1 2011-01-13 00:00:14 <ArtForzZz> 3rd party IP
  2 2011-01-13 00:00:38 <ArtForzZz> I think the open radeon -dev tree already has prelim support for 68xx
  3 2011-01-13 00:00:40 <lfm> to embarassed to let people see shit code
  4 2011-01-13 00:00:47 <ArtForzZz> yep
  5 2011-01-13 00:00:52 <xelister> are they stupid, they don't see how GPU computing community would love them for that? hint hint, scientists/hackers/security experts/advanced programming/math guys usually do not use crap like windows or mac
  6 2011-01-13 00:01:06 <ArtForzZz> real coders use CAL IL
  7 2011-01-13 00:01:13 <xelister> :O
  8 2011-01-13 00:01:19 <xelister> FUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
  9 2011-01-13 00:01:44 <xelister> ArtForzZz: what third party IP.. they bought Ati
 10 2011-01-13 00:01:57 <xelister> also what, they can sell but can not sell the code to like FOSS ?
 11 2011-01-13 00:02:01 <ArtForzZz> and ATI licensed that crap from other companies
 12 2011-01-13 00:02:03 <xelister> or to FSF or smth
 13 2011-01-13 00:02:19 <ArtForzZz> also probably fear of patent suits
 14 2011-01-13 00:02:25 <xelister> they need a serious bitchslap in their fucking head to come to their senses
 15 2011-01-13 00:02:53 <tcatm> luke-jr: okay, but from now in the spec won't change anymore
 16 2011-01-13 00:02:53 <xelister> in the end, they will loose money for not publishing drviers that would open way to be loved <3 obvious choice amongst all cluster computing users
 17 2011-01-13 00:02:54 <ArtForzZz> 3D rendering is a patent minefield nearly as bad as video encoding
 18 2011-01-13 00:03:18 <xelister> well then just release it in EU or something, USA can suck our collective dick
 19 2011-01-13 00:03:21 <lfm> ati was hardware outfit, their drivers always sucked even since ati wonder bios where I had to reset 5 times to get it to start
 20 2011-01-13 00:03:26 <luke-jr> tcatm: I'm making one final change I think
 21 2011-01-13 00:03:53 <xelister> luke-jr: haha yea I rememver All-In-Wonder cards. they REALLY MADE ME WONDER how such crap can crash my game each 15 minutes
 22 2011-01-13 00:04:04 <luke-jr> tcatm: I am modifying it such that X<digits> is always required in the URI, but that implementations should handle the X8/X4 defaults
 23 2011-01-13 00:04:25 <xelister> only good thing from Ati, is that we can say, that linux driver support can be actually BETTER then on windows ;)
 24 2011-01-13 00:04:51 <tcatm> luke-jr: does that change the code for parseAmount?
 25 2011-01-13 00:05:10 <ArtForzZz> ?
 26 2011-01-13 00:05:37 <ArtForzZz> their windows drivers more-or-less work for accelerated video decoding
 27 2011-01-13 00:05:37 <luke-jr> tcatm: not since the X4 change, no
 28 2011-01-13 00:05:47 <tcatm> luke-jr: ok
 29 2011-01-13 00:05:48 <ArtForzZz> which can't be said of fglrx
 30 2011-01-13 00:05:48 <luke-jr> m[9] ? Math.pow(16, parseInt(m[9], 16)) : 0x10000
 31 2011-01-13 00:05:52 <luke-jr> that is the last changed line
 32 2011-01-13 00:06:02 <xelister> wikileaks should just leak the code that may or may not be the actuall code from Ati, and Ati should release statement they will not sue for using it no matter is it their code or not ... then probably they are off the hook?
 33 2011-01-13 00:06:22 <tcatm> luke-jr: https://github.com/tcatm/bitcoin-js-remote/commit/c27a3149aeaa5038e28ecd1d09310dc70587aea7
 34 2011-01-13 00:06:26 <luke-jr> ArtForzZz: the free drivers are almost supporting video decoding and even DirectX
 35 2011-01-13 00:06:34 <xelister> "it's not our code that is infiringing imaginary patents - it is some anon's driver"
 36 2011-01-13 00:07:01 <ArtForzZz> wtf does "almost" supporting video decoding do?
 37 2011-01-13 00:07:33 <luke-jr> ArtForzZz: it's a WIP, and should work soon ;P
 38 2011-01-13 00:07:42 <ArtForzZz> wonder how well that'll work
 39 2011-01-13 00:08:03 <ArtForzZz> iirc the problem is AMd can't realse docs on UVD because it's pretty tightly coupled to HDCP crap
 40 2011-01-13 00:08:03 <lfm> means it wont lock up more than once per hour?
 41 2011-01-13 00:08:29 <luke-jr> oh
 42 2011-01-13 00:08:33 <luke-jr> I think it's more along the lines of vdpau
 43 2011-01-13 00:08:48 <lfm> sigh, oh ya hdcp, thatd fould em up
 44 2011-01-13 00:08:56 <luke-jr> tcatm: it would be much easier to contribute if you use gitorious or something free :
 45 2011-01-13 00:09:05 <ArtForzZz> I think I read about the effort, dont they pretty much emulate that using shaders?
 46 2011-01-13 00:09:20 <luke-jr> tcatm: do you need my help with fonts?
 47 2011-01-13 00:09:25 <tcatm> luke-jr: fonts?
