <MT`AwAy> xelister: no no no
195 2010-11-17 03:35:44 <MT`AwAy> they are not
196 2010-11-17 03:35:52 <MT`AwAy> read the discussion :D
197 2010-11-17 03:35:55 <xelister> so what you mean you send them to the faucet
198 2010-11-17 03:36:03 <MT`AwAy> xelister: there's a test faucet
199 2010-11-17 03:36:11 <MT`AwAy> http://freebitcoins.appspot.com/TEST/ <- there
200 2010-11-17 03:36:19 <MT`AwAy> so you can send test coins to the test faucet
201 2010-11-17 03:37:19 <xelister> ah ok I thought you ment https://freebitcoins.appspot.com/
202 2010-11-17 03:37:25 <MT`AwAy> mh, someone is generating like crazy on test, been 70 blocks since my last generation (was generating every 1~2 blocks before)
203 2010-11-17 03:37:37 <ArtForz> thats a 5770
204 2010-11-17 03:37:48 <MT`AwAy> xelister: test & live cannot communicate, and even if it did, coins you have on test are not on the live blockchain, so it wouldn't work anyway
205 2010-11-17 03:38:08 <xelister> MT`AwAy: yeah that is what I expected; Just thought of the real faucet :) carry on
206 2010-11-17 03:38:21 <MT`AwAy> I will :)
207 2010-11-17 03:41:00 <Edogaa> Hey
208 2010-11-17 03:41:04 <Edogaa> what ports should i foreward
209 2010-11-17 03:41:20 tcatm has quit (Quit: quit)
210 2010-11-17 03:41:23 <Edogaa> new modem me forgot to check what ports i forewarded the last time
211 2010-11-17 03:41:25 <Edogaa> so yeah
212 2010-11-17 03:41:34 <ArtForz> 8333
213 2010-11-17 03:41:34 <Diablo-D3> ALL OF THEM
214 2010-11-17 03:41:46 * Edogaa kicks Diablo-D3.
215 2010-11-17 03:42:27 <Edogaa> UDP or TCP?
216 2010-11-17 03:42:38 <Edogaa> it be faster if I only need to foreward one or the other lol
217 2010-11-17 03:42:44 <theymos> TCP
218 2010-11-17 03:42:45 <ArtForz> TCP
219 2010-11-17 03:43:53 <Edogaa> will this icnrease my connection rate
220 2010-11-17 03:44:00 <ArtForz> TEST faucet now has >150k available ;)
221 2010-11-17 03:44:16 <Edogaa> If only they were real bitcoins
222 2010-11-17 03:44:20 <Edogaa> :|
223 2010-11-17 03:44:32 <Diablo-D3> see
224 2010-11-17 03:44:35 <Diablo-D3> bitcoin sucks
225 2010-11-17 03:44:38 <Diablo-D3> the port isnt over 9000
226 2010-11-17 03:44:47 <Edogaa> :|
227 2010-11-17 03:44:51 * Edogaa kicks Diablo-D3.
228 2010-11-17 03:44:52 <ArtForz> test is 18333
229 2010-11-17 03:45:36 <xelister> btw, so the miners seem to make some of the first rounds "in python" (for m0m's)
230 2010-11-17 03:45:43 <ArtForz> it's *twice* OVER 9000!!11one
231 2010-11-17 03:45:52 <xelister> is that a good idea, why not do all rounds on gpu
232 2010-11-17 03:45:58 <ArtForz> </meme>
233 2010-11-17 03:46:02 <kisom_dev> w/in 2
234 2010-11-17 03:46:19 <xelister> chain splits in my wallet?? More likelly then you think.
235 2010-11-17 03:46:29 <ArtForz> xelister: because thats actually slower
236 2010-11-17 03:46:47 <xelister> ArtForz: how come? they are more complex math not working well on gpu?
237 2010-11-17 03:46:49 <Diablo-D3> ArtForz: now thats hardcore
238 2010-11-17 03:48:20 davex__ has quit (Quit: Ex-Chat)
239 2010-11-17 03:50:17 <ArtForz> xelister: initial 3 rounds are identical for all workitems
240 2010-11-17 03:50:42 <ArtForz> cheaper to do em on cpu than to recalc em for each nonce
241 2010-11-17 03:57:29 <xelister> ah right
242 2010-11-17 03:58:10 <xelister> Diablo-D3: your miner is based on jocl? how it is easy to install on linuxes (not just ubuntu) and on windows?
243 2010-11-17 03:58:35 <xelister> because mom's miner appeared to have trouble with dependency on pyopencl package
244 2010-11-17 03:59:11 <Diablo-D3> jocl?
245 2010-11-17 03:59:15 <Diablo-D3> are you fucking kidding me?
246 2010-11-17 03:59:34 <Diablo-D3> triple fuck you
247 2010-11-17 03:59:37 <Diablo-D3> I use lwjgl
248 2010-11-17 04:00:37 <Edogaa> if you lose bitcoins forever
249 2010-11-17 04:00:41 <xelister> so you write kernel in...? opencl?
250 2010-11-17 04:00:42 <Edogaa> And if they are never regenerated :|
251 2010-11-17 04:00:48 <Edogaa> Isn't that bad D:
252 2010-11-17 04:01:11 <ArtForz> nope
253 2010-11-17 04:04:09 <Diablo-D3> xelister: yes
254 2010-11-17 04:04:27 <Diablo-D3> xelister: everything needed is in the zip except a jre
255 2010-11-17 04:05:38 <Edogaa> i am running 0.3.12
256 2010-11-17 04:05:45 <Edogaa> is there a reasoni should upgrade for the time being
257 2010-11-17 04:07:40 <theymos> Edogaa: Someone could send you a special transaction that would "poison" your wallet (any transactions you send would be lost).
258 2010-11-17 04:07:58 <Edogaa> ...
259 2010-11-17 04:08:03 <Edogaa> thats a reason for upgrading
260 2010-11-17 04:08:04 <Edogaa> ._.
261 2010-11-17 04:08:51 <ArtForz> not quite
262 2010-11-17 04:08:52 <theymos> The fix for that was in 0.3.13. A system for safer backups was implemented in 0.3.14, too.
263 2010-11-17 04:09:03 <Edogaa> UPGRADING
264 2010-11-17 04:09:07 <Edogaa> UPGRADING VERY QUICKLY
265 2010-11-17 04:09:30 Maccer has quit (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
266 2010-11-17 04:09:53 <ArtForz> I still feel like the "fix" for 0/unconfirmed TX is wrong
267 2010-11-17 04:10:21 <theymos> What do you think the fix should have been?
268 2010-11-17 04:11:07 <ArtForz> client receiving a tx that has one or more 0/unconf as prev also puts those in its wallet and rebroadcasts em
269 2010-11-17 04:11:08 alystair has quit (Ping timeout: 276 seconds)
270 2010-11-17 04:11:56 <ArtForz> that way any part of a 0/unconf tree isnt lost as long as the receiver is still on the network
271 2010-11-17 04:12:39 <theymos> You'd still have transactions with fee issues that would take a long time to clear. And now there's priority problems.
272 2010-11-17 04:12:47 <Edogaa> so I just use the bin
273 2010-11-17 04:13:07 <ArtForz> yes
274 2010-11-17 04:13:15 <Edogaa> k
275 2010-11-17 04:13:22 <Edogaa> then idont need to compile from source :O
276 2010-11-17 04:13:43 <ArtForz> and the way it's now that tx depending on a lost 0/unconf TX will just sit there in your wallet for-fucking-ever
277 2010-11-17 04:14:41 <theymos> I wonder how easy it would be to fill up someone's HDD by sending them a billion tiny transactions...
278 2010-11-17 04:14:46 Maccer has joined
279 2010-11-17 04:14:58 <Edogaa> https://www.bitcoin4cash.com/ cannot buy US dollars :\
280 2010-11-17 04:16:09 <Edogaa> 9519 KHASH/s o_o
281 2010-11-17 04:16:24 <Edogaa> Go CPU go! D:
282 2010-11-17 04:16:28 <Edogaa> faster! lol
283 2010-11-17 04:17:27 <ArtForz> wow, I'm only throwing like 250x that on TEST for fun
284 2010-11-17 04:17:47 * xelister kicks Edogaa while he's down
285 2010-11-17 04:18:05 <Edogaa> what i do D:
286 2010-11-17 04:18:13 <Edogaa> also, why do i have such high khash atm?
287 2010-11-17 04:18:14 <xelister> rob ArtForz, duh
288 2010-11-17 04:18:51 <Edogaa> I wish I had a core i7 or an AMD hexacore
289 2010-11-17 04:18:53 <Edogaa> :|
290 2010-11-17 04:19:21 <xelister> best cpu gets like 10 Mhash
291 2010-11-17 04:19:30 <xelister> decent GPU gets 160 Mhash
292 2010-11-17 04:19:32 <Edogaa> :O
293 2010-11-17 04:19:41 <xelister> go go radeon
294 2010-11-17 04:20:08 * Edogaa wants a board with 3-4 CPU sockets for a hexacore assuming this thing can utilize all those threads.
295 2010-11-17 04:20:15 <xelister> Edogaa: no, this is stupid
296 2010-11-17 04:20:19 <xelister> you are not paying attention
297 2010-11-17 04:20:27 <Edogaa> :| A gpu is better?
298 2010-11-17 04:20:50 <xelister> gpu that costs 500 USD, will get you 600 Mhash
299 2010-11-17 04:21:02 <Edogaa> But you need a client that supports the GPU?
300 2010-11-17 04:21:10 <xelister> yea
301 2010-11-17 04:21:21 <Edogaa> and doesn't that take away 5 BC's for every single block completed
302 2010-11-17 04:21:21 <ArtForz> yea, m0mchils opencl miner or diablo-d3s
303 2010-11-17 04:21:32 <xelister> no, it takes your virginity
304 2010-11-17 04:21:39 <Edogaa> No seriously
305 2010-11-17 04:21:48 <ArtForz> 4x HD 5970 = about as fast as 150 hexcore phenoms
306 2010-11-17 04:22:11 <xelister> ye seriously. diablo is a big guy, the guy that runs miner for me, couldnt sit down for a week
307 2010-11-17 04:22:14 <Edogaa> :| The only client I heard that does that takes away a lot of Bitcoins from the users, not ven a 'fair amount' per successful BC collection.
308 2010-11-17 04:22:29 <Edogaa> THat kind of makes it useless :\
309 2010-11-17 04:22:32 <Diablo-D3> 4x 5970 == about 9000 watts
310 2010-11-17 04:22:37 <xelister> it does not. also, obviously it is better to get 500-5 = 495 BTC then get 100 BTC in same time
311 2010-11-17 04:22:39 <ArtForz> thats that stupid CUDA miner
312 2010-11-17 04:22:47 <Edogaa> oh?
313 2010-11-17 04:22:56 <Edogaa> So if I get a radeon, I could earn more BCs?
314 2010-11-17 04:22:56 <xelister> Diablo-D3: say what? one eats up to 400 W
315 2010-11-17 04:23:05 <Diablo-D3> xelister: I said about, not over.
316 2010-11-17 04:23:18 <Edogaa> Does a radeon HD 3000 support the necessary requirements?
317 2010-11-17 04:23:21 <Diablo-D3> no
318 2010-11-17 04:23:23 <theymos> Edogaa: The client that took a fee was long ago open-sourced. Nothing charges a fee any more.
319 2010-11-17 04:23:29 <Diablo-D3> Edogaa: 4000 and up only
320 2010-11-17 04:23:33 <xelister> Diablo-D3: 4x 5970 is about 1200 W not about 9000 W
321 2010-11-17 04:23:34 <Edogaa> FUDGE
322 2010-11-17 04:23:38 <Diablo-D3> theymos: can I charge a fee?
323 2010-11-17 04:24:13 <ArtForz> OCed to 900 @ 1.15V 4x5970 is pretty exactly 1600W at the plug
324 2010-11-17 04:24:14 <theymos> Diablo-D3: No one is using your client when it's free... ;)
325 2010-11-17 04:24:41 <xelister> ArtForz: watch out for wires in walls catching fire
326 2010-11-17 04:24:47 <Diablo-D3> theymos: :<
327 2010-11-17 04:24:56 <Diablo-D3> I think I wased my time writing it
328 2010-11-17 04:25:00 <xelister> rms runs it on hurd
329 2010-11-17 04:25:03 <Edogaa> hmm
330 2010-11-17 04:25:07 <xelister> there there Diablo-D3
331 2010-11-17 04:25:09 <xelister> I wil run it.
332 2010-11-17 04:25:09 <Edogaa> THen where can I get the cuda client?
333 2010-11-17 04:25:15 <Edogaa> I have a GTX 260 :/
334 2010-11-17 04:25:17 <xelister> Edogaa: cuda sucks donkey cock
335 2010-11-17 04:25:31 <ArtForz> xelister: european 230V = t3h win
336 2010-11-17 04:25:36 <Edogaa> you know
337 2010-11-17 04:25:49 <Edogaa> the fact that I didn't go with ATI is sort of making me realize what a dumbass I was
338 2010-11-17 04:25:53 <xelister> Edogaa: miners use openCL. they work with both nv and ati
339 2010-11-17 04:25:54 <Edogaa> For believing all the FUD
340 2010-11-17 04:25:59 <Edogaa> :O
341 2010-11-17 04:26:07 <xelister> so it works both card
342 2010-11-17 04:26:09 <Diablo-D3> Edogaa: nvidia has been surviving on FUD from faggot fans for half a decade
343 2010-11-17 04:26:12 <ArtForz> standard circuit is 230V 16A, = 8 5970s and change
344 2010-11-17 04:26:16 <xelister> though with radeon you are x4 faster in same price range
345 2010-11-17 04:26:24 <Diablo-D3> xelister: actually, the real issue isnt money
346 2010-11-17 04:26:27 <Diablo-D3> its watts and heat output
347 2010-11-17 04:26:41 * xelister wants to quote Diablo-D3
348 2010-11-17 04:26:43 <Diablo-D3> 5/6xxx is like 5x faster per watt
349 2010-11-17 04:26:51 <xelister> <Diablo-D3> nvidia has been surviving on FUD from faggot fans for half a decade
350 2010-11-17 04:26:53 <Edogaa> SO
351 2010-11-17 04:26:54 <xelister> hehe
352 2010-11-17 04:26:55 <Edogaa> the client is called
353 2010-11-17 04:26:58 <Edogaa> 'Miners' client?
354 2010-11-17 04:27:00 <Edogaa> or what
355 2010-11-17 04:27:04 <Edogaa> Miners OpeNCL CLIENT?
356 2010-11-17 04:27:07 <Diablo-D3> Edogaa: no
357 2010-11-17 04:27:08 <ArtForz> I have pretty much every socket on it's own circuit anyway
358 2010-11-17 04:27:12 <xelister> diablo client or m0m client
359 2010-11-17 04:27:17 <Diablo-D3> Edogaa: theres two miners, m0's and mine
360 2010-11-17 04:27:22 <xelister> perhaps diablo's one is easier to install, try it
361 2010-11-17 04:27:25 <xelister> since its java
362 2010-11-17 04:27:28 <Diablo-D3> mine doesnt run on nvidia until nvidia fixes their broken drivers
363 2010-11-17 04:27:32 <Diablo-D3> fucking nvidia
364 2010-11-17 04:27:38 <xelister> Diablo-D3: what they did?
365 2010-11-17 04:27:40 * Diablo-D3 hates them so much
366 2010-11-17 04:27:46 <Edogaa> Can I get a link to m0m's client then?
367 2010-11-17 04:27:48 <Diablo-D3> xelister: they STILL dont have opencl 1.1 support
368 2010-11-17 04:27:49 <Edogaa> or do I jsut google
369 2010-11-17 04:27:52 <Edogaa> m0m bitcoins
370 2010-11-17 04:28:06 <xelister> Diablo-D3: prove?
371 2010-11-17 04:28:07 <ArtForz> http://www.bitcoin.org/smf/index.php?topic=1334.0
372 2010-11-17 04:28:07 <Diablo-D3> Edogaa: google poclbm
373 2010-11-17 04:28:15 <xelister> URL/PICS or it didnt happen
374 2010-11-17 04:28:22 <xelister> btw, what a stupid name is that "poclbm"
375 2010-11-17 04:28:26 <Diablo-D3> xelister: run my client, it runs up to the point it calls the callback register
376 2010-11-17 04:28:27 <xelister> sounds like a franch man farting
377 2010-11-17 04:28:36 <Diablo-D3> xelister: which is the only 1.1 function in the whole thing
378 2010-11-17 04:28:38 <ArtForz> Python Open CL Bitcoin Miner ?
379 2010-11-17 04:28:47 <Diablo-D3> ArtForz: yes
380 2010-11-17 04:28:50 <Edogaa> Hmph
381 2010-11-17 04:28:53 <Edogaa> exe
382 2010-11-17 04:28:55 <Edogaa> windows?
383 2010-11-17 04:28:58 <Diablo-D3> python
384 2010-11-17 04:29:01 <xelister> or falling down the stairs.. Paris Ouch *clung* *lub* *BUM* mmmmmom!
385 2010-11-17 04:29:01 <Diablo-D3> though check the forum thread
386 2010-11-17 04:29:05 <Diablo-D3> someone built it with py2exe
387 2010-11-17 04:29:18 <ArtForz> if you want you can compile it yourself from source, the link is right there in the 1st post
388 2010-11-17 04:29:31 <Diablo-D3> "compile"
389 2010-11-17 04:29:32 <Diablo-D3> its python
390 2010-11-17 04:29:35 <Diablo-D3> you interpret it
391 2010-11-17 04:29:40 <theymos> My nvidia driver frequently crashes when Flash asks me whether some site can store a LSO. I'm not sure which company I should hate more.
392 2010-11-17 04:29:41 <xelister> Diablo-D3: then how come m0ms miner works on nvidias
393 2010-11-17 04:29:48 <ArtForz> I mean the client patch
394 2010-11-17 04:29:48 <Diablo-D3> xelister: he doesnt use the event callbacks
395 2010-11-17 04:30:02 <xelister> well... give a switch to use or not use them?
396 2010-11-17 04:30:05 <ArtForz> of course the python miner is python, duh
397 2010-11-17 04:30:07 <xelister> to support your agenta
398 2010-11-17 04:30:11 <xelister> you can always
399 2010-11-17 04:30:17 <Diablo-D3> xelister: his polls the card if its done (with blocking), mine has the card tell my miner its done
400 2010-11-17 04:30:32 <xelister> system.out.println("Using workaround for BROKEN nvidia drivers!!! NVIDIA SUCKS COCKS"); or sometinhg
401 2010-11-17 04:30:34 <Diablo-D3> xelister: with my miner, if I DONT use them.... nothing happens.