 48 2011-01-13 00:09:30 <ArtForzZz> = once again, DRM crazyness annoying consumers
 49 2011-01-13 00:09:40 <luke-jr> tcatm: yeah, you probably need a webfont or something, for most phones
 50 2011-01-13 00:09:43 <xelister> heh HDCP is shite
 51 2011-01-13 00:09:50 <ArtForzZz> not to mention HDCP is completely. fucking. pointless.
 52 2011-01-13 00:09:55 <tcatm> luke-jr: on android it looks fine
 53 2011-01-13 00:10:00 <xelister> it makes me regret buying any original and never do that mistake again
 54 2011-01-13 00:10:01 <luke-jr> tcatm: tried a TBC amount?
 55 2011-01-13 00:10:04 <lfm> digit restrictions mismanagement
 56 2011-01-13 00:10:11 <xelister> I was like... should I buy or pirate?
 57 2011-01-13 00:10:16 <tcatm> luke-jr: yep
 58 2011-01-13 00:10:28 <xelister> hmm.. rememer the last time I played LEGAL MOVIE and was forced to watch like 10 commercialls with no-skip? fuuuuuuuuu
 59 2011-01-13 00:10:31 <luke-jr> tcatm: (39*4194304).formatBitcoin()
 60 2011-01-13 00:10:36 <ArtForzZz> most fun part: HDCP doesnt even stop someone from capturing the uncompressed stream
 61 2011-01-13 00:10:46 <luke-jr> tcatm: try amount=163577856X0
 62 2011-01-13 00:11:04 <tcatm> luke-jr: Oh, i'm not using that formatBitcoin yet as it would need a little more work. I'm converting everything to BTC for now
 63 2011-01-13 00:11:09 <ArtForzZz> so instead of grabbing the DVI pairs between computer and display I grab the *same* data off the LVDS link between display controller board and panel
 64 2011-01-13 00:11:15 <luke-jr> tcatm: k, that avoids the font issue :p
 65 2011-01-13 00:11:27 <xelister> ArtForzZz: btw the keys for HDCP are long-leaked
 66 2011-01-13 00:11:32 <xelister> as you probbl know
 67 2011-01-13 00:11:34 <ArtForzZz> yep
 68 2011-01-13 00:11:41 <ArtForzZz> but even without that it's fucking pointless
 69 2011-01-13 00:11:52 <lfm> art that voids your warantee! oh dear!
 70 2011-01-13 00:11:56 <ArtForzZz> they built a locked gate without a fence
 71 2011-01-13 00:12:08 <xelister> ;)
 72 2011-01-13 00:12:30 <ArtForzZz> yeah, the lock is now broken, but there never was a fucking fence
 73 2011-01-13 00:12:30 <tcatm> luke-jr: can you develop iPhone apps?
 74 2011-01-13 00:12:35 <xelister> how exactly would you catch it without the keys?
 75 2011-01-13 00:12:45 <luke-jr> tcatm: no idea, no interest since you can't write free apps for iPhone
 76 2011-01-13 00:12:49 <lfm> xelister inside the TV
 77 2011-01-13 00:13:09 <ArtForzZz> well, you always have the analog hole
 78 2011-01-13 00:13:14 <xelister> lfm: so inside TV there is a blackbox decoder, and from it clear line path in LVDS format to display?
 79 2011-01-13 00:13:26 <ArtForzZz> yep
 80 2011-01-13 00:13:57 <luke-jr> but you can't rip it compressed that way
 81 2011-01-13 00:13:59 <lfm> prolly even has a dip plug or edge connector to tap into
 82 2011-01-13 00:14:01 <ArtForzZz> clock, 4 pairs  each R/G/B pixel data, h/vsync
 83 2011-01-13 00:14:01 <xelister> ArtForzZz: all anti pirace is stupid, eventually I can capture it off my fucking eye - like stick HDTV camera in front of panel (and dont give crap about lost details)
 84 2011-01-13 00:14:04 <luke-jr> you have to recompress it, which is usually lossy
 85 2011-01-13 00:14:15 <ArtForzZz> yes
 86 2011-01-13 00:14:27 <lfm> luke rip[ it uncompressed do your own compression
 87 2011-01-13 00:14:31 <ArtForzZz> but thats the thing, HDCP protects the uncompressed stream
 88 2011-01-13 00:14:36 <Diablo-D3> [08:10:36] <ArtForzZz> most fun part: HDCP doesnt even stop someone from capturing the uncompressed stream
 89 2011-01-13 00:14:44 <Diablo-D3> it also doesnt stop them from unecrypting it
 90 2011-01-13 00:15:00 <xelister> HDCP doesn't stop from anything
 91 2011-01-13 00:15:05 <ArtForzZz> thats the point
 92 2011-01-13 00:15:09 <Diablo-D3> no I mean
 93 2011-01-13 00:15:12 <ArtForzZz> it's DRM without purpose
 94 2011-01-13 00:15:12 <Diablo-D3> HDCP was broken
 95 2011-01-13 00:15:25 <xelister> well, perhaps from cops to say  OHHH NOO  that PEDORIST is haxoring the copyright, protect them, fire!
 96 2011-01-13 00:15:28 <ArtForzZz> HDCPs DESIGN was already broken
 97 2011-01-13 00:15:28 <luke-jr> tcatm: can we kill x-btc yet? :D
 98 2011-01-13 00:15:32 <Diablo-D3> ArtForzZz: well
 99 2011-01-13 00:15:38 <Diablo-D3> COPYING the stream is fine and all
100 2011-01-13 00:15:48 <tcatm> luke-jr: We'll simple not use it :)
101 2011-01-13 00:15:55 <ArtForzZz> actually thats what HDCP was supposed to protect against
102 2011-01-13 00:15:56 <Diablo-D3> but who wants to download a torrent of 9000gb of video?
103 2011-01-13 00:15:56 <xelister> ArtForzZz: it though stops technicall illeterates from coping. They can not even set VCR clock
104 2011-01-13 00:16:15 <Diablo-D3> it was supposed to stop digital ripping out of the DVI jack
105 2011-01-13 00:16:22 <ArtForzZz> honestly, the loss isn't that bad
106 2011-01-13 00:16:28 <xelister> but not warez group, so it leaks to torrent, and there almos everyone can get it
107 2011-01-13 00:16:36 <xelister> btw I know few cops, they pirate warez like hell
108 2011-01-13 00:16:43 <ArtForzZz> only thing you lose over "normal" recoding is that you do 2 color space conversions
109 2011-01-13 00:16:47 <xelister> stupid dead laws, they should be removed
110 2011-01-13 00:16:57 <lfm> ya blue ray keys that they cant protect either
111 2011-01-13 00:16:59 <Diablo-D3> ArtForzZz: HDCP is lossless though
112 2011-01-13 00:17:03 <Diablo-D3> well
113 2011-01-13 00:17:04 <Diablo-D3> generally
114 2011-01-13 00:17:09 <Diablo-D3> DRM is fucked
115 2011-01-13 00:17:11 <ArtForzZz> so?
116 2011-01-13 00:17:13 <Diablo-D3> sony _almost_ had it
117 2011-01-13 00:17:16 <Diablo-D3> on the ps3
118 2011-01-13 00:17:25 <Diablo-D3> but why the fuck build a hypervisor that doesnt employ nx?
119 2011-01-13 00:17:26 <xelister> a robbery in MY COPYRIGHT PRIV KEYS???  more likelly then you think
120 2011-01-13 00:17:30 <ArtForzZz> not really
121 2011-01-13 00:17:38 <ArtForzZz> locked down hardware != DRM
122 2011-01-13 00:17:41 <brocktice> Here's a problem:
123 2011-01-13 00:17:54 <Diablo-D3> and why the fuck use ECDSA keys where the random nonce is used more than once?
124 2011-01-13 00:17:58 <ArtForzZz> it's the reverse of DRM
125 2011-01-13 00:18:08 <brocktice> how do you securely encrypt something, but ship millions of devices that can decrypt it without giving away the secret key?
126 2011-01-13 00:18:18 <Diablo-D3> brocktice: burn it to hardware
127 2011-01-13 00:18:20 <ArtForzZz> you can't
128 2011-01-13 00:18:25 <brocktice> exactly
129 2011-01-13 00:18:26 <Diablo-D3> which IBM did.
130 2011-01-13 00:18:32 <Diablo-D3> the problem is
131 2011-01-13 00:18:36 <brocktice> They're fucked before they even start.
132 2011-01-13 00:18:38 <Diablo-D3> the software part of the chain could be violated
133 2011-01-13 00:18:53 <ArtForzZz> and given enough time and resources you could probably crack a IBM crypto module, too
134 2011-01-13 00:19:04 <Diablo-D3> the _only_ thing that fucked sony was their security hypervisor allowed bugs to be used
135 2011-01-13 00:19:05 <brocktice> Oh sure
136 2011-01-13 00:19:17 <brocktice> Diablo-D3: The only thing that fucked them to this point.
137 2011-01-13 00:19:19 <lfm> they can only try to hide keys deep in firmware/hardware and hope no one figures out how to get at it too soon
138 2011-01-13 00:19:23 <ArtForzZz> yep
139 2011-01-13 00:19:28 <brocktice> If that hadn't been discovered, there would have been something else.
140 2011-01-13 00:19:37 <Diablo-D3> brocktice: yes, but no one would have known they had broken ECDSA keys without it
141 2011-01-13 00:19:42 <brocktice> caveat: I'm not a sony ps3 DRM expert.
142 2011-01-13 00:19:46 <Diablo-D3> they used a buffer overflow exploit to inject their own code in
143 2011-01-13 00:19:52 <Diablo-D3> NX would have stopped that dead.
144 2011-01-13 00:19:54 <brocktice> The DRM just keeps it from using unsigned software right?
145 2011-01-13 00:19:58 <Diablo-D3> yes
146 2011-01-13 00:20:04 <brocktice> That's different from trying to encrypt media
147 2011-01-13 00:20:04 <Diablo-D3> its just a security signature
148 2011-01-13 00:20:08 <ArtForzZz> yep
149 2011-01-13 00:20:09 <Diablo-D3> not true, brocktice
150 2011-01-13 00:20:10 <lfm> prolly no one would have tried to break ps3 if they had keept supporting linux on ps3
151 2011-01-13 00:20:14 <ArtForzZz> it's DRM-in-reverse
152 2011-01-13 00:20:14 <brocktice> Yeah much like trying to root android phones
153 2011-01-13 00:20:31 <brocktice> The stock bootloaders are finicky about signatures on updates.
154 2011-01-13 00:20:44 <brocktice> But, people find a way around.
155 2011-01-13 00:20:49 <Diablo-D3> the software part of the bootloader chain is AES encrypted
156 2011-01-13 00:21:03 <Diablo-D3> the IBM bootrom has the key to unlock the first stage of the bootloader
157 2011-01-13 00:21:24 <Diablo-D3> and then the first stage unlocks the second stage, and the second stage can insert and remove keys as needed, via updates
158 2011-01-13 00:21:26 <brocktice> Well, it did buy them a few years, right?