402 2010-11-17 04:30:51 <ArtForz> Diablo-D3: behind the scenes it's all the same event system anyways
403 2010-11-17 04:30:53 <Diablo-D3> xelister: as in, no kernel ever gets queued
404 2010-11-17 04:30:59 <xelister> and use a pooling loop instead waiting for notification or soemthing - as I suppose m0m's dones
405 2010-11-17 04:31:00 <Diablo-D3> ArtForz: yes and no
406 2010-11-17 04:31:01 <xelister> *does
407 2010-11-17 04:31:10 <Diablo-D3> ArtForz: my miner wakes up once a second
408 2010-11-17 04:31:26 <Diablo-D3> xelister: almost, m0's use blocking calls
409 2010-11-17 04:31:38 <xelister> what is the end result difference?
410 2010-11-17 04:31:40 <Diablo-D3> ArtForz: like I said, fastest I could get going m0's was 66, mine was 70
411 2010-11-17 04:31:41 <xelister> your uses less host CPU?
412 2010-11-17 04:31:45 <ArtForz> 0
413 2010-11-17 04:31:49 <ArtForz> my miner uses clWaitForEvents
414 2010-11-17 04:31:51 <xelister> or has better hashrate or something?
415 2010-11-17 04:32:04 <xelister> Diablo-D3: you run m0's with -f 1 ?
416 2010-11-17 04:32:07 <Diablo-D3> xelister: cpu usage on 2.2 is fucked no matter what
417 2010-11-17 04:32:11 <Diablo-D3> no. -f 2
418 2010-11-17 04:32:11 <ArtForz> yep
419 2010-11-17 04:32:20 <ArtForz> sdk 2.2 = teh suck
420 2010-11-17 04:32:26 <xelister> why so?
421 2010-11-17 04:32:31 <Diablo-D3> -f 2 gets about 68 on m0s, about 71 on mine
422 2010-11-17 04:32:34 <xelister> lets email them
423 2010-11-17 04:32:43 <xelister> to fix it
424 2010-11-17 04:32:50 <ArtForz> hahaha
425 2010-11-17 04:32:53 <Diablo-D3> both with 2 wide vector and 64 sized workgroups, using 10.9
426 2010-11-17 04:32:54 <xelister> or just #opencl
427 2010-11-17 04:33:08 <Diablo-D3> xelister: I think its due to the possible thread spamming
428 2010-11-17 04:33:12 <Diablo-D3> Im not entirely sure
429 2010-11-17 04:33:20 <Diablo-D3> its surprisingly efficient, it just eats all idle cpu time
430 2010-11-17 04:33:31 <ArtForz> Diablo-D3: try it with multiple GPUs
431 2010-11-17 04:33:39 <Diablo-D3> ArtForz: mine should actually have a total win there
432 2010-11-17 04:33:43 <ArtForz> 2.2 completely keels over
433 2010-11-17 04:33:47 <Diablo-D3> ArtForz: try mine?
434 2010-11-17 04:33:49 <xelister> all IDLE time? and does utilizing CPU slow down or not the 1.2 program
435 2010-11-17 04:33:55 <Diablo-D3> ArtForz: it should happily consume ALL cpu time
436 2010-11-17 04:34:02 <xelister> ALL CPU TIME
437 2010-11-17 04:34:11 <xelister> it will eart your CPU
438 2010-11-17 04:34:27 <xelister> then consume all nearby cupercomputing centers cpus =
439 2010-11-17 04:34:29 <xelister> =)
440 2010-11-17 04:34:30 <Edogaa> the forum thread seems to show only poclbm_py2exe_20101102 and another one called bitcoint-getwork-svn181-win32
441 2010-11-17 04:34:37 <Edogaa> :/ both of them seem to be windows only
442 2010-11-17 04:34:45 <xelister> Edogaa: you're on linux? ubuntu?
443 2010-11-17 04:34:50 <Edogaa> Slackware
444 2010-11-17 04:34:56 <xelister> again?
445 2010-11-17 04:35:02 <xelister> are you the same guy that asked 2 days ago? :)
446 2010-11-17 04:35:03 <Edogaa> Haven't felt like deinstalling
447 2010-11-17 04:35:06 <Edogaa> no
448 2010-11-17 04:35:09 theymos has quit (Remote host closed the connection)
449 2010-11-17 04:35:19 <xelister> it seems slackwarers are heaving harder time with this
450 2010-11-17 04:35:29 <xelister> because pyopencl is not packed for slack
451 2010-11-17 04:35:35 <xelister> if you figure it out, ping me please
452 2010-11-17 04:35:46 <Edogaa> so what I need is pyopencl
453 2010-11-17 04:35:49 <Edogaa> and something else?
454 2010-11-17 04:36:07 <xelister> pyopencl python library, some other EASY INSTALLABLE python libs (via easy_install)
455 2010-11-17 04:36:10 <xelister> and python itself ofcourse
456 2010-11-17 04:36:16 <xelister> get svn and build the bitcoind main program
457 2010-11-17 04:36:24 <xelister> after patching it with patch
458 2010-11-17 04:36:28 <Edogaa> :/
459 2010-11-17 04:36:29 <xelister> and then run the python application
460 2010-11-17 04:37:07 <Edogaa> I'm not in the mood for this. I'm just going to switch to my windows install since it will be a faster install and I can get back to studying instead of doing all of this.
461 2010-11-17 04:37:40 <xelister> Edogaa: please write how it goes with win installer
462 2010-11-17 04:38:01 <Edogaa> Hmm, alright
463 2010-11-17 04:38:09 <Edogaa> see ya in a bit.
464 2010-11-17 04:38:37 Edogaa has quit (Quit: KVIrc 4.0.2 Insomnia http://www.kvirc.net/)
465 2010-11-17 04:39:04 <xelister> PyCUDA and PyOpenCL at Nvidia GTC 2010
466 2010-11-17 04:39:05 <xelister> Like last year, I had the honor of being invited to present PyCUDA and PyOpenCL along with a few examples of their use to a great crowd at Nvidia's inaugural GPU Technology Conference 2010.
467 2010-11-17 04:39:09 <xelister> ^-- author of pyopencl
468 2010-11-17 04:39:15 <xelister> sounds like nvbribefag? ;)
469 2010-11-17 04:39:40 <xelister> NV appears to have better marketing
470 2010-11-17 04:39:55 <ArtForz> yeah, and thats about it
471 2010-11-17 04:40:25 <ArtForz> that and AMD can't fucking get their pile of shit drivers together
472 2010-11-17 04:40:53 <xelister> that is correct
473 2010-11-17 04:41:09 <xelister> since like 5+ years I used ATI
474 2010-11-17 04:41:13 <xelister> they drives always had problems
475 2010-11-17 04:41:16 <xelister> even on windows
476 2010-11-17 04:41:31 <xelister> perhaps this is why they opensourced them, so someone else can fianlly sort it out at least for linux
477 2010-11-17 04:41:49 <ArtForz> well, their windows drivers are crappy in different ways
478 2010-11-17 04:42:06 <ArtForz> aka "how to disable internal crossfire of a 5970 under windows"
479 2010-11-17 04:42:16 <ArtForz> AMD: can't be done
480 2010-11-17 04:42:28 <xelister> NO CAN'T DO
481 2010-11-17 04:42:46 <ArtForz> fun part, install ANY other 5xxx card and you magically get the crossfire config options in CCC
482 2010-11-17 04:42:51 <xelister> btw, is using 5970 causing any problems under linux? single 5970 card
483 2010-11-17 04:43:04 <ArtForz> disable crossfire and - surprise - crossfire is disabled
484 2010-11-17 04:43:08 <ArtForz> nope
485 2010-11-17 04:43:19 <xelister> so it will work like a charm? like 5770?
486 2010-11-17 04:43:25 <ArtForz> well, you have to make sure crossfire is disabled
487 2010-11-17 04:43:27 <xelister> instead higher need for PSU
488 2010-11-17 04:43:34 <xelister> why?
489 2010-11-17 04:43:34 <ArtForz> but thats dead simple with aticonfig
490 2010-11-17 04:43:43 <ArtForz> crosffire and OCL don't mix
491 2010-11-17 04:43:44 <xelister> otherwise it will CROSSFIRE in my face? ;)
492 2010-11-17 04:43:46 <xelister> OCL/
493 2010-11-17 04:43:51 <xelister> ok
494 2010-11-17 04:43:57 <xelister> what is the command for that with aticonfig?
495 2010-11-17 04:44:06 <ArtForz> OCL on any non-master card in a crossfire chain will *work* but return invalid results
496 2010-11-17 04:44:13 <xelister> wonderfull
497 2010-11-17 04:44:23 * xelister claps
498 2010-11-17 04:44:54 <xelister> perhaps at least miners can show they electricity bill as PROVE OF WORK ;) if all results where invalid and got 0 coins lol
499 2010-11-17 04:44:55 <ArtForz> aticonfig --adapter=all --cfd
500 2010-11-17 04:45:11 <ArtForz> deletes all crossfire chains
501 2010-11-17 04:45:31 <xelister> disabling it does NOT cut card performance by ½?
502 2010-11-17 04:45:39 <ArtForz> nope
503 2010-11-17 04:45:45 <xelister> and in 3d games?
504 2010-11-17 04:45:51 <Diablo-D3> [11:36:47] <ArtForz> OCL on any non-master card in a crossfire chain will *work* but return invalid results
505 2010-11-17 04:45:54 <ArtForz> who plays 3d games on linux?
506 2010-11-17 04:45:59 * xelister me
507 2010-11-17 04:46:02 <Diablo-D3> me
508 2010-11-17 04:46:09 <xelister> ArtForz you geek ;)
509 2010-11-17 04:46:22 <xelister> so does it affect normal video rendering speed by half?
510 2010-11-17 04:46:26 <Diablo-D3> thats why Im buying a 6xxx
511 2010-11-17 04:46:31 <ArtForz> doesnt look like it
512 2010-11-17 04:46:31 <xelister> if yes, well then no big problem, just re-enable it
513 2010-11-17 04:46:40 <Diablo-D3> xelister: yes, turning off crossfire turns off acceleration using other cards
514 2010-11-17 04:46:41 <Diablo-D3> BUT
515 2010-11-17 04:46:42 <Diablo-D3> BUT
516 2010-11-17 04:46:45 <Diablo-D3> you dont win anyhow
517 2010-11-17 04:46:50 <ArtForz> but then, half a 5970 is a underclocked 5870 = still a pretty beefy card
518 2010-11-17 04:46:53 <Diablo-D3> on 5xxx, each card is about a 85% increase
519 2010-11-17 04:46:55 <xelister> well I was speaking about single 5970 card
520 2010-11-17 04:46:58 a_meteorite has joined
521 2010-11-17 04:47:14 <xelister> not double
522 2010-11-17 04:47:17 <Diablo-D3> so 2x gives you 185%
523 2010-11-17 04:47:23 <xelister> Diablo-D3: when its available in usa?
524 2010-11-17 04:47:26 <Diablo-D3> 3x gives you about, eh, 2 and a half
525 2010-11-17 04:47:33 <Diablo-D3> geforce its worse
526 2010-11-17 04:47:47 <ArtForz> yep, 6xxx crossfire scaling = awesome
527 2010-11-17 04:47:47 <Diablo-D3> 1x = 1x, 2x = 1.5x, 3x = 2x, 4x = ... 2x
528 2010-11-17 04:48:02 <Diablo-D3> yeah, 6xxx is supposed to have better crossfire performance
529 2010-11-17 04:48:10 <xelister> NLogia ®
530 2010-11-17 04:48:14 <xelister> nlog()
531 2010-11-17 04:48:22 <ArtForz> thats why nv is pushing triSLI, they can't get decent performance out of 4 ;)
532 2010-11-17 04:48:56 <Diablo-D3> sli is stupid though
533 2010-11-17 04:49:14 <Diablo-D3> I'd rather, say, make a whole rendering pipeline in opencl, and use uncrossfired cards
534 2010-11-17 04:49:17 <Diablo-D3> and manage the calcs myself
535 2010-11-17 04:49:30 <ArtForz> you know, I've been wondering if AMD might go for 6 or 8way xfire
536 2010-11-17 04:49:32 <Diablo-D3> but thats insanity
537 2010-11-17 04:49:48 <Diablo-D3> ArtForz: 8 way xfire may end up getting around 5x performance
538 2010-11-17 04:49:54 <ArtForz> yep
539 2010-11-17 04:49:57 <Diablo-D3> unless they can setup a better link
540 2010-11-17 04:50:13 <ArtForz> it's just a REALLY weird coincidence 10.10 got support for > 4 GPUs
541 2010-11-17 04:50:14 <Diablo-D3> the real problem is they dont have a unified task manager
542 2010-11-17 04:50:48 <ArtForz> combine that with 6xxx xfire scaling a lot better than 5xxx ...
543 2010-11-17 04:51:30 <Diablo-D3> huh
544 2010-11-17 04:51:37 * Diablo-D3 jizzed in his pants
545 2010-11-17 04:51:42 <ArtForz> = 6990 with 2 xfire connectors ?
546 2010-11-17 04:51:54 <Diablo-D3> well Im wondering what the big 6990 secret is
547 2010-11-17 04:52:26 <ArtForz> dunno
548 2010-11-17 04:52:32 <Diablo-D3> I wonder if the link between the two is some sort of ultra-intergral link
549 2010-11-17 04:52:40 <Diablo-D3> or some other sort of low latency unified control
550 2010-11-17 04:52:43 <anarchyx> what does OCL mean? overclocking?
551 2010-11-17 04:52:47 <ArtForz> OpenCL
552 2010-11-17 04:52:49 <Diablo-D3> anarchyx: opencl
553 2010-11-17 04:52:52 <anarchyx> oh
554 2010-11-17 04:52:53 <Diablo-D3> anarchyx: see also, ogl
555 2010-11-17 04:53:05 <anarchyx> so if you use crossfire and ocl what happens?
556 2010-11-17 04:53:05 <ArtForz> wild speculation: coherent shared memory link
557 2010-11-17 04:53:13 <Diablo-D3> ArtForz: unlikely
558 2010-11-17 04:53:14 <Diablo-D3> BUT
559 2010-11-17 04:53:21 <Diablo-D3> anarchyx: you only get one gpu working
560 2010-11-17 04:53:27 <Diablo-D3> ArtForz: BUT
561 2010-11-17 04:53:34 <Diablo-D3> ArtForz: it may be a HTX link
562 2010-11-17 04:53:46 <anarchyx> even if you dont have crossfire connectors hooked up to the cards?
563 2010-11-17 04:53:52 <Diablo-D3> see, if the task controller is unified, then you magically get +100% for every gpu
564 2010-11-17 04:54:03 <Diablo-D3> anarchyx: that crossfire connector doesnt do what you think it does
565 2010-11-17 04:54:09 <Diablo-D3> anarchyx: its almost for show
566 2010-11-17 04:54:12 <anarchyx> really?
567 2010-11-17 04:54:19 <anarchyx> wtf
568 2010-11-17 04:54:27 <Diablo-D3> anarchyx: the cards can dma to each other across the main pci-e bus
569 2010-11-17 04:54:38 <ArtForz> yep
570 2010-11-17 04:54:39 <anarchyx> why make it harder then?
571 2010-11-17 04:54:48 <Diablo-D3> anarchyx: its just for synchro across major jobs
572 2010-11-17 04:55:02 <ArtForz> the xfire connector bandwidth is a joke
573 2010-11-17 04:55:12 <Diablo-D3> the connector is a single pci-e link
574 2010-11-17 04:55:15 <ArtForz> yep
575 2010-11-17 04:55:18 <Diablo-D3> its enough to deal with final results
576 2010-11-17 04:55:27 <Diablo-D3> like split frame rendering, and sending the chunks to the main framebuffer
577 2010-11-17 04:55:41 <ArtForz> true
578 2010-11-17 04:55:51 <Diablo-D3> the bandwidth isnt the real issue
579 2010-11-17 04:55:53 <Diablo-D3> its the latency
580 2010-11-17 04:56:23 <Diablo-D3> xfire/sli links arent low latency enough to unify task management
581 2010-11-17 04:56:35 <Diablo-D3> if they were, you could collectively tile render
582 2010-11-17 04:56:47 <Diablo-D3> or split opencl wavefronts
583 2010-11-17 04:56:56 <Diablo-D3> (or rather, split up a collection of them)
584 2010-11-17 04:57:06 <Diablo-D3> with no penalty
585 2010-11-17 04:58:05 <Diablo-D3> collective tile rendering means, basically, 2x cards == 200% speed in all use cases
586 2010-11-17 04:58:18 <Diablo-D3> since the job is split down to the 16x16 tile
587 2010-11-17 04:58:39 <ArtForz> btw, I found *good* cheap PCIe risers
588 2010-11-17 04:58:40 <Diablo-D3> and then card 1 would merge results normally
589 2010-11-17 04:58:44 <ArtForz> http://pc-adapter.net/products/450.html
590 2010-11-17 04:59:16 <Diablo-D3> hah 4 layer pcb
591 2010-11-17 04:59:27 <ArtForz> they even work at 2.0 speeds without problems
592 2010-11-17 04:59:38 <Diablo-D3> I love 4 layer cheap boards you know why?
593 2010-11-17 04:59:51 <Diablo-D3> both outer layers are essentially blank
594 2010-11-17 05:00:04 <ArtForz> depends on the stackup
595 2010-11-17 05:00:14 <Diablo-D3> well yeah, Im implying you dont NEED the layers
596 2010-11-17 05:00:20 <ArtForz> you do
597 2010-11-17 05:00:23 <xelister> ArtForz: what for you need such connector
598 2010-11-17 05:00:28 <Diablo-D3> something like this, you only need the two inner layers
599 2010-11-17 05:00:32 <ArtForz> nope
600 2010-11-17 05:00:36 <Diablo-D3> ArtForz: no?
601 2010-11-17 05:00:41 <Diablo-D3> pci-e is double sided
602 2010-11-17 05:00:44 <ArtForz> you need a ground plane under the PCIe lanes
603 2010-11-17 05:00:56 <Diablo-D3> hrm
604 2010-11-17 05:01:01 <Diablo-D3> should be 6 layer then =/
605 2010-11-17 05:01:11 <Diablo-D3> actually 5
606 2010-11-17 05:01:19 <ArtForz> yep, but 4-layer seems to work for such a short distance
607 2010-11-17 05:01:24 <Diablo-D3> put the the grounds on the outside
608 2010-11-17 05:01:26 <Diablo-D3> :D
609 2010-11-17 05:01:44 <Diablo-D3> seriously though
610 2010-11-17 05:01:51 <Diablo-D3> we're all nuts
611 2010-11-17 05:01:53 <ArtForz> I *think* they put the PCIe lanes on the outside layers and the 2 inside layers are GND and split +3.3/+12
612 2010-11-17 05:02:05 <ArtForz> xelister: quad 5970
613 2010-11-17 05:02:10 <Diablo-D3> ArtForz: yeah, Im obviously no electrical engineer
614 2010-11-17 05:02:36 <ArtForz> well, I know a bit about stuff like that
615 2010-11-17 05:02:44 <Diablo-D3> ArtForz: btw
616 2010-11-17 05:02:47 <Diablo-D3> I just wrote something awesome
617 2010-11-17 05:02:48 <Diablo-D3> a rable
618 2010-11-17 05:02:52 <Diablo-D3> ribbon cable -> rable
619 2010-11-17 05:02:59 <ArtForz> hahaha
620 2010-11-17 05:03:01 <Diablo-D3> cant unsee!