159 2011-01-13 00:21:32 <Diablo-D3> thats all they have to do
160 2011-01-13 00:21:33 <brocktice> that was probably all they expected.
161 2011-01-13 00:21:34 <Diablo-D3> is buy a few years
162 2011-01-13 00:21:42 <ArtForzZz> so it takes a while to find someone with too much time and access to a FIB station
163 2011-01-13 00:21:52 <Diablo-D3> the number of years they need to buy needs to exceed the numbers of years they intend on selling it
164 2011-01-13 00:22:05 <ArtForzZz> you can't securely store anything in a chip really
165 2011-01-13 00:22:05 <Diablo-D3> which if the PS2 was any indication, it would have been ten years
166 2011-01-13 00:22:11 <Diablo-D3> ArtForzZz: no
167 2011-01-13 00:22:13 <Diablo-D3> the chip didnt fail
168 2011-01-13 00:22:16 <Diablo-D3> the software did.
169 2011-01-13 00:22:17 <ArtForzZz> I know
170 2011-01-13 00:22:22 <xelister> ArtForzZz: even trusted computing chip got facked
171 2011-01-13 00:22:25 <Diablo-D3> so Im not going to fault IBM here
172 2011-01-13 00:22:28 <ArtForzZz> but guess how a lot of sat smart cards get cracked?
173 2011-01-13 00:22:37 <Diablo-D3> well, they start attacking the chips there
174 2011-01-13 00:22:40 <ArtForzZz> yep
175 2011-01-13 00:22:43 <Diablo-D3> IBM however knows how to build this shit properly
176 2011-01-13 00:22:52 <Diablo-D3> they have employed similar shit on milspec harware
177 2011-01-13 00:22:53 <luke-jr> what happens when they figure out how to combine verify/decrypt?
178 2011-01-13 00:22:54 <brocktice> I see no need for a PS3
179 2011-01-13 00:22:55 <Diablo-D3> its generally shit proof
180 2011-01-13 00:22:56 <ArtForzZz> you can't protect a chip alone from offline attacks
181 2011-01-13 00:23:02 <Diablo-D3> luke-jr: er, they do
182 2011-01-13 00:23:04 <ArtForzZz> it's physically impossible
183 2011-01-13 00:23:16 <Diablo-D3> ArtForzZz: you "can"
184 2011-01-13 00:23:17 <xelister> ArtForzZz: it can scream for help
185 2011-01-13 00:23:18 <ArtForzZz> you NEED a independent power source
186 2011-01-13 00:23:26 <brocktice> ArtForzZz: unless you build some kind of mechanical self-destruct.
187 2011-01-13 00:23:27 <Diablo-D3> one arcade machine did it for the lulz
188 2011-01-13 00:23:30 <ArtForzZz> how can you scream when you have ... no power?
189 2011-01-13 00:23:31 <brocktice> But even that could be protected maybe
190 2011-01-13 00:23:34 <xelister> ArtForzZz: it can scream for help using micro atomic powerplant
191 2011-01-13 00:23:36 <brocktice> err, worked around
192 2011-01-13 00:23:38 <Diablo-D3> they had an encrypted rom chip set
193 2011-01-13 00:23:45 <luke-jr> Diablo-D3: key failure = meltdown?
194 2011-01-13 00:23:48 <xelister> or, just
195 2011-01-13 00:23:54 <xelister> give 1 guard with each device
196 2011-01-13 00:23:55 <Diablo-D3> the key was stored in _ram_ on the cpu
197 2011-01-13 00:23:59 <ArtForzZz> which is exactly what the ibm crypto modules do, they have their own internal battery
198 2011-01-13 00:24:00 <Diablo-D3> the cpu had a built in battery
199 2011-01-13 00:24:01 <luke-jr> IIRC any recent SIM card destroys itself if you try to clone it
200 2011-01-13 00:24:04 <xelister> want to watch TV @ home -> guard comes in with the TV
201 2011-01-13 00:24:07 <Diablo-D3> unplug the chip
202 2011-01-13 00:24:11 <Diablo-D3> the battery is also unplugged
203 2011-01-13 00:24:11 <xelister> commencing entertainment time!
204 2011-01-13 00:24:24 <ArtForzZz> yes, they try to play games with light-sensitive transistors on modern chips to deter decapping
205 2011-01-13 00:24:25 <Diablo-D3> its a very old trick
206 2011-01-13 00:24:34 <Diablo-D3> theres tons of fucking shit to do
207 2011-01-13 00:24:38 <ArtForzZz> yep
208 2011-01-13 00:24:44 <xelister> but eventually the hacking wins
209 2011-01-13 00:24:46 <xelister> ?
210 2011-01-13 00:24:48 <ArtForzZz> and tons of ways to bypass it
211 2011-01-13 00:24:48 <Diablo-D3> the IBM cells work well here because the initial part of the chain is on the chip
212 2011-01-13 00:24:49 <Diablo-D3> as in
213 2011-01-13 00:24:54 <Diablo-D3> you can no longer boot the chip at all
214 2011-01-13 00:25:05 <Diablo-D3> it doesnt know what to do when you turn it on
215 2011-01-13 00:25:23 <Diablo-D3> it'd be like starting a normal x86 box with a blank bios
216 2011-01-13 00:25:26 <ArtForzZz> iirc the cell also has the "keys to the kingdom" in ROM
217 2011-01-13 00:25:33 <xelister> you can have chip auto-destroy when it looses power reserves?
218 2011-01-13 00:25:40 <Diablo-D3> ArtForzZz: no
219 2011-01-13 00:25:46 <Diablo-D3> only do the first stage bootloader
220 2011-01-13 00:25:51 <ArtForzZz> so?
221 2011-01-13 00:25:54 <Diablo-D3> *to
222 2011-01-13 00:26:02 <ArtForzZz> aaand?
223 2011-01-13 00:26:11 <Diablo-D3> and the bootloader cant read the keys from the cpu.
224 2011-01-13 00:26:17 <ArtForzZz> huh?
225 2011-01-13 00:26:33 <Diablo-D3> theres no way to access the keys inside the cell from anything.
226 2011-01-13 00:27:02 <lfm> hardware debuggers?
227 2011-01-13 00:27:06 <Diablo-D3> lfm: nope
228 2011-01-13 00:27:10 <ArtForzZz> lower level
229 2011-01-13 00:27:32 <Diablo-D3> lfm: the cell also prevents you from tampering with sony's security hypervisor that runs in the one SPE
230 2011-01-13 00:27:37 <Diablo-D3> you cant hardware debug it
231 2011-01-13 00:28:12 <ArtForzZz> if someone can build it, someone can take it apart
232 2011-01-13 00:28:18 <Diablo-D3> the problem is
233 2011-01-13 00:28:27 <Diablo-D3> we no longer have to crack the cell
234 2011-01-13 00:28:31 <Diablo-D3> the bootloader chain is broken
235 2011-01-13 00:28:35 <lfm> i spoze they disable hardware debug on consumer chips, then we have to wait for insider leaks
236 2011-01-13 00:28:39 <ArtForzZz> yep, it's broken
237 2011-01-13 00:28:46 <Diablo-D3> lfm: no, theres no insider leak
238 2011-01-13 00:28:57 <Diablo-D3> lfm: once the SPE goes into supervisor mode, it cant be taken out
239 2011-01-13 00:29:01 <lfm> someone has a backup of the key
240 2011-01-13 00:29:02 <ArtForzZz> it might take a few years and a few M$, but it's certainly possible to RE hardware
241 2011-01-13 00:29:09 <Diablo-D3> lfm: its not a key
242 2011-01-13 00:29:16 <Diablo-D3> lfm: it simply ignores hardware debug commands
243 2011-01-13 00:29:29 <Diablo-D3> the hardware is designed this way
244 2011-01-13 00:29:30 <ArtForzZz> unless you find the FF storing the "SPE in supervisor mode" bit, create a contact with a FIB and bond it out
245 2011-01-13 00:29:44 <lfm> diablo how do they make new versions of firmware? someone has a key dont they?
246 2011-01-13 00:29:45 <Diablo-D3> ArtForzZz: well, you'd have to open the chip
247 2011-01-13 00:29:50 <ArtForzZz> err... yes
248 2011-01-13 00:29:53 <Diablo-D3> lfm: depends which part
249 2011-01-13 00:29:54 <ArtForzZz> thats generally how you attack crypto processors
250 2011-01-13 00:30:15 <Diablo-D3> lfm: sony cant update the first part of the bootloader without IBM
251 2011-01-13 00:30:19 <Diablo-D3> lfm: and they havent updated it
252 2011-01-13 00:30:25 <Diablo-D3> and updating it wont fix the problem
253 2011-01-13 00:30:47 <lfm> anything they can update we can use to boot our stuff if it is leaked
254 2011-01-13 00:30:54 <Diablo-D3> yes
255 2011-01-13 00:31:04 <Diablo-D3> the real problem is
256 2011-01-13 00:31:08 <ArtForzZz> ;;bc,mtgox
257 2011-01-13 00:31:08 <gribble> {"ticker":{"high":0.3996,"low":0.3176,"vol":39878,"buy":0.33,"sell":0.3996,"last":0.3996}}
258 2011-01-13 00:31:09 <Diablo-D3> to resecure this
259 2011-01-13 00:31:16 <Diablo-D3> we need to replace every ps3 AND every printed game
260 2011-01-13 00:31:25 <Diablo-D3> because we now know the key for game drm
261 2011-01-13 00:31:43 <Diablo-D3> because we can decrypt the part of the firmware that manages game drm
262 2011-01-13 00:31:47 <ArtForzZz> yep
263 2011-01-13 00:31:58 <Diablo-D3> all because of _one_ bug
264 2011-01-13 00:31:59 <Diablo-D3> no nx.
265 2011-01-13 00:32:07 <ArtForzZz> iirc we can now decrypt the whole chain down to the 0bl
266 2011-01-13 00:32:19 <Diablo-D3> yup
267 2011-01-13 00:32:41 <xelister> what is the bug against which NX is needed here?