621 2010-11-17 05:03:18 <ArtForz> the construction on these is quite interesting
622 2010-11-17 05:04:04 <ArtForz> looks like they use the ribbon cable with the same spacing as 80-wire ATA
623 2010-11-17 05:04:10 <xelister> ArtForz: but that changes 1 slot into 2 slots? or just expanding so there is physical space to put the big card?
624 2010-11-17 05:04:52 <Diablo-D3> ArtForz: huh
625 2010-11-17 05:05:01 <ArtForz> and thats directly soldered to the board
626 2010-11-17 05:05:01 <Diablo-D3> it doesnt look like one though
627 2010-11-17 05:05:08 davex__ has joined
628 2010-11-17 05:05:14 <Diablo-D3> although its kind of obvious theres an outer sheath
629 2010-11-17 05:05:17 <ArtForz> they glue alu sheets to both sides of the cable
630 2010-11-17 05:05:21 <Diablo-D3> yeah
631 2010-11-17 05:05:26 <ArtForz> pretty stiff ones
632 2010-11-17 05:05:43 <ArtForz> = these are quite inflexible flexible risers
633 2010-11-17 05:05:48 <Diablo-D3> lol
634 2010-11-17 05:05:51 <Diablo-D3> they're bend once ones
635 2010-11-17 05:05:55 <Diablo-D3> 6 inchs too
636 2010-11-17 05:05:56 <Diablo-D3> thats not bad
637 2010-11-17 05:05:58 <ArtForz> yep
638 2010-11-17 05:06:05 <Diablo-D3> thats enough to take the 4th pci-e and move it off the board
639 2010-11-17 05:06:20 <ArtForz> they're perfect for my application
640 2010-11-17 05:06:22 <Diablo-D3> if your case has a 8th, 9th, whatever slot and the mobo doesnt go out that far
641 2010-11-17 05:06:25 <Diablo-D3> yeah rly
642 2010-11-17 05:06:40 <ArtForz> = just raising cards #2 and #4 straight up 4"+
643 2010-11-17 05:06:40 <Diablo-D3> thats perfect for inside servers too if the card mounts properly to the chassis plate
644 2010-11-17 05:06:51 <Diablo-D3> like, 90 degree 1U installs etc
645 2010-11-17 05:06:53 <ArtForz> yep
646 2010-11-17 05:07:09 <xelister> ArtForz: so you need a mainboar with actually 4 PCI-X x16 slots in it?
647 2010-11-17 05:07:15 <Diablo-D3> wanna piss of a rack monkey? hand him a box with risers in it in a short case
648 2010-11-17 05:07:25 <ArtForz> xelister: yep
649 2010-11-17 05:07:32 <Diablo-D3> xelister: no!
650 2010-11-17 05:07:32 <xelister> ArtForz: heh, are there any?
651 2010-11-17 05:07:33 <Diablo-D3> pci-e!
652 2010-11-17 05:07:42 <xelister> yea pci-e x16
653 2010-11-17 05:07:57 <Diablo-D3> pci-x is a form of 64bit pci that exceeds 66mhz
654 2010-11-17 05:08:18 <ArtForz> http://www.legitreviews.com/images/reviews/1004/msi_790fx_gd70_fullboard.jpg
655 2010-11-17 05:08:59 <Diablo-D3> the bizzaro thing is they had 233mhz pci-x
656 2010-11-17 05:09:06 <xelister> wow crazy
657 2010-11-17 05:09:12 <xelister> this mobo
658 2010-11-17 05:09:16 <Diablo-D3> remember that 64bit 33mhz, 32bit 66mhz, and x1 pci-e 1.0 are the same speed
659 2010-11-17 05:09:24 <ArtForz> yep
660 2010-11-17 05:09:27 <xelister> btw, ArtForz, encountered any problems with linuxes as OS for mining?
661 2010-11-17 05:09:32 <ArtForz> nope
662 2010-11-17 05:09:36 <Diablo-D3> wait maybe it was 266
663 2010-11-17 05:09:37 <Diablo-D3> I forget
664 2010-11-17 05:09:39 <ArtForz> 266
665 2010-11-17 05:09:45 <ArtForz> 64-bit 266MHz
666 2010-11-17 05:09:47 <Diablo-D3> yeah, so thats a 8x pci-e slot
667 2010-11-17 05:09:47 <xelister> and multi-card setups work well too?
668 2010-11-17 05:09:55 <ArtForz> xelister: yep
669 2010-11-17 05:09:59 <Diablo-D3> oh and I forgot one
670 2010-11-17 05:10:00 <Diablo-D3> agp 1x
671 2010-11-17 05:10:06 <ArtForz> or a x4 2.0 slot ;)
672 2010-11-17 05:10:32 <Diablo-D3> so 64bit 33mhz == 32bit 66mhz == agp 1x == x1 pci-e 1.0 == half a x1 pci-e 2.0
673 2010-11-17 05:10:41 Animeking has joined
674 2010-11-17 05:10:41 <ArtForz> and afair PCIe 3 should be out soon
675 2010-11-17 05:10:49 <Diablo-D3> yeah its absurd
676 2010-11-17 05:10:53 <Animeking> I wish you guys made it simpler to install then having a 13 page thread :/
677 2010-11-17 05:10:54 <Diablo-D3> you dont need the bandwidth
678 2010-11-17 05:10:57 <Diablo-D3> they do it for the latency
679 2010-11-17 05:10:58 Animeking is now known as Edogaa
680 2010-11-17 05:11:02 <Diablo-D3> BUT
681 2010-11-17 05:11:11 <Diablo-D3> if you need the latency that bad, you really want native htx gear
682 2010-11-17 05:11:19 <Diablo-D3> like, they make native htx 10gbit controllers
683 2010-11-17 05:11:30 <ArtForz> well, it should be useful for "cheap" low-latency interconnects
684 2010-11-17 05:11:37 <Diablo-D3> ArtForz: well
685 2010-11-17 05:11:39 <Diablo-D3> if 3.0 doesnt suck
686 2010-11-17 05:11:46 <Diablo-D3> we could be looking at cheaper IB
687 2010-11-17 05:11:52 <ArtForz> yep, IB RDMA
688 2010-11-17 05:12:12 <Diablo-D3> cheap IB gear usually fucks you in the ass though
689 2010-11-17 05:12:48 <ArtForz> if I see this right PCIe3 should be 8Gbit/s effective per lane
690 2010-11-17 05:13:04 <Diablo-D3> its 4x 1.0
691 2010-11-17 05:13:05 <Diablo-D3> iirc
692 2010-11-17 05:13:10 <ArtForz> well, not quite
693 2010-11-17 05:14:23 <ArtForz> I think 1.0 was 2.5GHz w/ 8b/10b encoding = 2Gbit/s per lane
694 2010-11-17 05:14:46 <ArtForz> 2.0 5GHz w/ 8/10 encoding = 4Gbit/lane
695 2010-11-17 05:15:28 <ArtForz> 3.0 8GHz w/ 128/130 encoding = 7.87Gbit/lane
696 2010-11-17 05:15:39 <Diablo-D3> yeah ignore that encoding shit
697 2010-11-17 05:15:45 <ArtForz> huh?
698 2010-11-17 05:16:05 <xelister> bad move to not make it like 3.0 8.1 GHz
699 2010-11-17 05:16:20 <Diablo-D3> ArtForz: well, look at pci vs pci-e
700 2010-11-17 05:16:23 <xelister> now they cant say in commercials "TWO TIMES FASTER" they have to say like "/almost" two times faster"
701 2010-11-17 05:16:32 <ArtForz> so a x16 3.0 link should be ~ 15.75GB/s full duplex
702 2010-11-17 05:16:35 <Diablo-D3> pci 66mhz doesnt quite == x1 pci 1.0
703 2010-11-17 05:16:50 <ArtForz> yea
704 2010-11-17 05:16:54 <Diablo-D3> but the numbers are so close, who cares
705 2010-11-17 05:17:20 <xelister> the guy that sues for missinformation cares ;)
706 2010-11-17 05:17:38 <ArtForz> thats actually not that bad
707 2010-11-17 05:17:47 <Diablo-D3> its a smidgen of difference
708 2010-11-17 05:17:58 <ArtForz> faster than a 32-bit HT2.0 link
709 2010-11-17 05:18:24 <Diablo-D3> http://en.wikipedia.org/wiki/List_of_device_bandwidths
710 2010-11-17 05:18:28 <Diablo-D3> the best page on wikipedia
711 2010-11-17 05:19:05 <ArtForz> wtf? a PCIe x32 link?
712 2010-11-17 05:19:11 <Diablo-D3> yes they make them
713 2010-11-17 05:19:16 <Diablo-D3> Ive only seen a picture of one
714 2010-11-17 05:19:36 <ArtForz> I dont think thats actually in the PCIe spec
715 2010-11-17 05:19:39 <Diablo-D3> it is
716 2010-11-17 05:19:49 <Diablo-D3> but its not in the general specification known as REALITY
717 2010-11-17 05:19:56 <Edogaa> :/ are there installation instructions, like do I have tio replace the .exe and a few other folders in my current install?
718 2010-11-17 05:19:57 <Edogaa> or?
719 2010-11-17 05:20:07 <Diablo-D3> Edogaa: what are you talking about?
720 2010-11-17 05:20:19 <Edogaa> The m0m client
721 2010-11-17 05:20:51 <Diablo-D3> you get a getworked patched bitcoin, and then you get m0's client
722 2010-11-17 05:21:03 * Edogaa headdesk.
723 2010-11-17 05:21:11 <Diablo-D3> thats all you need
724 2010-11-17 05:21:42 <Diablo-D3> or if you use mine, you still need a getwork patched bitcoin
725 2010-11-17 05:21:50 <ArtForz> I have the spec here somewhere...
726 2010-11-17 05:22:03 <Diablo-D3> ArtForz: well, look at the subtle implication in the design
727 2010-11-17 05:22:21 <Diablo-D3> ArtForz: since each lane is separate, you can have unlimited lanes
728 2010-11-17 05:22:22 <xelister> btw
729 2010-11-17 05:22:25 <ArtForz> yea
730 2010-11-17 05:22:33 <xelister> how to bit bitcoind on linux for windows, anyone?
731 2010-11-17 05:22:39 <Diablo-D3> ArtForz: problem is, there is no system physically big enough for that
732 2010-11-17 05:22:50 <Diablo-D3> ArtForz: you'd use that shit for some sort of unified backplane blade setup
733 2010-11-17 05:22:52 <xelister> there are problems with crosscompiling boost I best
734 2010-11-17 05:22:53 <ArtForz> well, just build a 42 lane link ;)
735 2010-11-17 05:22:54 <xelister> bet
736 2010-11-17 05:23:08 <Diablo-D3> ArtForz: like, ever see normal blades?
737 2010-11-17 05:23:13 <ArtForz> yea
738 2010-11-17 05:23:19 <Diablo-D3> ArtForz: a lot of them use plastic connectors not meant for the purpose
739 2010-11-17 05:23:40 <Diablo-D3> like, some use 3 or 4 pci connectors, and only one is an actual pci bus
740 2010-11-17 05:24:10 <Diablo-D3> so theres nothing stopping you from using pci x32 connections, and not wiring them all up as pci-e
741 2010-11-17 05:24:20 <ArtForz> well, standard connectors are cheap
742 2010-11-17 05:24:24 <Diablo-D3> yeah exactly
743 2010-11-17 05:24:37 <Diablo-D3> and pci-e connectors are nice
744 2010-11-17 05:24:41 <ArtForz> yep
745 2010-11-17 05:25:01 <ArtForz> 1.0mm grid, controlled impedance
746 2010-11-17 05:25:06 <Diablo-D3> so just use a gigantic x32 connector, and use it to boost your penis size... er... sales bonus
747 2010-11-17 05:25:17 <ArtForz> and fully specd for up to 5GHz
748 2010-11-17 05:25:32 <Diablo-D3> well, if you use it for what it was planned for
749 2010-11-17 05:25:50 <Edogaa> poclbm.py is that the client :/
750 2010-11-17 05:25:50 <ArtForz> sure
751 2010-11-17 05:25:50 <Diablo-D3> its still a piece of plastic with zif pins
752 2010-11-17 05:26:00 <Diablo-D3> Edogaa: no, its the miner
753 2010-11-17 05:26:10 <Diablo-D3> Edogaa: bitcoin is the client
754 2010-11-17 05:26:13 <ArtForz> yeah, but a piece of plastic specified for high frequency stuff
755 2010-11-17 05:26:16 <Edogaa> Okay
756 2010-11-17 05:26:32 <ArtForz> which is quite useful even if it's not PCIe you're running over it
757 2010-11-17 05:26:37 <ArtForz> see HTX slot ;)
758 2010-11-17 05:26:41 <Diablo-D3> yeah
759 2010-11-17 05:26:42 <Edogaa> I thought by 'client' it be a completely new client
760 2010-11-17 05:26:51 <Diablo-D3> Edogaa: no, its just a getwork patched bitcoin
761 2010-11-17 05:27:03 <Diablo-D3> Edogaa: without the getwork patch, the miner cant talk to bitcoin
762 2010-11-17 05:28:14 <Edogaa> with the getwork patch all I need to do is download it, un7zip it then throw it into my installation directory where the client is and repalce the .exe and two dll files
763 2010-11-17 05:28:40 <Diablo-D3> Edogaa: no
764 2010-11-17 05:28:49 <Diablo-D3> Edogaa: the word "patch" refers to an actual patch
765 2010-11-17 05:29:02 <xelister> you need to compile it
766 2010-11-17 05:29:06 <Diablo-D3> m0 has builds already prepatched, see the appropriate thread
767 2010-11-17 05:29:10 <xelister> probably. isnt there ready exe
768 2010-11-17 05:29:11 <Diablo-D3> for windows, anyways
769 2010-11-17 05:29:13 <xelister> yea
770 2010-11-17 05:33:02 anarchyx has quit (Ping timeout: 265 seconds)
771 2010-11-17 05:35:13 Maccer has quit (Ping timeout: 240 seconds)
772 2010-11-17 05:35:31 Maccer has joined
773 2010-11-17 05:37:36 <Edogaa> alright
774 2010-11-17 05:39:53 <Edogaa> Okay, appropriate thread is this, https://www.bitcoin.org/smf/index.php?topic=1334.0 ??? If so, I got the bitcoin-getwork-svn181-win32 which I assume, only assume is a patched .exe client and two dlls which is what is in the zip or I assume that is what it is after that assuming I got the irght idea I don't know what to do next :/
775 2010-11-17 05:40:08 <Edogaa> Or what to do with the miner
776 2010-11-17 05:40:22 <Diablo-D3> Edogaa: you run the bitcoin in that zip, then you start the miner
777 2010-11-17 05:40:31 <Edogaa> ok
778 2010-11-17 05:40:32 <Diablo-D3> theres py2exe builds of the miner too
779 2010-11-17 05:40:39 <Diablo-D3> since you windows users are too dumb to use python
780 2010-11-17 05:40:43 <Edogaa> -_-
781 2010-11-17 05:41:09 <Diablo-D3> well, lets face it, windows has sucked all your brains out
782 2010-11-17 05:41:13 <Edogaa> -_-
783 2010-11-17 05:41:14 <Diablo-D3> otherwise you'd just switch to linux already
784 2010-11-17 05:41:17 <Edogaa> -_-
785 2010-11-17 05:41:29 <Edogaa> ignored
786 2010-11-17 05:42:06 <Diablo-D3> what was that Edogaa? did you say you wanted to be banned?
787 2010-11-17 05:45:30 <nanotube> Diablo-D3: what was that, did you say you have chanop? :P
788 2010-11-17 05:46:12 <Diablo-D3> nanotube: I bet I could pay bitcoins for bans.
789 2010-11-17 05:47:08 <nanotube> i bet otherwise - i think it's unlikely the current chanops would take bribes for unfair bans. that would ruin the channel.
790 2010-11-17 05:48:21 <nanotube> of course, there's a price for almost anything... :) so with a sufficiently large bribe, anything's possible.
791 2010-11-17 05:48:37 <nanotube> but it'd probably take more coins than you have. :)
792 2010-11-17 05:49:43 * Diablo-D3 looks in his wallet
793 2010-11-17 05:49:48 * Diablo-D3 sigh
794 2010-11-17 05:49:52 <eureka^> If you go into the first or second rooms, you could search around for a hour and still find nothing.
795 2010-11-17 05:49:53 <eureka^> If you go to first or second room, you still can find the time to search for Rino Amane.
796 2010-11-17 05:49:53 <eureka^> Even if you go first or second room, then, Rino can find the time to find the circumference.
797 2010-11-17 05:49:54 <Diablo-D3> I have not generated one block with my miner
798 2010-11-17 05:49:59 <Diablo-D3> I wonder if it even works
799 2010-11-17 05:50:25 <nanotube> hm, what's your 50% time to generation? and how long have you been generating?
800 2010-11-17 05:51:27 <Diablo-D3> something like 3 days
801 2010-11-17 05:51:49 <Diablo-D3> 95% is 10 days 9 hours x_x
802 2010-11-17 05:55:39 <nanotube> hm, so it's still inconclusive, then...
803 2010-11-17 05:56:38 <Diablo-D3> yes
804 2010-11-17 05:56:40 <Diablo-D3> Ive done maybe
805 2010-11-17 05:56:42 <Diablo-D3> 4 days?
806 2010-11-17 05:56:50 <Diablo-D3> I dont keep the machine running 24/7
807 2010-11-17 06:01:04 <davex__> do you make 1 coin per block?
808 2010-11-17 06:02:03 Kiba has quit (Ping timeout: 272 seconds)
809 2010-11-17 06:03:14 <xelister> Diablo-D3: whats your speed
810 2010-11-17 06:04:41 <xelister> if like 70M then you need arpund 6 days +1(verif) so 2 more
811 2010-11-17 06:04:45 <xelister> (diff will change soon)
812 2010-11-17 06:05:18 <xelister> Diablo-D3: guys on #opencl as about the bug in openCL you mentioned
813 2010-11-17 06:05:26 <xelister> davex__: 50btc per block
814 2010-11-17 06:06:00 <xelister> I will use Diablo's miner...... <Diablo-D3> I have not generated one block with my miner I wonder if it even works ..... or perhaps not :D
815 2010-11-17 06:06:43 <ArtForz> well, thats quite easy to test
816 2010-11-17 06:07:23 <ArtForz> just try on test
817 2010-11-17 06:10:52 <nanotube> davex__: currently 50 coins per generated block.
818 2010-11-17 06:11:19 <nanotube> ah xelister has already fielded that one
819 2010-11-17 06:11:21 <nanotube> nvm :)
820 2010-11-17 06:11:36 <xelister> Edogaa: well have to agree with Diablo-D3
821 2010-11-17 06:11:47 <Edogaa> hmm?
822 2010-11-17 06:12:16 <xelister> not using linux (or other good alternative do default crappy windows), on own home computer, often can be a sign of being stupid about computers
823 2010-11-17 06:12:41 <Edogaa> What arrogance you portray
824 2010-11-17 06:12:43 <Edogaa> that is all it is
825 2010-11-17 06:13:22 <xelister> Edogaa: for example using bitcoin, building it, mining, appears easier here on linux
826 2010-11-17 06:13:42 <xelister> as do most open-source related, and programing/development tasks
827 2010-11-17 06:13:59 <davex__> nah
828 2010-11-17 06:14:18 <nanotube> xelister: there can be valid reasons to use windows, though... so no need to be so insulting about it. :) (gaming, win-only software like autocad and crap, support for various specific hardware, etc.)
829 2010-11-17 06:14:25 <ArtForz> yea
830 2010-11-17 06:14:42 <xelister> nanotube: thats why I said about /home/ computer :)
831 2010-11-17 06:14:46 <ArtForz> but for open source dev work windows plain sucks
832 2010-11-17 06:14:57 <xelister> ok some people do 3d studio max from home but thats other case
833 2010-11-17 06:16:12 <ArtForz> and yeah, hardware with win-only drivers = teh fail
834 2010-11-17 06:16:33 <xelister> fortunatelly I somehow avoid this problem
835 2010-11-17 06:16:52 <Diablo-D3> xelister: its around 60mhash now
836 2010-11-17 06:16:55 <Edogaa> 3DS Max, Adobe, games, and linux does have a lot of open source tools as well, even GCC has been ported as mingw32. Although it uses a different set of windows specific libs if I recall. Regardless of what you have to say Linux is not superior to windows, anyone who says that is just a freetard, or a linux-fanatic, likely the latter. -_-
837 2010-11-17 06:16:58 <Diablo-D3> xelister: due to 10.10 sucking
838 2010-11-17 06:17:02 <Diablo-D3> well not sucking sucking
839 2010-11-17 06:17:08 <Diablo-D3> they fixed a bug which had unforunate implications
840 2010-11-17 06:17:22 <ArtForz> Edogaa: install a dev env to build a C python extension
841 2010-11-17 06:17:44 <Diablo-D3> why doesnt he just use the pyopencl binary for windows?
842 2010-11-17 06:17:49 <Diablo-D3> Edogaa: btw, mingw sucks
843 2010-11-17 06:18:02 <ArtForz> on debian/ubuntu: apt-get install python-dev
844 2010-11-17 06:18:22 <Diablo-D3> on debian/ubuntu: apt-get install python-opencl or whatever it is
845 2010-11-17 06:18:34 <ArtForz> on windows: grab a few packaghes from various sites, get EXACTLY this version of the MSVC compiler, ...
846 2010-11-17 06:19:12 <Edogaa> https://www.bitcoin.org/smf/index.php?topic=1334.0 if you are talking about, poclbm_py2exe I have no clue what to do with that now that its downloaded. There aren't even any instructions on the page, most of it is guesswork not even a read me page.
847 2010-11-17 06:19:35 * Diablo-D3 just facepalms
848 2010-11-17 06:19:39 <Diablo-D3> Edogaa is hopeless
849 2010-11-17 06:19:50 <Edogaa> What
850 2010-11-17 06:19:57 <Edogaa> that you aren't very informative :P
851 2010-11-17 06:20:33 <xelister> Diablo-D3: 10.10 ?
852 2010-11-17 06:21:05 * xelister beats the crap out of Edogaa with a 2x8
853 2010-11-17 06:21:15 <Diablo-D3> xelister: yes, 10.9 worked far better
854 2010-11-17 06:21:24 <xelister> 10.10 / 10.9 - what?
855 2010-11-17 06:21:37 <ArtForz> yes, ubuntu 10.10
856 2010-11-17 06:21:42 <ArtForz> and debian testing
857 2010-11-17 06:22:02 <ArtForz> and debian unstable, and probably also ubuntu 10.4, can't be bothered to check
858 2010-11-17 06:22:09 <Diablo-D3> no, cat
859 2010-11-17 06:22:47 <xelister> Edogaa: I think there is a reason for most developers, most adv. users, and most good miners here are on linux not windowscrap ;)
860 2010-11-17 06:26:22 <nanotube> xelister: if you work with say, autocad, it's not unreasonable to want a copy on your home computer. :) but anyway, gaming is as home as it comes. :P
861 2010-11-17 06:27:04 <nanotube> xelister: i'm a big fan of linux myself, use it on my box... but no need to dis people, that's all i'm saying.
862 2010-11-17 06:31:51 <ArtForz> duh, not all windows users are stupid
863 2010-11-17 06:32:33 <ArtForz> but nearly all stupid users are running windows
864 2010-11-17 06:32:42 <Edogaa> Ubuntu
865 2010-11-17 06:32:44 <Edogaa> 'nuff said
866 2010-11-17 06:32:49 <ArtForz> what?
867 2010-11-17 06:32:59 <Edogaa> a lot of 'stupid users' run linux too
868 2010-11-17 06:33:10 <ArtForz> really, not that many
869 2010-11-17 06:33:27 <Edogaa> and don't say I am wrong, I've seen a lot of stuff said by many of these users, by a LOT of users that obviouslty have no clue what they're talking about.
870 2010-11-17 06:33:36 <ArtForz> sure
871 2010-11-17 06:34:11 <Edogaa> I'd like to think the majority of linux users aren't that bad yet though :/
872 2010-11-17 06:34:17 <ArtForz> yep
873 2010-11-17 06:34:33 <Edogaa> But saying, 'not that many' is quite an understatement...
874 2010-11-17 06:34:40 <ArtForz> no
875 2010-11-17 06:35:07 <davex__> so is there a way to tell bitcoin to automatically update a backed-up wallet every time a transaction is processed?
876 2010-11-17 06:35:21 <ArtForz> davex__: nope
877 2010-11-17 06:35:37 <ArtForz> but with keypool it's not that much of a problem anymore
878 2010-11-17 06:36:02 <davex__> will check thato ut
879 2010-11-17 06:36:38 <ArtForz> = backing up every 50 or so transactions is plenty
880 2010-11-17 06:37:36 <ArtForz> keypool always keeps 100 keypairs around, so you need to do >100 generations/sends after the last backup to actually use keypairs not in that backup
881 2010-11-17 06:37:39 <Edogaa> lets see, I have python 2.6.5, Microsoft VIsual Studios C++ 2008, no numpty and scipy yet though :/ and the opencl gpu driver and opencl sdk doesn't seem to be available for normal none-developer users :/
882 2010-11-17 06:39:10 <ArtForz> huh?
883 2010-11-17 06:39:59 <ArtForz> doesnt cuda 3.2 already have ocl support?
884 2010-11-17 06:40:13 <Edogaa> i do not know.
885 2010-11-17 06:40:28 <Edogaa> http://wiki.tiker.net/PyOpenCL/Installation/Windows i am just reading this
886 2010-11-17 06:44:10 <davex__> ArtForz: It looks like the keypool is also stored locally?
887 2010-11-17 06:44:28 <ArtForz> huh?
888 2010-11-17 06:44:38 <ArtForz> keypool is stored in wallet
889 2010-11-17 06:45:25 <davex__> ArtForz, i'm wondering what happens if you have an old wallet on a flash drive, and your current wallet has a few transactions on it, but gets wiped out somehow
890 2010-11-17 06:46:02 <davex__> ArtForz, how can those transactions get reconstructed if the keypool is also stored locally (in the current wallet)
891 2010-11-17 06:46:13 <ArtForz> the keypool is a FIFO
892 2010-11-17 06:46:20 <ArtForz> = oldest keys are used first
893 2010-11-17 06:47:13 <ArtForz> so the "random" keys used in the new transactions (as long as its less than 100) are already in the backup
894 2010-11-17 06:47:45 <ArtForz> plop that backup wallet dat in a fresh datadir, client downlaods block chain and sees "it's" tranascvtions ebcause they match the keys
895 2010-11-17 06:47:48 <davex__> oh... so with a known key i can go on the network and get the transaction info?
896 2010-11-17 06:48:06 <davex__> ah, that's what the block is for.
897 2010-11-17 06:48:48 <ArtForz> well, not directly, but yeah
898 2010-11-17 06:49:49 <davex__> hmm... if someone figures out the keys in your keypool somehow, couldn't they enter malicious transactions?
899 2010-11-17 06:49:53 <ArtForz> yes
900 2010-11-17 06:50:09 <ArtForz> thats why they're in the wallet
901 2010-11-17 06:50:13 <nanotube> davex__: that's why you keep them safe :)
902 2010-11-17 06:50:24 <ArtForz> if someone has access to your wallet, you're fucked
903 2010-11-17 06:50:32 <davex__> ok i see.
904 2010-11-17 06:52:59 <xelister> btw
905 2010-11-17 06:53:21 <xelister> is there a know which or how many bitcoins where generated by given user
906 2010-11-17 06:53:34 <xelister> like, how to find out how many coins did ArtForz produce
907 2010-11-17 06:53:48 <xelister> what I need to know, some public key of the node, tht remains unchanged?
908 2010-11-17 06:53:48 <ArtForz> nope
909 2010-11-17 06:53:51 <nanotube> xelister: only if you know all his keys.
910 2010-11-17 06:54:00 <Edogaa> great now my HDD is full of junk files...
911 2010-11-17 06:54:06 <xelister> ok for user with just 1 normal generator
912 2010-11-17 06:54:14 <xelister> what I need to know? his... ?
913 2010-11-17 06:54:19 <ArtForz> nope
914 2010-11-17 06:54:38 <xelister> for user that is not trying to hide this information - how to? ArtForz?
915 2010-11-17 06:54:43 <ArtForz> each generation uses a new "random" key from the keypool
916 2010-11-17 06:55:06 <xelister> ok how to track how many keys are generated by my friend that will run my generator?
917 2010-11-17 06:55:22 <ArtForz> you really can't
918 2010-11-17 06:55:24 <xelister> in a way that can not be forged though s/friend/buddy
919 2010-11-17 06:55:38 <ArtForz> see the pooled mining htread for that one
920 2010-11-17 06:55:41 <xelister> well if the generator would auto-transfer all to me, then I would know
921 2010-11-17 06:56:06 <xelister> except, he can buy bitcoins and transfer that. And I would like to know just how many where actually generated by him
922 2010-11-17 06:56:39 <xelister> the generator could just signal me, but such information is unverificable, like message "I generated 50 BTC" is not a prove
923 2010-11-17 06:57:31 <ArtForz> shrug
924 2010-11-17 06:57:43 <xelister> perhaps generator could send me like some private key that togeather with public avaialble information is a prove that no more BTCs where actually /generated/ by given person
925 2010-11-17 06:57:43 <ArtForz> really, read the pooled mining thread
926 2010-11-17 06:58:04 <ArtForz> afair they pretty much solved it
927 2010-11-17 07:02:15 <xelister> ArtForz: that is other topic that I had
928 2010-11-17 07:02:21 <xelister> if you ment http://www.bitcoin.org/smf/index.php?topic=1458.20
929 2010-11-17 07:02:40 <xelister> the question is not the prove that client is actually generating hashes and tries to solve something
930 2010-11-17 07:03:15 <ArtForz> yea, that one
931 2010-11-17 07:03:47 <xelister> instead: in mode where slaves will send all generated BTC to central server, how the server can check if the send BTCs where actually generated by them - or "just" bought by them. The generating program can be customized so it can do something to help such verification
932 2010-11-17 07:03:59 <ArtForz> thats easy
933 2010-11-17 07:04:03 <xelister> :]
934 2010-11-17 07:04:27 <xelister> so, how?
935 2010-11-17 07:04:30 <Diablo-D3> http://dagobah.net/flash/know_your_destiny.swf
936 2010-11-17 07:04:32 <Diablo-D3> this is just absurd
937 2010-11-17 07:04:44 <Diablo-D3> Im pretty sure they've improved it
938 2010-11-17 07:04:54 <xelister> I thought os just adding up incoming transactions from given address, but that allows also to send buyed coins
939 2010-11-17 07:05:17 <ArtForz> sending generated coins means those coins havent been spent before, so the inputs are all coinbases
940 2010-11-17 07:06:07 <ArtForz> bought coins have been sent before, so the input(s) will be a normal transaction
941 2010-11-17 07:06:11 <xelister> ad how exactly can I see what actuall coins where sent to me, and if they where used before? I use mostly that simple RPC client (wget)
942 2010-11-17 07:06:24 <ArtForz> hrrrmmm, not sure
943 2010-11-17 07:06:36 <xelister> or using any online tool for bc
944 2010-11-17 07:06:56 <ArtForz> I know the gui can do it
945 2010-11-17 07:07:11 <ArtForz> hrrrm, BBE or similar should also work
946 2010-11-17 07:08:30 <xelister> bbe?
947 2010-11-17 07:09:06 <ArtForz> http://theymos.ath.cx:64150/bbe/
948 2010-11-17 07:09:51 <xelister> hm if you would send me now 1 BTC for example, then how we check?
949 2010-11-17 07:10:09 <joe_1> send me 1 BTC and i'll tell you if it was generated by you or was bought
950 2010-11-17 07:10:17 <ArtForz> well, in gui it's easy
951 2010-11-17 07:10:24 <xelister> 16b2Hzi3akNSdgmuAm7QvfKBMDd68oBuAD
952 2010-11-17 07:10:31 <ArtForz> run with -debug, double click transaction to get tx hash
953 2010-11-17 07:11:01 <xelister> joe_1: Im rather interested about knowing how to check that exactly
954 2010-11-17 07:11:33 <joe_1> ok, i dont know how to do it, but you look at the block chain log file
955 2010-11-17 07:11:35 <xelister> overall, I will need to get statistics of who send me how much BTC overall per sender, and how much generated - per each sender
956 2010-11-17 07:11:46 <joe_1> the log file will tell you which block the coin you just got came from
957 2010-11-17 07:12:30 <ArtForz> here we have a TX from 2 coinbases: http://theymos.ath.cx:64150/bbe/tx/d57e7559f870df88e82dfec81d9570d66e959dd29c0d85adc1e9060f12502326
958 2010-11-17 07:13:05 <joe_1> xelister: the people sending you money, are they other natural people or other computers you control?
959 2010-11-17 07:13:37 <ArtForz> check the previous output links, both diectly lead to coinbase TXes
960 2010-11-17 07:14:08 <xelister> joe_1: people
961 2010-11-17 07:14:22 <xelister> we have sort of bet about who can actually generate the most coins in given time
962 2010-11-17 07:15:04 <joe_1> and to prove what they generated, they have to pass them through you?
963 2010-11-17 07:15:26 <xelister> yes
964 2010-11-17 07:15:37 <xelister> well I will probably change them to money at some point
965 2010-11-17 07:16:01 <xelister> what do you think about BTCs value in around 2011.03 and .04, what is likelly? more then 0.30?
966 2010-11-17 07:19:40 joe_1 has quit (Ping timeout: 265 seconds)
967 2010-11-17 07:19:47 <ArtForz> well, if they send you all the coins anyways, why not have them send you a copy of their wallet?
968 2010-11-17 07:20:49 <xelister> so the wallet will store some private information, unique only to the generator of given bitcoins?
969 2010-11-17 07:21:03 <ArtForz> yep
970 2010-11-17 07:21:06 <xelister> that is not ever let go to anyone even after coin was transfered away
971 2010-11-17 07:21:09 <xelister> what is that information?
972 2010-11-17 07:21:25 <ArtForz> the private key of the coinbase
973 2010-11-17 07:21:54 joe_1 has joined
974 2010-11-17 07:32:16 a_meteorite has quit (Quit: a_meteorite)
975 2010-11-17 08:14:17 alystair has joined
976 2010-11-17 08:40:00 XxMalouinxX has quit (Read error: Connection reset by peer)
977 2010-11-17 08:40:01 OneFixt has quit (Read error: Connection reset by peer)
978 2010-11-17 08:40:08 XxMalouinxX has joined
979 2010-11-17 08:40:09 OneFixt has joined
980 2010-11-17 08:40:35 OneFixt is now known as Guest21087
981 2010-11-17 08:42:51 XxMalouinxX has quit (Read error: Connection reset by peer)
982 2010-11-17 08:42:51 LobsterMan has quit (Read error: Connection reset by peer)
983 2010-11-17 08:42:51 XxMalouinxX has joined
984 2010-11-17 08:42:56 tg has quit (Read error: Connection reset by peer)
985 2010-11-17 08:43:01 LobsterMan has joined
986 2010-11-17 08:43:02 LobsterMan has quit (Changing host)
987 2010-11-17 08:43:02 LobsterMan has joined
988 2010-11-17 08:43:17 tg has joined
989 2010-11-17 08:51:08 dwdollar has quit (Ping timeout: 250 seconds)
990 2010-11-17 08:52:17 dwdollar has joined
991 2010-11-17 08:53:42 stearns_ has joined
992 2010-11-17 08:54:35 stearns_ has quit (Client Quit)
993 2010-11-17 08:59:30 stearns_ has joined
994 2010-11-17 09:05:14 Teppy2 has quit (Ping timeout: 252 seconds)
995 2010-11-17 09:06:02 Teppy has quit (Ping timeout: 240 seconds)
996 2010-11-17 09:09:09 alystair has quit (Ping timeout: 264 seconds)
997 2010-11-17 09:21:19 Teppy has joined
998 2010-11-17 09:22:26 joe_1 has quit (Ping timeout: 265 seconds)
999 2010-11-17 09:37:40 XxMalouinxX has quit (Read error: Connection reset by peer)
1000 2010-11-17 09:37:44 XxMalouinxX has joined
1001 2010-11-17 09:42:46 XxMalouinxX has quit (Read error: Connection reset by peer)
1002 2010-11-17 09:42:46 XxMalouinxX has joined
1003 2010-11-17 09:57:37 altamic has joined
1004 2010-11-17 10:09:00 altamic has quit (Quit: altamic)
1005 2010-11-17 10:16:23 a_meteorite has joined
1006 2010-11-17 10:16:25 a_meteorite has quit (Excess Flood)
1007 2010-11-17 10:18:41 altamic has joined
1008 2010-11-17 10:22:11 Akiraa has quit (Quit: Daisy, Daisy, give me your answer, do...)
1009 2010-11-17 10:25:28 a_meteorite has joined
1010 2010-11-17 10:38:42 stearns_ has quit (Ping timeout: 276 seconds)
1011 2010-11-17 11:01:18 Teppy2 has joined
1012 2010-11-17 11:01:40 stearns_ has joined
1013 2010-11-17 11:51:29 noagendamarket has quit (Read error: Connection reset by peer)
1014 2010-11-17 12:10:41 <UukGoblin> right, got m0mchil's thing to work on one 5770
1015 2010-11-17 12:10:45 soultcer has quit (Ping timeout: 260 seconds)
1016 2010-11-17 12:10:50 <UukGoblin> diablo's still showing function unimplemented
1017 2010-11-17 12:11:19 <Diablo-D3> UukGoblin: on 2.1?