268 2011-01-13 00:32:43 <Diablo-D3> you can make sony's first boot loader stage directly load a new OS
269 2011-01-13 00:32:46 <Diablo-D3> and sony cant stop you
270 2011-01-13 00:32:53 <ArtForzZz> yep
271 2011-01-13 00:33:00 <Diablo-D3> it'd have full access to the hardware
272 2011-01-13 00:33:11 <lfm> replace hypervisor then
273 2011-01-13 00:33:16 <brocktice> osnap
274 2011-01-13 00:33:17 <Diablo-D3> lfm: it never loads
275 2011-01-13 00:33:20 <brocktice> almost hit 0.4 on mtgox
276 2011-01-13 00:33:26 <ArtForzZz> this is the stuff that normally LOADS the hypervisor
277 2011-01-13 00:33:27 <Diablo-D3> the security hypervisor is in the second bootloader stage
278 2011-01-13 00:33:34 <lfm> k
279 2011-01-13 00:33:44 <Diablo-D3> the security hypervisor only does important shit like manage game drm
280 2011-01-13 00:33:57 <Diablo-D3> it doesnt do anything as part of the OS security chain
281 2011-01-13 00:34:01 <ArtForzZz> and iirc it even doesnt do that properly
282 2011-01-13 00:34:04 <Diablo-D3> it just needs to not leak shit
283 2011-01-13 00:34:04 <lfm> and it sandboxed the old linux
284 2011-01-13 00:34:14 <ArtForzZz> yep, and failed at it ;)
285 2011-01-13 00:34:24 <Diablo-D3> lfm: yes, OtherOS ran as a hypervisor guest, same way Game OS does
286 2011-01-13 00:34:36 <Diablo-D3> the hardware managed the SPE protection
287 2011-01-13 00:34:39 <Diablo-D3> so there wasnt a real issue
288 2011-01-13 00:35:08 <Diablo-D3> IBM must be pissed though
289 2011-01-13 00:35:20 <ArtForzZz> why?
290 2011-01-13 00:35:20 <Diablo-D3> at no point did they drop the ball
291 2011-01-13 00:35:38 <Diablo-D3> and they're going to have to end up making it harder for solution vendors to fuck shit up
292 2011-01-13 00:35:44 <ArtForzZz> yep
293 2011-01-13 00:36:00 <Diablo-D3> Sony only fucked up because IBM didnt supply the whole fucking thing
294 2011-01-13 00:36:06 <Diablo-D3> which now they'll essentially have to do
295 2011-01-13 00:36:12 <Diablo-D3> because the Sonys of the world cant be trusted to do it right
296 2011-01-13 00:36:28 <ArtForzZz> I still dont get how you can botch ECDSA so badly
297 2011-01-13 00:36:34 <Diablo-D3> ArtForzZz: the random key
298 2011-01-13 00:36:35 <Diablo-D3> er
299 2011-01-13 00:36:39 <Diablo-D3> the random part of the key
300 2011-01-13 00:36:44 <Diablo-D3> k or whatever
301 2011-01-13 00:36:47 <ArtForzZz> yeah, the nonrandom nonce
302 2011-01-13 00:36:48 <Diablo-D3> its supposed to be a nonce
303 2011-01-13 00:36:51 <Diablo-D3> it wasnt.
304 2011-01-13 00:37:03 <ArtForzZz> yep
305 2011-01-13 00:37:03 <Diablo-D3> someone didnt fucking check their shit
306 2011-01-13 00:37:08 <ArtForzZz> it was a nalways
307 2011-01-13 00:37:14 <Diablo-D3> lol
308 2011-01-13 00:37:18 <Diablo-D3> it was the number 4.
309 2011-01-13 00:37:25 <Diablo-D3> chosen fairly by a die.
310 2011-01-13 00:37:39 <ArtForzZz> but serously, why fucking implement ECDSA sig GENERATION yourself?
311 2011-01-13 00:37:50 <Diablo-D3> sometimes I just want to punch someone's blood out
312 2011-01-13 00:37:56 <ArtForzZz> crypto libraries exist for a purpose
313 2011-01-13 00:38:01 <Diablo-D3> ArtForzZz: well
314 2011-01-13 00:38:03 <Diablo-D3> like I said
315 2011-01-13 00:38:10 <ArtForzZz> NIH syndrome I guess
316 2011-01-13 00:38:12 <Diablo-D3> IBM is going to have to handhold on future shit
317 2011-01-13 00:38:31 <Diablo-D3> IBM has a fucking history of doing shit perfectly.
318 2011-01-13 00:38:32 <ArtForzZz> yep
319 2011-01-13 00:38:46 <Diablo-D3> and you know what
320 2011-01-13 00:38:56 <Diablo-D3> I'd even buy into a trusted platform of IBM's if I needed one
321 2011-01-13 00:39:09 <Diablo-D3> there is huge security implications in not having shit that breaks at a drop of a hat
322 2011-01-13 00:39:14 <ArtForzZz> well, they generally do what they're supposed to
323 2011-01-13 00:39:26 <xelister> TRADE|                         50 @ $0.395
324 2011-01-13 00:39:28 <xelister> huhuhu
325 2011-01-13 00:39:55 <ArtForzZz> though I had to implement some ugly hacks for enterprise IBM hardware
326 2011-01-13 00:40:25 <lfm> ya IBM was pretty evil when they were on top with 370 etc
327 2011-01-13 00:41:15 <lfm> like ebcidic! really!
328 2011-01-13 01:24:05 <luke-jr> is Genjix ever here?
329 2011-01-13 01:25:46 <gribble> genjix was last seen in #bitcoin-dev 3 days, 4 hours, 38 minutes, and 3 seconds ago: <genjix> ic
330 2011-01-13 01:25:46 <nanotube> ;;seen genjix
331 2011-01-13 01:25:53 <nanotube> yes
332 2011-01-13 01:28:09 <EvanR> "if you hire a hooker and pay her in bitcoins, is it technically hooking"
333 2011-01-13 01:29:34 <lfm> if you hire a hooker and pay here in buttons, is it technicly protitution?