1018 2010-11-17 12:11:23 <Diablo-D3> you cant use 2.1 with mine
1019 2010-11-17 12:11:39 <UukGoblin> can't get it to work on 2 5770s - tried with and without crossfire, on one xorg server, tried on two, but it only works on the currently active one
1020 2010-11-17 12:11:42 <UukGoblin> Diablo-D3, yeah, 2.1
1021 2010-11-17 12:11:50 <Diablo-D3> 2.1 doesnt provide opencl 1.1
1022 2010-11-17 12:11:53 <UukGoblin> right
1023 2010-11-17 12:12:12 <Diablo-D3> btw, to get it working on 2 5770's with m0s
1024 2010-11-17 12:12:21 <Diablo-D3> turn crossfire off, select two different devices
1025 2010-11-17 12:12:28 <Diablo-D3> as in, run it twice
1026 2010-11-17 12:12:31 <UukGoblin> m0's only showing CPU and one Juniper
1027 2010-11-17 12:12:38 <Diablo-D3> orly?
1028 2010-11-17 12:12:40 <UukGoblin> yeah
1029 2010-11-17 12:12:41 <Diablo-D3> I wonder why
1030 2010-11-17 12:12:52 <Diablo-D3> sure you turned crossfire off?
1031 2010-11-17 12:12:52 <UukGoblin> even though the X desktop spans across two cards (I think)
1032 2010-11-17 12:13:06 <UukGoblin> I disconnected the crossfire link physically so yes
1033 2010-11-17 12:13:09 <Diablo-D3> no
1034 2010-11-17 12:13:16 <Diablo-D3> the crossfire link has no control over it being on or off
1035 2010-11-17 12:13:21 <UukGoblin> orly
1036 2010-11-17 12:13:29 <Diablo-D3> its just an extra pci-e lane thats routed between the two
1037 2010-11-17 12:13:40 <Diablo-D3> they can still DMA directly to each other, or bounce through system ram
1038 2010-11-17 12:13:45 <UukGoblin> well I ran two Xorg servers each on different card and could switch between them with ctrl-alt-f7 and -f8
1039 2010-11-17 12:13:46 <Diablo-D3> use aticonfig to turn it off
1040 2010-11-17 12:13:50 <Diablo-D3> UukGoblin: doesnt matter
1041 2010-11-17 12:13:59 <UukGoblin> right
1042 2010-11-17 12:14:06 <Diablo-D3> I can run two X servers too
1043 2010-11-17 12:14:08 <Diablo-D3> and I have one card
1044 2010-11-17 12:14:25 <UukGoblin> well each monitor was plugged to a different card... let me try though
1045 2010-11-17 12:15:09 <Diablo-D3> aticonfig --cf=off should do it I think
1046 2010-11-17 12:15:16 <UukGoblin> cool thanks trying it now
1047 2010-11-17 12:15:21 <Diablo-D3> I dont have anything to crossfire so I cant check
1048 2010-11-17 12:15:23 <UukGoblin> will take a while cause I have to re-plug it
1049 2010-11-17 12:16:24 <Diablo-D3> UukGoblin: I wish they'd replace the crossfire link with something else
1050 2010-11-17 12:16:33 <Diablo-D3> you missed the earlier discussion
1051 2010-11-17 12:16:56 <Diablo-D3> to get the most out of crossfire, you'd need a unified event and task manager
1052 2010-11-17 12:17:11 <Diablo-D3> so you could, say, scatter/gather tile renders
1053 2010-11-17 12:17:24 <Diablo-D3> or the same with wavefronts in opencl
1054 2010-11-17 12:17:34 <Diablo-D3> but the latency would be killer
1055 2010-11-17 12:18:16 <Diablo-D3> you COULD do it on one of the dual chip cards
1056 2010-11-17 12:18:23 <Diablo-D3> but AMD hasnt yet
1057 2010-11-17 12:19:55 <Diablo-D3> but if they did, that'd be, say, 400% speed with 4 cards in any game or any workload
1058 2010-11-17 12:20:02 <Diablo-D3> or 800% with 8 chips
1059 2010-11-17 12:21:57 <ArtForz> you need to have all cards on one x server
1060 2010-11-17 12:22:28 <ArtForz> the xorg.conf created by aticonfig --adapter=all --initial -f is a good starting point
1061 2010-11-17 12:29:50 <UukGoblin> ArtForz, yup, did that, thanks
1062 2010-11-17 12:29:57 <UukGoblin> gimme few more mins ;-)
1063 2010-11-17 12:41:14 <UukGoblin> yeah, crossfire is definitely off, and Xorg is running on two cards, and m0's thing only sees one Juniper device
1064 2010-11-17 12:41:43 <ArtForz> whats your DISPLAY set to?
1065 2010-11-17 12:41:52 <UukGoblin> :0.0
1066 2010-11-17 12:41:57 <ArtForz> try :0
1067 2010-11-17 12:41:59 <UukGoblin> but it's just one Xorg
1068 2010-11-17 12:42:01 XxMalouinxX has quit (Changing host)
1069 2010-11-17 12:42:01 XxMalouinxX has joined
1070 2010-11-17 12:42:32 <ArtForz> yes, :0.0 = first xserver, first screen/device/whatever
1071 2010-11-17 12:42:45 <ArtForz> :0.1 = first xserver, second screen
1072 2010-11-17 12:42:57 <ArtForz> :0 = first xserver, all screens
1073 2010-11-17 12:43:37 <UukGoblin> ah, good point
1074 2010-11-17 12:43:51 <UukGoblin> yeah that's it
1075 2010-11-17 12:43:56 <UukGoblin> thanks a lot guys :-D
1076 2010-11-17 12:44:13 <UukGoblin> now ph34r the powah of my two 5770s ;-)
1077 2010-11-17 12:44:23 <UukGoblin> difficulty will probably go up by about 10 because of that ;-)
1078 2010-11-17 12:45:34 <ArtForz> *shrug*
1079 2010-11-17 12:46:29 <ArtForz> I'm running a 5770 on TEST for fun
1080 2010-11-17 12:48:22 <UukGoblin> :->
1081 2010-11-17 12:48:30 <UukGoblin> you think test bitcoins will go up in value? ;-]
1082 2010-11-17 12:48:35 <ArtForz> nope
1083 2010-11-17 12:49:02 <ArtForz> just want to get difficulty to a sensible level so I can run a few TX flood tests
1084 2010-11-17 12:49:35 <UukGoblin> getting 155Mhash from each (at -f 10), both cards are at 70 degrees C :-/
1085 2010-11-17 12:49:41 <ArtForz> a TX flood at < 10sec/block is kinda hard to pull off
1086 2010-11-17 12:49:49 <UukGoblin> hrm ;-)
1087 2010-11-17 12:50:05 <ArtForz> 155 sounds about right
1088 2010-11-17 12:50:42 <ArtForz> I'm getting 160
1089 2010-11-17 12:51:38 <ArtForz> stock, that is
1090 2010-11-17 12:52:14 <ArtForz> at 1010 core I get 190
1091 2010-11-17 12:52:30 <UukGoblin> m0's client isn't optimal too
1092 2010-11-17 12:52:37 <ArtForz> yep
1093 2010-11-17 12:52:54 <UukGoblin> I'm kinda worried about temperature
1094 2010-11-17 12:53:11 <ArtForz> it's usually 3-4%& slower than mine, so 155 sounds right
1095 2010-11-17 12:53:17 <UukGoblin> these were quite cheap and the radiator doesn't connect to GPU RAM at all - only the GPU
1096 2010-11-17 12:53:39 <ArtForz> yeah, same here
1097 2010-11-17 12:53:54 <UukGoblin> hopefully RAM won't heat up much
1098 2010-11-17 12:54:06 <ArtForz> as long as core temp stays under 80 or so it should be fine
1099 2010-11-17 13:04:06 joe_1 has joined
1100 2010-11-17 13:11:37 omglolbbq has left ()
1101 2010-11-17 13:26:24 <ArtForz> TEST faucet now has 222222.22 btc
1102 2010-11-17 13:34:59 tcatm has joined
1103 2010-11-17 13:45:50 <altamic> brocktice here
1104 2010-11-17 13:45:52 <altamic> ?
1105 2010-11-17 13:58:30 <xelister> hmm
1106 2010-11-17 13:58:34 <xelister> ArtForz: did you seen that
1107 2010-11-17 13:58:50 <xelister> Radeon HD 5970 has Core clock lower, just 725
1108 2010-11-17 13:59:05 <xelister> why this does not affect generations? is there separate clock for shader
1109 2010-11-17 13:59:32 <xelister> http://en.wikipedia.org/wiki/Comparison_of_AMD_graphics_processing_units#Evergreen_.28HD_5xxx.29_series
1110 2010-11-17 14:01:44 <tcatm> xelister: you can change the clock
1111 2010-11-17 14:02:11 <ArtForz> yes, stock 5970 only has 725 core clk
1112 2010-11-17 14:02:53 <ArtForz> up the core voltage from their default 1.05 to 1.15 (5870 is 1.1625) and they OC to 900 or so
1113 2010-11-17 14:04:39 <xelister> ArtForz: but your 600 M estimation was with default clock?
1114 2010-11-17 14:05:04 <xelister> bedcause based on wikipedia I get
1115 2010-11-17 14:05:18 <ArtForz> stock is prolly ~530m using m0s miner
1116 2010-11-17 14:05:28 <xelister> 3200 cores * 725 clock = 2320000 == 546
1117 2010-11-17 14:05:30 <xelister> yey so Im correct
1118 2010-11-17 14:05:47 <xelister> and the coeef is 4250
1119 2010-11-17 14:06:09 <xelister> but I thought you mentioned yesterday something about 6500 ops to calculate sha, so why is my coeef here just 4250
1120 2010-11-17 14:06:21 <xelister> expected M = cores * clock / 4250
1121 2010-11-17 14:06:48 <ArtForz> on 5xxx its ~ 4150ops
1122 2010-11-17 14:06:53 <xelister> ok
1123 2010-11-17 14:06:58 <ArtForz> on 4xxx and nvidia ~6350 ops
1124 2010-11-17 14:07:03 <xelister> and on nv 2xx its like 16000 it seems
1125 2010-11-17 14:08:11 <Diablo-D3> xelister: yes but
1126 2010-11-17 14:08:18 <Diablo-D3> ops is related to compiler efficiency
1127 2010-11-17 14:08:24 <ArtForz> yep
1128 2010-11-17 14:08:30 <ArtForz> but this is theoretical best case
1129 2010-11-17 14:08:45 <Diablo-D3> UukGoblin: btw, yes, you fucked your DISPLAY up
1130 2010-11-17 14:08:57 <UukGoblin> Diablo-D3, wasn't me! it's ubuntu's fault!
1131 2010-11-17 14:09:23 <Diablo-D3> UukGoblin: ubuntu doesnt set display to :0.0
1132 2010-11-17 14:10:02 <UukGoblin> so what did?
1133 2010-11-17 14:10:54 <Diablo-D3> you?
1134 2010-11-17 14:10:59 <xelister> whats the new diff?
1135 2010-11-17 14:11:07 <xelister> 6500?
1136 2010-11-17 14:11:17 <ArtForz> estnextdiff... 6622
1137 2010-11-17 14:11:26 <ArtForz> and rising
1138 2010-11-17 14:11:44 <Diablo-D3> welp, I guess there goes my new video card
1139 2010-11-17 14:15:10 <xelister> Diablo-D3: goes?
1140 2010-11-17 14:15:25 <xelister> ArtForz: why dont you take a vaction for month or a week ;)
1141 2010-11-17 14:15:29 <brocktice> altamic: What's up?
1142 2010-11-17 14:15:38 * xelister slaps ArtForz for rising the diff faster
1143 2010-11-17 14:15:46 <altamic> hey I am looking into the android thinghie
1144 2010-11-17 14:16:30 <ArtForz> what? I'm only doing 9Gh/s total
1145 2010-11-17 14:16:34 <xelister> full client in Android? sounds like 10.000 BTC worth of work to me
1146 2010-11-17 14:16:34 soultcer has joined
1147 2010-11-17 14:16:36 <altamic> and I was evaluating how to put up a light client
1148 2010-11-17 14:16:45 <xelister> assuming it is a full reimplementation
1149 2010-11-17 14:16:58 <altamic> a sort of frontend
1150 2010-11-17 14:17:06 <altamic> for a bitcoind
1151 2010-11-17 14:17:15 <altamic> speaking with json/rpc
1152 2010-11-17 14:17:15 <xelister> heh :) we need a quote grabbing bot for this channel. "<ArtForz> what? I'm only doing 9Gh/s total"
1153 2010-11-17 14:17:20 <Diablo-D3> xelister: well, if Im no longer generating coins, there goes my funds.
1154 2010-11-17 14:17:28 <Diablo-D3> isnt the total network like 20 ghash?
1155 2010-11-17 14:17:35 <ArtForz> 60
1156 2010-11-17 14:17:52 <brocktice> altamic: OK, yeah I was thinking, you know how there are mobile webkit sites that look a lot like real apps?
1157 2010-11-17 14:17:52 <altamic> I am using phonegap + jquerymobile
1158 2010-11-17 14:17:55 <Diablo-D3> so what if all you gpu farm faggots turned your shit off for a month
1159 2010-11-17 14:18:11 <ArtForz> at the start of this diff it was ~40
1160 2010-11-17 14:18:22 <ArtForz> so once again the usual x1.5
1161 2010-11-17 14:18:34 <altamic> I have just deployed a sample app on the emulator
1162 2010-11-17 14:18:35 <brocktice> altamic: Oh, cool, hadn't heard of phonegap
1163 2010-11-17 14:18:47 <brocktice> xelister: yeah 10k btc sounds not bad
1164 2010-11-17 14:18:55 <altamic> but the real problem is the fat client
1165 2010-11-17 14:19:00 <brocktice> I can't spring for all of that though, we need other people.
1166 2010-11-17 14:19:23 <xelister> brocktice: if you would find funds, probably I would get it done; even more for gayphone
1167 2010-11-17 14:19:36 <brocktice> gayphone?
1168 2010-11-17 14:19:41 <xelister> my friends develops for geyphone
1169 2010-11-17 14:19:44 <xelister> you know, eye-phone
1170 2010-11-17 14:19:49 <brocktice> oh
1171 2010-11-17 14:20:29 <altamic> I mean all that shas would burn battery
1172 2010-11-17 14:21:04 <brocktice> altamic: Well, you can run a fat client without generating.
1173 2010-11-17 14:21:14 <brocktice> Most people do, I think.
1174 2010-11-17 14:21:31 <brocktice> depending on your definition of 'fat'
1175 2010-11-17 14:21:41 <Diablo-D3> well, like brocktice's mom
1176 2010-11-17 14:21:42 <Diablo-D3> fat
1177 2010-11-17 14:21:44 <brocktice> There's a lot of support for removing mining from the mainline client anyway.
1178 2010-11-17 14:22:03 <brocktice> Diablo-D3: har har
1179 2010-11-17 14:22:57 <altamic> brocktice: http://en.wikipedia.org/wiki/Client_(computing)#Fat
1180 2010-11-17 14:23:32 <brocktice> altamic: Right, but you can run bitcoin without generating, right?
1181 2010-11-17 14:23:41 <altamic> ...yes! :)
1182 2010-11-17 14:23:48 <brocktice> As far as I'm concerned, the issue is wallet online vs. wallet on the phone.
1183 2010-11-17 14:23:48 <altamic> you are definitely right :)
1184 2010-11-17 14:24:06 <brocktice> Wallet and processing of transactions on the phone = fat client
1185 2010-11-17 14:24:10 <brocktice> generating on the phone = really fat client
1186 2010-11-17 14:24:36 <brocktice> Everything on a server, simple interface to it on the phone = thin client.
1187 2010-11-17 14:24:56 <brocktice> The problem with a fat client on the phone is that it'll have to maintain a network connection at all times or download blocks whenever you connect it.
1188 2010-11-17 14:25:14 <brocktice> I think a thin client will be the way to go, because otherwise it's still going to burn battery, even without the hashing.
1189 2010-11-17 14:25:20 <brocktice> The radio will drain a lot.
1190 2010-11-17 14:25:22 <ArtForz> something in between would also make sense
1191 2010-11-17 14:25:52 <ArtForz> server keeps block chain, client keeps wallet
1192 2010-11-17 14:25:59 <brocktice> Whether that thin client is a web app or a native app, I don't really care, although I think QR code stuff is going to be hard with the web app.
1193 2010-11-17 14:26:14 <brocktice> particularly using the camera in a user-friendly way.
1194 2010-11-17 14:26:40 <ArtForz> client asks server "any TX for keys x,y,z since block X"
1195 2010-11-17 14:27:43 Akiraa has joined
1196 2010-11-17 14:29:01 <altamic> QR stuff requires a scheme URI for bitcoins
1197 2010-11-17 14:29:20 <brocktice> It ought to be just fine to encode an address in a QR code
1198 2010-11-17 14:29:32 <ArtForz> server might also hand the client block headers (80 bytes/block) and full blocks (or just merkle tree hashes + TX belonging to client), so the client can still verify the block chain
1199 2010-11-17 14:29:57 <dsg> altamic: It's been proposed on the forums
1200 2010-11-17 14:29:59 <ArtForz> imo that'd also be a good setup for a hub-leaf network
1201 2010-11-17 14:30:24 <brocktice> http://chart.apis.google.com/chart?cht=qr&chs=230x230&chl=14FCnThVHJSRpSKAGFvThhaucm8rcvgocA
1202 2010-11-17 14:30:25 RazielZ has joined
1203 2010-11-17 14:30:27 <brocktice> That's an address
1204 2010-11-17 14:31:21 <brocktice> we can bikeshed all day about URIs and the best way to do things, but I think simple and functional is a good start.
1205 2010-11-17 14:31:49 <altamic> yes I agree
1206 2010-11-17 14:32:02 <dsg> brocktice: Encoding an address directly in a QR code without any indication that it's a bitcoin address is a really bad idea.
1207 2010-11-17 14:32:09 <altamic> but I don't think the priority is on QR
1208 2010-11-17 14:32:13 <dsg> There's more uses for QR codes than bitcoin
1209 2010-11-17 14:32:46 <brocktice> dsg: OK, so add "BTC_ADDR:" to the beginning before encoding it
1210 2010-11-17 14:32:50 <brocktice> It can always be changed later.
1211 2010-11-17 14:32:52 <dsg> That's much better
1212 2010-11-17 14:33:16 <dsg> http://pastebin.com/VsBbmXQx
1213 2010-11-17 14:33:30 <dsg> There it is, by zarutian and theymos
1214 2010-11-17 14:33:47 <dsg> This scheme seems fairly logical and throught-through. Why not use it?
1215 2010-11-17 14:33:53 <brocktice> Sure
1216 2010-11-17 14:34:24 <brocktice> When you say, "It's been proposed on the forums", I hear "someone said it would be a good idea", not "someone already came up with a scheme"
1217 2010-11-17 14:34:48 <altamic> dsg: great!
1218 2010-11-17 14:35:58 anarchy has joined
1219 2010-11-17 14:36:49 <brocktice> Well well, brucewagner *did* step up, nice.
1220 2010-11-17 14:37:43 <brocktice> biab
1221 2010-11-17 14:42:21 <gavinandresen> Anybody here running latest bleeding-edge svn bitcoin on a PC? Or anybody willing to? I'm worried a change in svn r182 might cause problems.
1222 2010-11-17 14:48:09 <gavinandresen> Anybody? Beuller?
1223 2010-11-17 14:48:58 anarchy is now known as anarchyx
1224 2010-11-17 14:51:48 <Diablo-D3> so
1225 2010-11-17 14:52:13 <Diablo-D3> I think its official
1226 2010-11-17 14:52:18 <Diablo-D3> I wasted my time making an efficient client
1227 2010-11-17 14:53:34 Guest21087 is now known as OneFixt
1228 2010-11-17 14:53:36 OneFixt has quit (Changing host)
1229 2010-11-17 14:53:36 OneFixt has joined
1230 2010-11-17 15:08:11 genjix has joined
1231 2010-11-17 15:08:13 <genjix> hey
1232 2010-11-17 15:11:01 hippich_ has quit (Ping timeout: 260 seconds)
1233 2010-11-17 15:23:11 <xelister> gavinandresen: hi
1234 2010-11-17 15:23:18 <gavinandresen> Good morning
1235 2010-11-17 15:23:22 <xelister> gavinandresen: Im on linux, I can run from svn perahps, what is the question. are you devel?
1236 2010-11-17 15:24:18 <gavinandresen> I changed the way bitcoin writes to debug.log in svn r182 (latest) and want to make sure it doesn't have unintended consequences, especially on PCs
1237 2010-11-17 15:24:56 <gavinandresen> (I run on Mac and Linux, don't do PC stuff any more)
1238 2010-11-17 15:25:01 <xelister> gavinandresen: I suppose I can try, run bitcoin on linux (generator, slow) from rev182
1239 2010-11-17 15:25:08 <xelister> oh.
1240 2010-11-17 15:25:13 <xelister> by PC you mean "windows"?
1241 2010-11-17 15:25:16 <xelister> its not correct ;)
1242 2010-11-17 15:25:21 <gavinandresen> Yeah, sorry, windows
1243 2010-11-17 15:25:47 <xelister> gavinandresen: ok I will give bitcoin svn try on windows
1244 2010-11-17 15:26:05 <xelister> if you help me to set up how to build it; best tommorow; sounds good? or, to cross-compile it
1245 2010-11-17 15:26:23 <xelister> crosscompile from linux should be easy, IF we get the boost especially crossbuilded
1246 2010-11-17 15:26:41 <gavinandresen> I know nothing about building bitcoin for windows......
1247 2010-11-17 15:26:51 <xelister> oh
1248 2010-11-17 15:27:05 <xelister> if anyone comes areound that does (or want to try to) then ping me, I will help
1249 2010-11-17 15:27:15 <gavinandresen> thanks again!
1250 2010-11-17 15:27:20 <xelister> n/p
1251 2010-11-17 15:56:58 djoot has quit (Changing host)
1252 2010-11-17 15:56:58 djoot has joined
1253 2010-11-17 16:04:24 dwdollar has left ()
1254 2010-11-17 16:07:54 <appamatto> Diablo-D3, why was it a waste?
1255 2010-11-17 16:17:39 <altamic> davout around?
1256 2010-11-17 16:17:41 <altamic> no..
1257 2010-11-17 16:25:42 <brocktice> appamatto: I think because difficulty is getting so high it's not worth running it on his card?
1258 2010-11-17 16:25:47 <brocktice> appamatto: Also, nobody else uses it.
1259 2010-11-17 16:27:00 <altamic> I find names like getreceivedbylabel quite disturbing
1260 2010-11-17 16:27:26 <UukGoblin> hmm looks like aticonfig doesn't report temperature when screen is blanked or disconnected
1261 2010-11-17 16:27:34 <UukGoblin> just hangs there doing nothing
1262 2010-11-17 16:27:58 <ArtForz> are you sure it's only aticonfig hanging?
1263 2010-11-17 16:28:13 <UukGoblin> on read(3, 0x2416824, 4096) = -1 EAGAIN (Resource temporarily unavailable)
1264 2010-11-17 16:28:18 <UukGoblin> miners are mining...
1265 2010-11-17 16:28:44 <ArtForz> ahh, yeah, I've noticed that too sometimes
1266 2010-11-17 16:29:22 <ArtForz> disabling screen blanking did the trick for me
1267 2010-11-17 16:29:39 <UukGoblin> ah, cool
1268 2010-11-17 16:30:11 <ArtForz> I also get weird aticonfig hangs when I hotplug/unplug screens
1269 2010-11-17 16:30:20 <UukGoblin> yeah, I did that, actually
1270 2010-11-17 16:30:32 <UukGoblin> remote xsetting dpms force on didn't help
1271 2010-11-17 16:30:38 <UukGoblin> I bet it's the hot-unplug then
1272 2010-11-17 16:30:43 <ArtForz> usually have to restart X to get aticonfig to work again
1273 2010-11-17 16:30:56 <UukGoblin> it's annoying that you have to have X running for opencl to work
1274 2010-11-17 16:31:00 <ArtForz> yeah
1275 2010-11-17 16:31:30 <ArtForz> not only running, but active
1276 2010-11-17 16:33:01 <brocktice> Diablo-D3: Hey, I tried your miner again
1277 2010-11-17 16:33:10 <brocktice> Diablo-D3: It doesn't quite max out if I run it on one core
1278 2010-11-17 16:33:23 <brocktice> Diablo-D3: If I run it on all 4 it uses about 3 of them 30% each, and the other 0%
1279 2010-11-17 16:33:45 <brocktice> Diablo-D3: I think it's doing something funny when distributing jobs to multiple cores
1280 2010-11-17 16:33:48 <ArtForz> brocktice: what speeds are you getting?
1281 2010-11-17 16:34:03 <brocktice> ArtForz: Well, his counter is screwy, it takes along average, so it was inching up slowly
1282 2010-11-17 16:34:12 <brocktice> ArtForz: but it only hit 97-98% GPU load
1283 2010-11-17 16:34:26 <ArtForz> as I noticed horrible multiGPU scaling with SDK 2.2 and my miner
1284 2010-11-17 16:34:40 <brocktice> Yeah I'd rather stick with 2.1
1285 2010-11-17 16:34:54 <brocktice> When I killed it it was around 320mhash/s on one core
1286 2010-11-17 16:35:02 <brocktice> at current clocks I normally get 330 with m0's
1287 2010-11-17 16:35:15 <brocktice> I'm going to start pushing the voltage back up a little along with the clocks now that I know I didn't fry it.
1288 2010-11-17 16:35:21 <ArtForz> clock speed?
1289 2010-11-17 16:35:26 <brocktice> 900/1200
1290 2010-11-17 16:35:30 <ArtForz> not too shabby
1291 2010-11-17 16:35:51 <ArtForz> I get 338.2 @ 900
1292 2010-11-17 16:36:04 johnyh has quit (Remote host closed the connection)
1293 2010-11-17 16:36:12 <brocktice> Even my bum core can run at 900
1294 2010-11-17 16:36:23 <ArtForz> at stock voltage?
1295 2010-11-17 16:36:26 <brocktice> If I try to go up to 925/950 at 1.1625v they start tanking
1296 2010-11-17 16:36:28 <brocktice> no
1297 2010-11-17 16:36:30 <brocktice> 1.1625
1298 2010-11-17 16:36:33 <ArtForz> ahhh
1299 2010-11-17 16:36:44 <brocktice> at 1.2 all but one could hit 1 GHz
1300 2010-11-17 16:36:48 <ArtForz> my worst core can do 900 @ 1.15
1301 2010-11-17 16:36:54 <brocktice> But I should be able to get it lower than that.
1302 2010-11-17 16:37:01 <ArtForz> at 1.1625 it can do 910
1303 2010-11-17 16:37:13 <ArtForz> and thats one core out of 12 cards
1304 2010-11-17 16:37:20 <brocktice> At 1.9 they can all go pretty fast
1305 2010-11-17 16:37:24 <brocktice> but only for a half a second :)
1306 2010-11-17 16:37:27 <ArtForz> lol
1307 2010-11-17 16:37:39 <ArtForz> actually some maniacs are running 1.4+
1308 2010-11-17 16:37:50 <brocktice> I wonder how long I could run it at 1.3
1309 2010-11-17 16:38:00 <brocktice> Before shit started breaking
1310 2010-11-17 16:38:11 <ArtForz> wild guess... until the VRMs blow up
1311 2010-11-17 16:38:13 <brocktice> If you're just gaming for an hour or whatever I can see really abusing it.
1312 2010-11-17 16:38:18 <ArtForz> yeah
1313 2010-11-17 16:38:24 <brocktice> But running it 24/7 at full blast is another story
1314 2010-11-17 16:38:27 <ArtForz> gaming really doesn't push the cards that hard
1315 2010-11-17 16:38:41 <ArtForz> mining is nearly as bad as 24/7 furmark
1316 2010-11-17 16:38:48 <brocktice> furmark is worse?
1317 2010-11-17 16:38:55 <brocktice> I guess it uses more than just the shaders
1318 2010-11-17 16:38:58 <ArtForz> yeah, but not by much
1319 2010-11-17 16:39:24 <ArtForz> yep
1320 2010-11-17 16:39:47 <ArtForz> it pretty much pushes all units
1321 2010-11-17 16:39:47 <brocktice> 2010-11-17 10:18:11 INFO Device[0]: 330331khash/s
1322 2010-11-17 16:39:47 <brocktice> 2010-11-17 10:18:12 INFO Device[1]: 330298khash/s
1323 2010-11-17 16:39:47 <brocktice> 2010-11-17 10:18:13 INFO Device[3]: 330295khash/s
1324 2010-11-17 16:39:48 <brocktice> 2010-11-17 10:18:14 INFO Device[2]: 330294khash/s
1325 2010-11-17 16:39:53 <brocktice> That's at 900/1200
1326 2010-11-17 16:40:22 <brocktice> 1.1625/1.15
1327 2010-11-17 16:41:05 <ArtForz> 342013 342012 338289 342046 342078 342074 342065 342073
1328 2010-11-17 16:41:18 <ArtForz> thats the box with one core @ 900
1329 2010-11-17 16:41:24 <brocktice> nice
1330 2010-11-17 16:41:28 <ArtForz> all others @ 910, mem @ 1000
1331 2010-11-17 16:41:37 <brocktice> I went back and looked at RAM some more
1332 2010-11-17 16:41:43 <brocktice> I don't think it gives me much after all
1333 2010-11-17 16:42:06 theymos has joined
1334 2010-11-17 16:42:07 <brocktice> I think if we could just get double-buffering in m0's we should make up most of the remaining difference.
1335 2010-11-17 16:42:18 <ArtForz> with my miner 1000 to 1200 was a 0.3% diff
1336 2010-11-17 16:42:31 <ArtForz> and thats with core @ 900
1337 2010-11-17 16:42:46 <brocktice> ArtForz: how's your 6u build coming?
1338 2010-11-17 16:42:48 <ArtForz> core @ 725 it was < 0.2%
1339 2010-11-17 16:43:15 <ArtForz> not too good
1340 2010-11-17 16:43:35 <ArtForz> I have to change the card stack ordering
1341 2010-11-17 16:44:56 <ArtForz> and I don't like how hot the PSUs are getting
1342 2010-11-17 16:45:48 <brocktice> ArtForz: could you fit some ducting in there or something to get them some fresh air?
1343 2010-11-17 16:46:19 <ArtForz> nope
1344 2010-11-17 16:46:57 <brocktice> Any way you could fit them up front like I did with my radiator?
1345 2010-11-17 16:47:04 <ArtForz> they are up front
1346 2010-11-17 16:47:07 <brocktice> Would that obstruct your airflow too much?
1347 2010-11-17 16:47:08 <brocktice> Oh
1348 2010-11-17 16:47:16 <brocktice> fans forward?
1349 2010-11-17 16:47:19 <brocktice> (intake)
1350 2010-11-17 16:47:26 <ArtForz> nope
1351 2010-11-17 16:47:39 <ArtForz> the basic setup is pretty simple
1352 2010-11-17 16:48:04 <ArtForz> PSUs are in front at the left/right exhausting to front
1353 2010-11-17 16:48:21 <brocktice> Where are they pulling from? Case ambient?
1354 2010-11-17 16:48:30 <ArtForz> yep
1355 2010-11-17 16:48:37 <brocktice> you're going to feed the PSU exhaust right back into your intake fans :(
1356 2010-11-17 16:48:40 <ArtForz> yep
1357 2010-11-17 16:48:49 astaroth has joined
1358 2010-11-17 16:48:52 <ArtForz> and at the rear I have the same problem in reverse
1359 2010-11-17 16:49:02 <brocktice> could you flip them so they're drawing from the front and exhausting into the case?
1360 2010-11-17 16:49:11 <ArtForz> I guess so
1361 2010-11-17 16:49:32 hippich has joined
1362 2010-11-17 16:49:35 <ArtForz> well... cables will be too short
1363 2010-11-17 16:49:46 <brocktice> get extenders?
1364 2010-11-17 16:49:50 <anarchyx> how many units are you trying to put into 1 box?
1365 2010-11-17 16:49:58 <anarchyx> how many cards i mea
1366 2010-11-17 16:49:59 <anarchyx> n
1367 2010-11-17 16:50:00 <ArtForz> yeah, at least the 4+4 could get a problem
1368 2010-11-17 16:50:04 <ArtForz> 8 cards in 6u
1369 2010-11-17 16:50:10 <brocktice> anarchyx: he's nuts, he's mounting one motherboard on the lid
1370 2010-11-17 16:50:11 dwdollar1 has joined
1371 2010-11-17 16:50:12 <anarchyx> 8x5970?
1372 2010-11-17 16:50:19 <ArtForz> yep
1373 2010-11-17 16:50:21 <anarchyx> lol
1374 2010-11-17 16:50:25 <anarchyx> good luck with that :D
1375 2010-11-17 16:50:40 <anarchyx> 4 in 4u is already a bit of a challenge imo :)
1376 2010-11-17 16:50:50 <brocktice> anarchyx: Yeah very much so
1377 2010-11-17 16:51:03 <brocktice> My motherboard has stupid slot placement, so it's impossible for me, basically
1378 2010-11-17 16:51:03 <ArtForz> 4 in 5u on air is no problem
1379 2010-11-17 16:51:37 <anarchyx> why are you trying to save out on the 2U?
1380 2010-11-17 16:51:49 <anarchyx> just do 2 times 4U
1381 2010-11-17 16:51:49 <ArtForz> dunno really
1382 2010-11-17 16:51:58 <anarchyx> if youre thinking about a datacenter
1383 2010-11-17 16:52:03 <anarchyx> power = the cost nowadays
1384 2010-11-17 16:52:08 <anarchyx> not so much the space
1385 2010-11-17 16:53:38 <ArtForz> yeah, actually I think you could go even higher density
1386 2010-11-17 16:53:45 <altamic> could anybody provide the url of a running bitcoin server and let me try some interaction?
1387 2010-11-17 16:53:54 <altamic> I would like to test my json/rpc calls
1388 2010-11-17 16:54:02 <brocktice> altamic: I can fire one up, hang on
1389 2010-11-17 16:54:10 <ArtForz> if it werent for the stupid side intake 2 in 1u would be possible
1390 2010-11-17 16:55:34 <ArtForz> honestly it wouldnt be too bad power density wise compared to some HPC setups
1391 2010-11-17 16:56:01 <ArtForz> I've seen designs >1kW/U
1392 2010-11-17 16:56:18 <anarchyx> most datacenters dont support more than 32 amps per rack, or youre paying premiums.. 40 5970s in a rack.. dont know how much amp that would be
1393 2010-11-17 16:56:34 <anarchyx> but thats 10x4U
1394 2010-11-17 16:56:51 <ArtForz> 1 5970 = about 350W, OCed about 400 at the plug
1395 2010-11-17 16:57:14 <bonsaikitten> usually you don't get more than 10A anymore
1396 2010-11-17 16:57:19 <brocktice> Hey, if you set up bitcoind with an rpc user and pass, does it only listen to localhost?
1397 2010-11-17 16:57:24 <ArtForz> for general hosting, yea
1398 2010-11-17 16:57:24 <brocktice> Or does it listen to all incoming connections?
1399 2010-11-17 16:57:28 <brocktice> without further config
1400 2010-11-17 16:57:33 <bonsaikitten> it's quite funny when you use 5U of a 40U rack because it's the cheapest setup :)
1401 2010-11-17 16:57:35 <anarchyx> so with 220Volt that would be...
1402 2010-11-17 16:58:05 <bonsaikitten> ~2.5kW
1403 2010-11-17 16:58:24 <altamic> brocktice: server=true
1404 2010-11-17 16:58:24 <altamic> rpcuser=admin
1405 2010-11-17 16:58:25 <altamic> rpcpassword=password
1406 2010-11-17 16:58:27 <ArtForz> for high density stuff 3-phase to the rack is pretty normal
1407 2010-11-17 16:58:29 <bonsaikitten> and 32A ends up around 20kW
1408 2010-11-17 16:58:33 <altamic> on your bitcoin.conf
1409 2010-11-17 16:59:01 <ArtForz> yep
1410 2010-11-17 16:59:07 <brocktice> altamic: doesn't bitcoind automatically run in server mode?
1411 2010-11-17 16:59:07 <anarchyx> 7 amps per 4U?
1412 2010-11-17 16:59:13 <brocktice> It's listening on 8332, I can see it
1413 2010-11-17 16:59:35 <brocktice> with the username and pass I msged to you
1414 2010-11-17 16:59:39 <bonsaikitten> anarchyx: no, ~3A per 4U ... 600W
1415 2010-11-17 16:59:45 <ArtForz> beyond 20kW or so per rack cooling also becomes a bit of an issue
1416 2010-11-17 17:00:35 <bonsaikitten> ArtForz: only financial issue ... 50kW per rack can be had easily, closed rack with a fullsize rack heat exchanger
1417 2010-11-17 17:00:58 <anarchyx> bonsaikitten: 4X5970 in a 4U would be 1500Watt right?