334 2011-01-13 01:29:45 <EvanR> yes, the next question
335 2011-01-13 01:30:20 <lfm> same thing
336 2011-01-13 01:30:30 <EvanR> ok now whats the answer ;)
337 2011-01-13 01:30:44 <lfm> yes
338 2011-01-13 01:30:55 <EvanR> what about legally
339 2011-01-13 01:31:28 <EvanR> bitcoin legal advise, 10000BTC per hour
340 2011-01-13 01:31:40 <lfm> prolly depends on local law. they try not to prosecute husbands that put their wife on an allowance
341 2011-01-13 01:31:55 <EvanR> lol
342 2011-01-13 01:32:42 <luke-jr> lfm: that is clearly not prostitution, since the wife retains the marital debt even if she is given no allowance
343 2011-01-13 01:33:07 <EvanR> marriage is like grand prostitution
344 2011-01-13 01:33:24 <lfm> but if the husband cuts off the allowance if the wife cuts off the nookie, is that prostitution?
345 2011-01-13 01:33:41 <luke-jr> wife can't "cut off the nookie"
346 2011-01-13 01:33:53 <lfm> depends on the jurisdiction
347 2011-01-13 01:33:58 <luke-jr> no
348 2011-01-13 01:34:16 <luke-jr> a wife is bound under natural law to have relations so long as the husband's request is reasonable
349 2011-01-13 01:34:19 <luke-jr> (and vice versa)
350 2011-01-13 01:34:30 <lfm> luke-jr we are talking civl law here, not biblical law
351 2011-01-13 01:34:35 <luke-jr> natural law applies to all
352 2011-01-13 01:34:50 <lucky> huh?
353 2011-01-13 01:34:53 <lucky> says who? that's stupid.
354 2011-01-13 01:34:57 <lfm> you are deluded Luke
355 2011-01-13 01:35:05 <luke-jr> that's part of the nature of marriage
356 2011-01-13 01:35:14 <lfm> in your dreams
357 2011-01-13 01:35:57 <lucky> ... says who?
358 2011-01-13 01:36:17 <luke-jr> lucky: the infallible Catholic Church ofc
359 2011-01-13 01:36:41 <lucky> uh... huh.
360 2011-01-13 01:36:43 <lfm> infallible? dont make us laugh
361 2011-01-13 01:36:59 <luke-jr> lfm: why not? you gonna die if you laugh?
362 2011-01-13 01:40:09 <lfm> luke-jr you make it sound like you are one of those people who think a husband cannot rape his own wife by definition?
363 2011-01-13 01:41:17 <luke-jr> lfm: that is correct, although it does not excuse battery of course; if a wife refuses to meet her obligations, it doesn't justify the husband being abusive
364 2011-01-13 01:42:19 <luke-jr> 2 wrongs don't make a right
365 2011-01-13 01:42:31 <lfm> well luke-jr I guess you have been asleep for the last 30 years or so cuz that would be you only excuse for thinking that way
366 2011-01-13 01:43:10 <luke-jr> lfm: natural law and morality never changes
367 2011-01-13 01:43:15 <nevezen> bitcoin's dev & technical discussion forum is an interesting read.. now only starting on the second page :)
368 2011-01-13 01:43:37 <lfm> morality changes all the time. you just arnt paying attention
369 2011-01-13 01:44:23 <lfm> as for "natural law" there isnt much natural about marriage and other similar human activities
370 2011-01-13 01:50:22 <luke-jr> lfm: nope
371 2011-01-13 01:50:32 <luke-jr> nothing perfect changes
372 2011-01-13 01:52:47 <lfm> luke, you're a hopeless troll. I appologize to the rest of the channel for getting roped into your stupid debate
373 2011-01-13 02:06:23 <kiba> finally
374 2011-01-13 02:06:30 <kiba> now all I need to do is sell my article for bitcoin
375 2011-01-13 02:08:26 <EvanR> too many L names flying around
376 2011-01-13 02:11:13 <luke-jr> lfm: just the facts
377 2011-01-13 02:13:52 <luke-jr> anyhow& anything wrong with this? Mushkin 2x4GB DDR3 1333 + AMD Phenom II X4 955 + Gigabyte GA-890GPA-UD3H
378 2011-01-13 02:21:33 <Keefe> asking whether any of that is not compatible with each other?
379 2011-01-13 02:22:23 <luke-jr> whether it sounds like a good desktop
380 2011-01-13 02:22:36 <luke-jr> $344
381 2011-01-13 02:23:01 <Keefe> that's subjective
382 2011-01-13 02:23:18 <Keefe> for most people, the very cheapest stuff newegg sells is just fine
383 2011-01-13 02:25:19 <Keefe> i.e. 1gb ddr3, sempron 140, cheap ddr3 board
384 2011-01-13 03:18:27 <xelister> TRADE|                       4.18 @ $0.4
385 2011-01-13 03:18:28 <xelister> !!!
386 2011-01-13 03:20:04 <Cusipzzz> painting the tape obv....less than 2 USD :p
387 2011-01-13 03:27:30 <kupo> is it ok to run 0.3.19 ?
388 2011-01-13 03:27:34 <kupo> sorry 0.3.15 ?
389 2011-01-13 03:29:59 <nanotube> kupo: iirc it's fine, no security fixes in .15+ that i remember. but best to run latest if you can.
390 2011-01-13 03:30:27 <kupo> ok ty
391 2011-01-13 03:30:34 <kupo> how do I join these pools btw?