1418 2010-11-17 17:01:10 <ArtForz> yeah
1419 2010-11-17 17:01:43 <ArtForz> a bit less
1420 2010-11-17 17:01:50 <ArtForz> stock about 1400W at the plug
1421 2010-11-17 17:02:02 <bonsaikitten> too much for most datacenters
1422 2010-11-17 17:02:03 <anarchyx> so its almost 7 amps per 4U at 220V
1423 2010-11-17 17:02:08 <ArtForz> yep
1424 2010-11-17 17:02:26 <bonsaikitten> which means with 32A feed you'd have only half the rack full
1425 2010-11-17 17:02:30 <bonsaikitten> and that's pushing it
1426 2010-11-17 17:02:34 <anarchyx> which would allow for less than 5 boxes in a full rack.. at 32 amp
1427 2010-11-17 17:02:38 <brocktice> biab
1428 2010-11-17 17:02:39 <anarchyx> yeah
1429 2010-11-17 17:03:02 <bonsaikitten> but you don't need the bandwidth of a datacenter
1430 2010-11-17 17:03:03 <anarchyx> you cant cool much more than that without specialised equipment i think
1431 2010-11-17 17:03:15 <bonsaikitten> you could set up your machines somewhere in an old industrial building
1432 2010-11-17 17:03:26 <anarchyx> yes i guess :)
1433 2010-11-17 17:03:42 <bonsaikitten> they usually have 200A+ three-phase feeds, and a large building allows for free cooling
1434 2010-11-17 17:03:50 <bonsaikitten> just keep the doors open at all times ;)
1435 2010-11-17 17:03:51 <anarchyx> or in a shed in winter.. but thats moist :)
1436 2010-11-17 17:04:13 <bonsaikitten> too cold would be bad, temperature difference can break hardware
1437 2010-11-17 17:04:30 <bonsaikitten> I heard funny stories when people brought stuff in from their car where it had cooled down to ~0C
1438 2010-11-17 17:04:38 <bonsaikitten> bad idea to fire up such a machine immediately :)
1439 2010-11-17 17:04:45 <ArtForz> also, 5970s dont like to start when they're < 10°C or so
1440 2010-11-17 17:05:20 <theymos> Bitcoin Block Explorer now has a testnet version: http://theymos.ath.cx:64150/testnet/bbe/
1441 2010-11-17 17:05:35 <ArtForz> people running chilled water found that out
1442 2010-11-17 17:05:43 <ArtForz> theymos: cool
1443 2010-11-17 17:07:50 <bonsaikitten> ArtForz: similar to the AMD coldbug I guess, processors got unhappy around the freezing point
1444 2010-11-17 17:07:54 <ArtForz> yep
1445 2010-11-17 17:08:15 <ArtForz> theres a beta bios floating around that seems to fix it
1446 2010-11-17 17:08:59 <ArtForz> I dont quite get why you'd run a 5970 on chilled water, but hey...
1447 2010-11-17 17:10:39 <dsg> ArtForz: Does your miner also use up a whole CPU core with stream SDK 2.2?
1448 2010-11-17 17:10:45 <ArtForz> yea
1449 2010-11-17 17:10:55 <dsg> Ok, then it's not a bug in m0's code.
1450 2010-11-17 17:11:06 <ArtForz> and multiGPU seems to scale horribly with 2.2
1451 2010-11-17 17:11:26 <dsg> I get better performance with 2.2 though (only 1 GPU)
1452 2010-11-17 17:11:45 <dsg> But I'm not sure it's worth it for losing one core
1453 2010-11-17 17:19:12 <ByteCoin> What's the current block number? I forget the command
1454 2010-11-17 17:19:19 <theymos> ;;bc,blocks
1455 2010-11-17 17:19:20 <gribble> 92454
1456 2010-11-17 17:20:24 <ByteCoin> Cheers! themos has spotted two coinbase transactions with identical hashes about 800 blocks ago!
1457 2010-11-17 17:20:51 <ByteCoin> That's bad!
1458 2010-11-17 17:21:03 <gavinandresen> Bad for whoever generated them....
1459 2010-11-17 17:21:04 <ArtForz> why?
1460 2010-11-17 17:21:08 <ArtForz> yea
1461 2010-11-17 17:21:22 <theymos> gavinandresen: So the duplicates are definitely invalid when spent?
1462 2010-11-17 17:21:30 <gavinandresen> One of them will be
1463 2010-11-17 17:21:57 <theymos> Excelent. That makes fixing BBE much easier.
1464 2010-11-17 17:22:13 <ArtForz> btw, why the hell are we putting block.nBits in coinbase txin?
1465 2010-11-17 17:22:29 <ByteCoin> Bad design. No idea about nBits
1466 2010-11-17 17:22:48 <ByteCoin> Arbitary and pointless
1467 2010-11-17 17:23:05 <theymos> nBits was even removed for like one SVN revision.
1468 2010-11-17 17:23:09 <ByteCoin> I should have spotted this flaw though....
1469 2010-11-17 17:23:33 <ArtForz> yeah, not needed, and if you want to include something from the block use height or start of prevblock hash or something to make it unique per block
1470 2010-11-17 17:24:29 <ArtForz> sure, 5 bytes wasted per block isnt THAT bad, but still 5 bytes wasted
1471 2010-11-17 17:25:17 <ByteCoin> Well if we're going to get annoyed about bytes wasted what about the prevout hash?
1472 2010-11-17 17:25:55 <ByteCoin> As I've said before. Bitcoin values clients diskspace as zero.
1473 2010-11-17 17:26:05 <ArtForz> yep
1474 2010-11-17 17:26:49 <Akiraa> bonsaikitten: Can you gradually overclock a system so it doesn't undergo thermal shock from the increasing cooling requirements?
1475 2010-11-17 17:27:00 <gavinandresen> "Let he who has written his own implementation cast the first stone?"
1476 2010-11-17 17:27:18 <ArtForz> *casts stone*
1477 2010-11-17 17:27:20 <anarchyx> this datacenter just quoted me 100 euro / amp (includes rackspace)
1478 2010-11-17 17:27:22 <ByteCoin> I wonder if satoishi will be tempted to release an updated client that, among the other changes artificially credits that public key with 59BTC
1479 2010-11-17 17:27:26 <anarchyx> thats expensive :P
1480 2010-11-17 17:27:27 <ByteCoin> 50BTC
1481 2010-11-17 17:27:43 <bonsaikitten> Akiraa: the biggest problem is material stress, so just waiting is usually enough
1482 2010-11-17 17:28:07 <ByteCoin> *casts stone* Thanks Artforz
1483 2010-11-17 17:28:09 <anarchyx> 100 euro / 1amp per month includes full usage of it
1484 2010-11-17 17:28:10 <brocktice> anarchyx: datacenters are damned expensive
1485 2010-11-17 17:28:33 <anarchyx> how much euro per kilowatth would that be
1486 2010-11-17 17:28:55 <anarchyx> or better, would it be profitable for bitcoin generation :)
1487 2010-11-17 17:28:58 <brocktice> i forget the conversion factor for RMS
1488 2010-11-17 17:29:27 <ArtForz> I'm guessing EU? = 230Veff ?
1489 2010-11-17 17:29:52 <anarchyx> belgium
1490 2010-11-17 17:29:56 <anarchyx> yah
1491 2010-11-17 17:30:10 <ArtForz> then it's 0.23kW
1492 2010-11-17 17:30:31 <brocktice> local datacenter here wants almost US$0.30/kwh
1493 2010-11-17 17:30:56 <brocktice> and that doesn't even cover the rack space
1494 2010-11-17 17:31:12 <ArtForz> so.. about 168kWh/month ?
1495 2010-11-17 17:31:22 <brocktice> much cheaper to just heat my basement
1496 2010-11-17 17:32:06 <anarchyx> and 1 5970 does about 750 BTC per month at the new rate?
1497 2010-11-17 17:32:53 <ArtForz> so 100 eur for 168kWh .. why does that sound like a really crappy deal?
1498 2010-11-17 17:32:59 <anarchyx> so thats 750BTC per 1,5 amp =
1499 2010-11-17 17:33:08 <anarchyx> 150 euro for 750 BTC
1500 2010-11-17 17:33:26 <anarchyx> unless i miscaculated
1501 2010-11-17 17:33:28 <ArtForz> thats like 0.595 eur/kWh
1502 2010-11-17 17:33:56 <brocktice> ouch
1503 2010-11-17 17:34:08 <brocktice> highway robbery
1504 2010-11-17 17:34:26 <ArtForz> unless I missed something
1505 2010-11-17 17:34:43 <brocktice> well it includes the rack space
1506 2010-11-17 17:34:47 <anarchyx> whats the standard price per kwh?
1507 2010-11-17 17:34:53 <anarchyx> 0.16?
1508 2010-11-17 17:35:00 <altamic> .14 here
1509 2010-11-17 17:35:06 <brocktice> I pay about US$0.09 before taxes
1510 2010-11-17 17:35:10 <xelister> I think around 0.15 euro here
1511 2010-11-17 17:35:19 <ArtForz> around here utility power is expensive, 0.23 EUR
1512 2010-11-17 17:35:23 <brocktice> but the rack space itself is not bad
1513 2010-11-17 17:35:27 <brocktice> I mean, that's included
1514 2010-11-17 17:35:28 <altamic> ~ .155 after taxxes
1515 2010-11-17 17:35:56 <anarchyx> think only way this is profitable = doing it in your house and leave windows open
1516 2010-11-17 17:35:58 <brocktice> What happns to the price after the first amp?
1517 2010-11-17 17:36:07 <anarchyx> it stays the same
1518 2010-11-17 17:36:38 <brocktice> Unless you have a friend with some free/cheap space in a datacenter, it's not worth it.
1519 2010-11-17 17:37:00 <brocktice> However, I will mention that if you want to rack 'em in your house Ikea Lack coffee tables make great Enterprise LackRacks.
1520 2010-11-17 17:37:16 <anarchyx> :D
1521 2010-11-17 17:37:21 <brocktice> One whole side of my desk is two of them bolted together
1522 2010-11-17 17:38:25 <anarchyx> i think im gonna have to tap power of the street lights
1523 2010-11-17 17:38:31 <anarchyx> ^^
1524 2010-11-17 17:38:42 <brocktice> anarchyx: Look up "The Radioactive Boy Scout"
1525 2010-11-17 17:38:49 <brocktice> Just think of it as an instruction manual. :)
1526 2010-11-17 17:38:51 <anarchyx> anyone every considered solar? its subsidized
1527 2010-11-17 17:38:59 <anarchyx> k :)
1528 2010-11-17 17:39:23 <brocktice> http://www.flickr.com/photos/brocktice/5184503481/
1529 2010-11-17 17:40:45 <brocktice> $20/table
1530 2010-11-17 17:41:08 <brocktice> and a few bucks for the metal joiner plates
1531 2010-11-17 17:51:08 <xelister> so anyone has formula for
1532 2010-11-17 17:51:28 <xelister> Wats needed for earning 1 USD from selling BTC?
1533 2010-11-17 17:51:39 <xelister> or Wats to produce 1 BTC
1534 2010-11-17 17:51:46 <xelister> ArtForz: ?
1535 2010-11-17 17:51:57 <ArtForz> yes
1536 2010-11-17 17:52:33 <xelister> what is the formula?
1537 2010-11-17 17:53:23 <ArtForz> sec, combining em
1538 2010-11-17 17:57:35 <ArtForz> watthoursperbtc = watts * difficulty / hashpersec * 2**32 / 50 / 3600
1539 2010-11-17 17:58:17 <xelister> uh
1540 2010-11-17 17:58:29 <xelister> ok so "now" for (soon) diff 6500, what is the outcome
1541 2010-11-17 17:58:33 <xelister> I calculated that
1542 2010-11-17 17:58:36 <ArtForz> sec
1543 2010-11-17 17:59:20 <ArtForz> I think I have a mistake in there somewhere
1544 2010-11-17 17:59:33 <ArtForz> no, looks right
1545 2010-11-17 18:00:12 <ArtForz> diff 6500, 5970 @ 400W and 680Mhps I get 91.2 Wh/btc
1546 2010-11-17 18:00:50 <ArtForz> that sounds wrong
1547 2010-11-17 18:01:00 <xelister> one Radeon HD 5770 is expected to generate around 118 USD in income (not counting any costs) in one month from now - this seems similar to your calculation, ArtForz?
1548 2010-11-17 18:01:43 <ArtForz> 91.2 Wh = 0.0912kWh, @ 0.20/kWh thats $0.0182.../btc
1549 2010-11-17 18:02:30 <xelister> how much does the card use on it own while mining, the max given in spec around, yes?
1550 2010-11-17 18:02:40 <brocktice> 400 or so when OCed
1551 2010-11-17 18:02:54 <ArtForz> yep, thats at the plug including the rest of the system
1552 2010-11-17 18:03:08 <ArtForz> with a 80+ gold PSU
1553 2010-11-17 18:03:09 <brocktice> oh
1554 2010-11-17 18:03:26 <ArtForz> a few W more @ 120V as PSU efficiency is a few % lower
1555 2010-11-17 18:03:31 <brocktice> whatsit for the card alone?
1556 2010-11-17 18:04:03 <ArtForz> around 350
1557 2010-11-17 18:04:37 <brocktice> One of these days, maybe when I reflash the card bioses, I need to install that kill-a-watt
1558 2010-11-17 18:05:24 <ArtForz> kill-a-watt = suck, get a true power meter
1559 2010-11-17 18:05:28 <xelister> ArtForz: what is your estimate of expected income in this month (diff 6700 assumed) from radeon-5970 ?
1560 2010-11-17 18:05:38 <ArtForz> remember, VA != W ;)
1561 2010-11-17 18:05:44 <ArtForz> sec
1562 2010-11-17 18:06:50 <ArtForz> 3100BTC
1563 2010-11-17 18:07:04 <brocktice> ArtForz: do you keep any or just sell them all?
1564 2010-11-17 18:07:15 <ArtForz> I keep a few 10k around
1565 2010-11-17 18:08:15 <xelister> ArtForz: so many? I estimated 1830 BTC
1566 2010-11-17 18:08:30 <ArtForz> oh, stock 5970?
1567 2010-11-17 18:08:31 <xelister> ArtForz: because diff now is 6700, and it will rise in 2 weeks by x2 for example
1568 2010-11-17 18:08:35 <brocktice> xelister: difficulty went up
1569 2010-11-17 18:08:44 <ArtForz> erm... now its 6700 ?
1570 2010-11-17 18:08:44 <brocktice> xelister: it was lower at the beginning of the month
1571 2010-11-17 18:08:53 <xelister> it will be 6700 in a day or so, no? you said
1572 2010-11-17 18:09:00 <ArtForz> ahhh, yes
1573 2010-11-17 18:09:06 <ArtForz> okay... then again
1574 2010-11-17 18:09:13 <brocktice> Right but if you're saying 'this month', that includes the earlier part of the month
1575 2010-11-17 18:09:33 <xelister> brocktice: Im advocating the opposite point to what you think I am :) I say it will be LESS - it will be just 1830 BTC and not 3100
1576 2010-11-17 18:09:45 <xelister> ah ok. In a month, starting from now
1577 2010-11-17 18:09:47 <ArtForz> so we assume x1.5 every 10 days or so?
1578 2010-11-17 18:09:50 <brocktice> xelister: I know
1579 2010-11-17 18:10:00 <ArtForz> well, x1.5 every 9.333 days
1580 2010-11-17 18:10:20 <xelister> I assumed that 1 month from now will be: 2 weeks from now will be 6700, and next 2 weeks will be x2 harder
1581 2010-11-17 18:10:56 <ArtForz> we've beed doing pretty much that recently, ~ x1.5 every 9.x days
1582 2010-11-17 18:10:56 <xelister> ok so starting frow now, you said 3100/2 it gives 1500 BTC... that is close to mine 1800 BTC
1583 2010-11-17 18:11:26 * xelister updates his formula for diff rising
1584 2010-11-17 18:11:44 <xelister> so 1 month is: 10days + 10days/1.5 + 10days/1.5^2
1585 2010-11-17 18:12:48 <ArtForz> yep
1586 2010-11-17 18:13:15 <brocktice> xelister: that can't go on indefinitely though
1587 2010-11-17 18:13:22 <xelister> so 1500 - 1700 BTC
1588 2010-11-17 18:13:37 <ArtForz> I get 2071
1589 2010-11-17 18:13:37 <xelister> brocktice: you think that soon rising of difficulty may slow down?
1590 2010-11-17 18:13:40 <brocktice> Only so many people are willing to lay out the cash up front for the hardware and run it all the time.
1591 2010-11-17 18:13:48 <xelister> brocktice: indeed
1592 2010-11-17 18:13:53 <brocktice> And it's about to become unprofitable on nVidia hardware per Art's calculations.
1593 2010-11-17 18:14:07 <ArtForz> well, not quite it seems
1594 2010-11-17 18:14:17 <ArtForz> nv is about 4x worse for hash/W
1595 2010-11-17 18:14:58 <xelister> ArtForz: how much % do you OC the cards/
1596 2010-11-17 18:15:09 <brocktice> ArtForz: you should start a FAQ
1597 2010-11-17 18:15:15 <ArtForz> 910 core... so 25.5%
1598 2010-11-17 18:15:25 <xelister> so much?
1599 2010-11-17 18:15:28 <ArtForz> yep
1600 2010-11-17 18:15:36 <brocktice> xelister: they ship quite underclocked
1601 2010-11-17 18:15:44 <brocktice> because of PCIe power specifications
1602 2010-11-17 18:15:46 <xelister> and it gives correct speed? because it could start doing errors quietly
1603 2010-11-17 18:16:01 <ArtForz> I have my own tool for burn-in testing em
1604 2010-11-17 18:16:08 <xelister> ArtForz: share it? :)
1605 2010-11-17 18:16:14 <brocktice> xelister: check out memtestCL
1606 2010-11-17 18:16:22 <brocktice> it's not as slick as Art's but it should be a decent check
1607 2010-11-17 18:16:54 <ArtForz> my code is really quite simple
1608 2010-11-17 18:17:33 <ArtForz> I basically replaced the (if H == 0) with output = H
1609 2010-11-17 18:18:03 <ArtForz> and instead of grabbing a block header from the client I just renerate 80 bytes of randomness
1610 2010-11-17 18:18:56 <ArtForz> save the result of the first kernel run and compare it against subsequent runs
1611 2010-11-17 18:19:15 <ArtForz> cry foul if it doesn't match
1612 2010-11-17 18:19:59 <ArtForz> but then for our work we don't need 100% perfect results
1613 2010-11-17 18:20:14 <ArtForz> = we dont really care if 0.01% of caclulated hashes are wrong
1614 2010-11-17 18:20:28 <xelister> thats what I thought too
1615 2010-11-17 18:20:34 <xelister> unless we damage the cards etc
1616 2010-11-17 18:20:37 <ArtForz> yep
1617 2010-11-17 18:20:44 <xelister> btw, ArtForz you plan on reselling all thease cards later?
1618 2010-11-17 18:20:55 <xelister> probably in 2011 Q1 if you would continue you anyway should switch to 6xxx
1619 2010-11-17 18:21:12 <ArtForz> depends on how good 6970/6990 ends up
1620 2010-11-17 18:21:34 <ArtForz> I honestly dont have too high hopes
1621 2010-11-17 18:22:27 <ArtForz> most likely I'll keep 8-10 of the 5970s or so
1622 2010-11-17 18:22:47 <xelister> perhaps there isnt enought deman in games for such cards yet
1623 2010-11-17 18:24:05 <tcatm> jgarzik: are you around?
1624 2010-11-17 18:24:08 <xelister> for what .. game cennter? ;)
1625 2010-11-17 18:24:23 <ArtForz> yep
1626 2010-11-17 18:24:41 <xelister> woot
1627 2010-11-17 18:26:10 <ArtForz> btw, modern demanding games + high AA + 2560 res or dual 1920 = dips to single-digit FPS even with 2x5970
1628 2010-11-17 18:27:24 <ArtForz> thats with a 6-core Phenom @ 3.8
1629 2010-11-17 18:31:13 <brocktice> ArtForz: As long as it can run Quake III I'll be happy.
1630 2010-11-17 18:31:25 <brocktice> And StarCraft II
1631 2010-11-17 18:31:32 <brocktice> And my old macbook can do the first one
1632 2010-11-17 18:31:37 <brocktice> the new laptop can do the second
1633 2010-11-17 18:38:45 <ArtForz> btw, next diff in ~24h
1634 2010-11-17 18:43:17 <brocktice> frack
1635 2010-11-17 18:43:21 altamic has quit (Ping timeout: 272 seconds)
1636 2010-11-17 18:45:04 <nanotube> hey, so i want to add the 'estimate' data to gribble, since bitbot is often dead lately... is there a nice place to pull that data from without having to calculate my own?
1637 2010-11-17 18:45:20 <ArtForz> dont think so :/
1638 2010-11-17 18:46:42 <brocktice> nanotube: shouldn't be that hard to calculate
1639 2010-11-17 18:46:46 <brocktice> who runs bitbot btw?
1640 2010-11-17 18:46:47 <nanotube> mmm sucks.
1641 2010-11-17 18:46:56 <nanotube> necrodearia runs bitbot. well, ran. until it died. heh.
1642 2010-11-17 18:47:00 <ArtForz> necrodearia
1643 2010-11-17 18:47:06 <theymos> nanotube: I'm going to add it to my real-time stats page one of these days.
1644 2010-11-17 18:47:16 <brocktice> nanotube: Just get Art's formula
1645 2010-11-17 18:47:31 <nanotube> brocktice: formulas are nice... but i'd have to actually access block chain data too
1646 2010-11-17 18:48:31 altamic has joined
1647 2010-11-17 18:48:33 <nanotube> theymos: i'll be waiting. let me know when you have it up. :)
1648 2010-11-17 18:48:36 <ArtForz> you'd only need block timestamps
1649 2010-11-17 18:48:54 <nanotube> ArtForz: yes... which is block chain data. :)
1650 2010-11-17 18:49:05 <ArtForz> well, it's not the full block chain data
1651 2010-11-17 18:49:17 <nanotube> is there a nice place i can access block timestamps, then? :)
1652 2010-11-17 18:49:46 <tcatm> blk.bitcoinwatch.com?
1653 2010-11-17 18:50:53 <theymos> You can get almost the same output as getblock (more precision, actually) with BBE's rawblock: http://theymos.ath.cx:64150/bbe/rawblock/00000000839a8e6886ab5951d76f411475428afc90947ee320161bbf18eb6048
1654 2010-11-17 18:51:03 <ByteCoin> theymos: While you're here. I've posted something new on BBE on the forum. My main short term request is to add block and transaction sizes in bytes to the block page.
1655 2010-11-17 18:51:17 <theymos> I'd have to modify getblock.
1656 2010-11-17 18:51:34 <theymos> I can probably figure out how to do it, but it's low priority.
1657 2010-11-17 18:54:12 <nanotube> theymos: is it possible to access blocks on your bbe using block number?
1658 2010-11-17 18:54:16 <nanotube> like on blk.bitcoinwatch?
1659 2010-11-17 18:55:16 <theymos> nanotube: No.
1660 2010-11-17 18:55:25 <nanotube> because to calculate estimate, i'd need to look at timestamp of block #(2016*n-1) and look at the time elapsed between latest block and the previous difficulty-change block
1661 2010-11-17 18:55:40 <ArtForz> yep
1662 2010-11-17 18:55:56 <nanotube> er... i mean #(2016*(n-1)) (forgot parens :P )
1663 2010-11-17 18:56:25 <ArtForz> I think last diff change block# = current block# - current block# mod 2016 - 1
1664 2010-11-17 18:57:00 <ByteCoin> nanotube: You can go to the root page and enter the block number in the search box manually.
1665 2010-11-17 18:57:19 <ByteCoin> That is theymos.ath.cx:64160/bbe
1666 2010-11-17 18:57:29 <ByteCoin> 64150
1667 2010-11-17 18:57:59 <nanotube> ByteCoin: i could... but it would be a pain to have my bot fill in forms. :) i'd rather just have it fetch "someurl/<blocknum>"
1668 2010-11-17 18:58:16 <theymos> I'll create the estimage page right now.
1669 2010-11-17 18:58:22 <ByteCoin> Yes I assumed as much.
1670 2010-11-17 18:58:22 <ArtForz> yep
1671 2010-11-17 18:58:35 <ArtForz> this also gets the corner cases right
1672 2010-11-17 18:58:37 <nanotube> theymos: hehe ok, you'll save me trouble, thanks :)
1673 2010-11-17 18:59:18 <nanotube> ArtForz: shouldn't the last diff change block be (cur block % 2016) * 2016 ?
1674 2010-11-17 18:59:27 <ArtForz> errr... no
1675 2010-11-17 18:59:36 <ByteCoin> themos: Add the block and transaction size stuff while you're at it hey won't you?
1676 2010-11-17 18:59:41 <ArtForz> current block, rounded down to next multiple of 2016, -1
1677 2010-11-17 18:59:41 <ByteCoin> Please
1678 2010-11-17 19:00:23 <nanotube> ArtForz: er heh yea sorry, i meant not mod, but "integer division"
1679 2010-11-17 19:00:40 <nanotube> which is of course a whole other operation. lol. but that's what i was thinking.
1680 2010-11-17 19:01:04 <ArtForz> so yeah, floor(cur block / 2016) * 2016 - 1
1681 2010-11-17 19:03:03 <nanotube> why -1? wasn't the previous diff-change-block 90720, which is 45*2016?
1682 2010-11-17 19:03:08 <ArtForz> nope
1683 2010-11-17 19:03:24 <nanotube> ah i see what you mean, you're talking about 'the first block calculated with new difficulty'
1684 2010-11-17 19:03:33 <nanotube> i was talking about 'last block calculated with old difficulty'
1685 2010-11-17 19:03:34 <nanotube> heh
1686 2010-11-17 19:03:36 <ArtForz> the %2016 == 0 block is already at the new diff
1687 2010-11-17 19:03:55 <nanotube> ah
1688 2010-11-17 19:04:18 <nanotube> i though %2016==0 was the 'last block at old diff'. ok
1689 2010-11-17 19:04:21 <nanotube> thought
1690 2010-11-17 19:04:26 <ArtForz> nope, it's a bit confusing
1691 2010-11-17 19:04:49 <nanotube> and the genesis block is block #0?
1692 2010-11-17 19:04:54 <ArtForz> -1
1693 2010-11-17 19:05:11 <ArtForz> at least I think so.. sec
1694 2010-11-17 19:05:43 <nanotube> if it's block #0, then the number of blocks before first difficulty change is 2016, which would keep things consistent... :)
1695 2010-11-17 19:06:05 <ArtForz> hrrrm
1696 2010-11-17 19:06:10 <nanotube> if it's -1, then... it would mean that first batch of blocks before difficulty change is 2017. so i bet it's block #0
1697 2010-11-17 19:06:23 <ArtForz> looks like it's -1
1698 2010-11-17 19:06:28 <nanotube> ah, bbe shows genesis as #0
1699 2010-11-17 19:06:38 <nanotube> or is that not the genesis block?
1700 2010-11-17 19:06:49 <theymos> Last retarget=90719?
1701 2010-11-17 19:07:19 <ArtForz> yep
1702 2010-11-17 19:07:39 <ArtForz> or... wait... ARGH
1703 2010-11-17 19:07:42 <nanotube> so 90720 was first with current diff eh?
1704 2010-11-17 19:07:51 <nanotube> haha yea this is a bit confusing.
1705 2010-11-17 19:08:01 <nanotube> ;;bc,nexttarget
1706 2010-11-17 19:08:02 <gribble> 92736
1707 2010-11-17 19:08:07 <ArtForz> yeah
1708 2010-11-17 19:08:19 <ArtForz> retarget after 90719
1709 2010-11-17 19:08:22 <nanotube> so, 92736 - is that "this block will be the first with new difficulty", or "this block will be the last with old difficulty" ?
1710 2010-11-17 19:08:43 <nanotube> i guess it should be the former
1711 2010-11-17 19:08:47 <ArtForz> former
1712 2010-11-17 19:08:49 <nanotube> since block counts start from 0
1713 2010-11-17 19:09:12 <ArtForz> last found at old diff is % 2016 == 2015
1714 2010-11-17 19:09:33 <ArtForz> and we need to start timing from that one
1715 2010-11-17 19:10:16 <nanotube> i see... aka n+1 %2016 == 0, to adjust for the 0-based counting.
1716 2010-11-17 19:10:22 <nanotube> it all makes sense now. :)
1717 2010-11-17 19:10:33 <ArtForz> and block 0 isnt genesis I think
1718 2010-11-17 19:10:47 <nanotube> yea, that's what i gathered from your earlier statements
1719 2010-11-17 19:10:55 <ArtForz> wait... it is? wtf
1720 2010-11-17 19:11:03 <nanotube> haha
1721 2010-11-17 19:11:11 * nanotube ungathers
1722 2010-11-17 19:11:32 <ArtForz> I start counting at genesis == -1, why the fuck is then my block height right?
1723 2010-11-17 19:11:49 <ArtForz> no, wait, genesis IS 0
1724 2010-11-17 19:11:56 <ByteCoin> How do you use block height?
1725 2010-11-17 19:12:23 <ArtForz> and retarget is after block with height % 2016 == 2015
1726 2010-11-17 19:13:01 <nanotube> yea, block 0 in bbe has prev block as all zeros... so it must be the genesis
1727 2010-11-17 19:13:15 <ArtForz> but then how the fuck did that work at the 2015 retarget?
1728 2010-11-17 19:13:30 <ArtForz> oh, special case
1729 2010-11-17 19:13:33 <nanotube> we had a total of 2016 blocks at the first difficulty level
1730 2010-11-17 19:13:37 <nanotube> 0-2015
1731 2010-11-17 19:13:41 <nanotube> right?
1732 2010-11-17 19:13:52 <nanotube> so the block batches remain consistently 2016 in size
1733 2010-11-17 19:14:13 <ArtForz> yes
1734 2010-11-17 19:14:36 <nanotube> \o/ i think i finally get it. :)
1735 2010-11-17 19:14:53 <ArtForz> but what timestamp was before block 0 ?
1736 2010-11-17 19:15:26 <nanotube> mm good question
1737 2010-11-17 19:15:43 <nanotube> guess /that/ was a special case
1738 2010-11-17 19:16:13 <ArtForz> yeah, I think we had a special case for that one
1739 2010-11-17 19:16:51 <nanotube> anyway, so theymos let me know when you have pretty next difficulty target output on your site. :)
1740 2010-11-17 19:17:53 <ArtForz> okay... we retarget when block# % 2016 is == 2015
1741 2010-11-17 19:18:20 <ArtForz> and from that we go back 2015 blocks
1742 2010-11-17 19:19:10 <ArtForz> so.. we end up at block % 2016 == 0
1743 2010-11-17 19:19:52 <ArtForz> urk, this is really confusing
1744 2010-11-17 19:20:17 <ArtForz> this means we just act like the timespan between 2015 and 0 never existed?
1745 2010-11-17 19:22:41 <ArtForz> *head hurts* *adds some debug prints to getnextworkrequired*
1746 2010-11-17 19:27:14 <genjix> anyone know a good tech for if clients want to be notified of changes in a list of lines (added and removed)
1747 2010-11-17 19:27:41 <genjix> like on a server you have 10 lines of text. when lines are removed/added, the clients get notified.
1748 2010-11-17 19:28:03 <genjix> atm im thinking hosting xml on server for when client initially connect, then use irc for broadcasting changes
1749 2010-11-17 19:28:08 <genjix> but that seems a bit clunky
1750 2010-11-17 19:28:15 <genjix> any better ideas?
1751 2010-11-17 19:36:08 <theymos> ArtForz: This look right? http://theymos.ath.cx:64150/q/estimate
1752 2010-11-17 19:37:27 <ArtForz> I get 6697.6
1753 2010-11-17 19:38:22 <theymos> The seconds between the blocks is 717485?
1754 2010-11-17 19:38:51 <ArtForz> blocks 90719 and 92480?
1755 2010-11-17 19:39:31 <theymos> Yes.
1756 2010-11-17 19:39:50 <ArtForz> sec
1757 2010-11-17 19:40:34 <ArtForz> yep
1758 2010-11-17 19:41:50 <ArtForz> 717485 sec for 1761 blocks -> ~407.43 sec/blk avg
1759 2010-11-17 19:44:26 <ArtForz> 600(target sec/blk) * 1761(blocks) / 717485(secs) = 1.4726...
1760 2010-11-17 19:45:15 <ArtForz> 600(target sec/blk) * 1761(blocks) / 717485(secs) * 4536.4 (current diff) = 6680
1761 2010-11-17 19:45:19 <ArtForz> now I'm completely confused
1762 2010-11-17 19:49:39 <xelister> diff is still at 4536, why so? (not that I'm complaining ;)
1763 2010-11-17 20:02:02 <theymos> ArtForz: Is this all correct? http://theymos.ath.cx:64150/q/estimate
1764 2010-11-17 20:11:05 Maccer has quit (Ping timeout: 250 seconds)
1765 2010-11-17 20:19:35 <theymos> nanotube: http://theymos.ath.cx:64150/q/estimate seems to be correct.
1766 2010-11-17 20:43:34 RazielZ has quit (Read error: Connection reset by peer)
1767 2010-11-17 20:55:59 * brocktice is annoyed at the whining about miners
1768 2010-11-17 20:59:14 * nameless !~root@weowntheinter.net|is annoyed at the whining ByteCoin
1769 2010-11-17 20:59:24 <nameless> !~root@weowntheinter.net|and by the tab-complete fail
1770 2010-11-17 21:13:32 <xelister> brocktice: what whining about miners?
1771 2010-11-17 21:18:47 Diablo-D3 has quit (Ping timeout: 265 seconds)
1772 2010-11-17 21:20:50 warner has joined
1773 2010-11-17 21:27:43 <brocktice> xelister: http://www.bitcoin.org/smf/index.php?topic=1771.0;all
1774 2010-11-17 21:31:47 <xelister> Radeon cluster - She take my money, well I'm in need
1775 2010-11-17 21:31:50 <xelister> Yeah she's a triflin' friend indeed
1776 2010-11-17 21:31:56 <xelister> he's a gold digger way over time
1777 2010-11-17 21:32:04 * xelister dances
1778 2010-11-17 21:32:12 <brocktice> Mining bitcoins comes at a cost
1779 2010-11-17 21:32:18 <brocktice> Whether you had the hardware to start with or not
1780 2010-11-17 21:32:26 <brocktice> At the very least it's power + opportunity cost
1781 2010-11-17 21:32:40 <brocktice> If you bought a Radeon HD 5970 to play games on, you can't play games and mine at the same time.
1782 2010-11-17 21:32:58 <brocktice> anyway, enough of that
1783 2010-11-17 21:33:05 <xelister> if some of winers make typo in "digger" we have opportunity to sue them ;)
1784 2010-11-17 21:34:29 <wumpus> brocktice: not that gaming is more profitable than mining, usually :)
1785 2010-11-17 21:34:46 <brocktice> wumpus: no, but you shelled out that money on a card for gaming
1786 2010-11-17 21:35:22 <brocktice> so you will have to reduce gaming if you want to mine more
1787 2010-11-17 21:35:23 <xelister> wumpus: not really, I have a friend that earned up to 300-500 USD playing on guys computers
1788 2010-11-17 21:35:32 <xelister> since she was like 18
1789 2010-11-17 21:35:42 <brocktice> lol
1790 2010-11-17 21:35:44 <xelister> although she had to be naked >_>
1791 2010-11-17 21:35:51 <wumpus> hehe
1792 2010-11-17 21:35:58 <wumpus> that's a different sort of game
1793 2010-11-17 21:36:09 <tcatm> i just noticed that mysql can do something in less than 5ms while my python script needs a few seconds
1794 2010-11-17 21:37:10 <xelister> tcatm: present you with a coding problem that you are unable to solve? ;)
1795 2010-11-17 21:38:11 <wumpus> tcatm: makes sense, databases are optimized for certain cases and it's very hard to beat them there
1796 2010-11-17 21:39:46 <tcatm> it's missing one feature, though
1797 2010-11-17 22:03:27 darksk1ez has quit (Ping timeout: 260 seconds)
1798 2010-11-17 22:27:38 <tcatm> bitcoincharts now calculates everything on demand using sql :)
1799 2010-11-17 23:19:28 <jgarzik> tcatm: pong
1800 2010-11-17 23:19:39 astaroth has quit (Quit: Page closed)
1801 2010-11-17 23:23:58 hippich has quit (Ping timeout: 255 seconds)
1802 2010-11-17 23:25:42 <tcatm> jgarzik: any news on api.bcw?
1803 2010-11-17 23:26:15 altamic has quit (Ping timeout: 276 seconds)
1804 2010-11-17 23:26:18 <jgarzik> tcatm: just switched from sqlite to mysql. it's getting there. need to put in abuse blocks (IP filters)
1805 2010-11-17 23:27:41 <tcatm> is someone abusing your site?!
1806 2010-11-17 23:28:42 kermit has quit (Quit: Leaving.)
1807 2010-11-17 23:28:59 kermit has joined
1808 2010-11-17 23:32:17 <tcatm> how fast will api.bcw be?
1809 2010-11-17 23:32:37 <jgarzik> tcatm: api.bcw doesn't exist yet, so no :) but it is trivial for bots to flood the site and DoS it
1810 2010-11-17 23:33:09 <jgarzik> tcatm: mysql has its advantages and disadvantages. unless I want to move to a multi-VPS reader/writer setup, gotta take some precautions
1811 2010-11-17 23:40:52 <tcatm> i don't like that multiple trades can have the same timestamp at mtgox. after every sql query the order is random and thus the closing price changes
1812 2010-11-17 23:56:26 <xelister> jgarzik: how are you blocking DoS? there are rules to limit amount of connection per IP block to given port
1813 2010-11-17 23:56:28 <davex__> someone on brocktice's link above was complaining about legal issues with converting dollars to bitcoins.
1814 2010-11-17 23:56:30 <davex__> How is that illegal?
1815 2010-11-17 23:57:47 <xelister> davex__: dunno. how is smoking a poted plant in own home illegal? now tell that to DEA.
1816 2010-11-17 23:58:10 <xelister> well this is more about: if usa gov wants, they will fuck you over "because they can"
1817 2010-11-17 23:58:20 <davex__> of course, i just don't think there's any law at present
1818 2010-11-17 23:58:30 <xelister> about actuall legality, seing how there are laws and restrictions for every fucking thibgs
1819 2010-11-17 23:58:44 <xelister> and even people are geting fines or jail for cursing online recently
1820 2010-11-17 23:59:19 <xelister> it seems liklelly usa gov would to also controll most important thing - money
1821 2010-11-17 23:59:22 <xelister> in every way possible
1822 2010-11-17 23:59:31 <davex__> yes. but bitcoin makes that difficult
1823 2010-11-17 23:59:32 <xelister> but ofcourse its just my guess - im not a lawyer ;)
1824 2010-11-17 23:59:34 <davex__> as far as i can tell