392 2011-01-13 03:30:42 <gribble> No fancy GPU farm, and don't want to wait for months for a block gen? Join the mining pool! http://mining.bitcoin.cz/
393 2011-01-13 03:30:42 <nanotube> see instructions on the ,,pool page
394 2011-01-13 03:31:46 <kupo> ack
395 2011-01-13 03:32:19 <nanotube> have fun. :)
396 2011-01-13 03:35:48 <kiba> so the price of bitcoin reached 0.40 USD
397 2011-01-13 03:35:53 <kiba> wee!
398 2011-01-13 03:36:13 <kiba> trying to sell my interview for bitcoin...
399 2011-01-13 03:36:20 <kiba> interview article...
400 2011-01-13 03:40:48 <kupo> hiya nanotube
401 2011-01-13 03:41:01 <Cusipzzz> 5770 for 120$ good deal?
402 2011-01-13 03:41:17 <Cusipzzz> at these rates, even i may have to get into mining
403 2011-01-13 03:42:21 <nanotube> nameless|: haha
404 2011-01-13 03:42:25 <nanotube> kiba: pong
405 2011-01-13 03:44:09 <nanotube> kupo: sup?
406 2011-01-13 03:45:04 <xelister> kiba: interview about what
407 2011-01-13 03:46:19 <kiba> MT's QBitcoin project
408 2011-01-13 03:46:45 <kupo> nanotube: nvm
409 2011-01-13 03:47:19 <nanotube> k
410 2011-01-13 03:50:20 <grondilu> nice spike in price on mtgox !
411 2011-01-13 03:50:35 <nanotube> yep
412 2011-01-13 03:50:42 <nanotube> it's been all the rage around here. :)
413 2011-01-13 03:52:56 <nanotube> heh
414 2011-01-13 03:54:59 <xelister> RAGGGEEEON
415 2011-01-13 03:55:54 <xelister> rageon 6999. Now with new, better drivers! Don't worry about lack of multi-GUI OpenCL, any OpenCL or Linux support, why bother when you have cool x64 anti-aliasing, woo
416 2011-01-13 03:56:07 <xelister> *gpu
417 2011-01-13 03:56:34 <kiba> yay!
418 2011-01-13 03:56:42 <kiba> a 27% rise in the price of bitcoin in one day!
419 2011-01-13 03:58:14 <newsham> bubble bobble
420 2011-01-13 03:58:38 <newsham> bid/ask spread is huge
421 2011-01-13 03:59:03 <newsham> 0.34 to 0.40
422 2011-01-13 03:59:45 <newsham> 0.34 to 0.414
423 2011-01-13 04:00:11 <Cusipzzz> when people wake up and seethe offer at > .4, they will move bids up...
424 2011-01-13 04:00:40 <kiba> I predict we end up at 0.50
425 2011-01-13 04:00:41 <kiba> tommorow
426 2011-01-13 04:00:54 <nanotube> kiba: whatcha willing to bet? :)
427 2011-01-13 04:01:07 <Cusipzzz> i'll take some of that as well =)
428 2011-01-13 04:01:07 <kiba> no bet
429 2011-01-13 04:01:08 <hacim> 5870's are about $200-250, i bet they are pretty decent still compared to a 5970
430 2011-01-13 04:01:21 <nanotube> kiba: hehe
431 2011-01-13 04:01:40 <newsham> there's a lot of shares up between 0.414 and 0.5
432 2011-01-13 04:01:45 <nanotube> Cusipzzz: that said, there are only 2000+ worth of asks between here and .5
433 2011-01-13 04:02:03 <newsham> at least 2kbtc
434 2011-01-13 04:02:06 <Cusipzzz> 1000 USD...then a brick wall at .51
435 2011-01-13 04:02:35 <nanotube> mm, well, 'brick wall' in the amount of 10kusd. considering recent volumes, not so much.
436 2011-01-13 04:02:42 <Cusipzzz> ya maybe
437 2011-01-13 04:02:51 <nanotube> the main question is... will the miners wake up and be like zomg, maybe now is a good time to sell before the bubble bursts. :)
438 2011-01-13 04:03:18 <Cusipzzz> hope so...voltatility is a good thing...and not straight up :p
439 2011-01-13 04:03:26 <Keefe> if they put in asks around 0.4, it won't affect the market much
440 2011-01-13 04:04:02 <newsham> i dont think 0.4 is the peak of hte bubble
441 2011-01-13 04:04:06 <nanotube> well, depends on if the buying pressure keeps up.
442 2011-01-13 04:04:54 <hacim> he's offline!
443 2011-01-13 04:04:59 <kiba> w00t!
444 2011-01-13 04:05:03 <kiba> we reach 2 million dollars
445 2011-01-13 04:05:13 <kiba> now, we're a 2 million dollars economy!
446 2011-01-13 04:05:19 <nevezen> is there a maximum amount of bitcoins that can be generated at the moment?
447 2011-01-13 04:05:33 <nanotube> nevezen: can you clarify your question?
448 2011-01-13 04:05:34 <kiba> 21 million
449 2011-01-13 04:05:42 <nanotube> nevezen: what do you mean by 'at the moment' ?
450 2011-01-13 04:05:52 <newsham> nev: 300 per hr?
451 2011-01-13 04:05:58 <nevezen> before generating becomes irrelevant, if ever..
452 2011-01-13 04:06:05 <kiba> 6 blocks per hours, translate into 300
453 2011-01-13 04:06:08 <kiba> 300 * 24
454 2011-01-13 04:06:17 <newsham> not quite a maximum (could be a lot more, if a bunch of miners came online quickly before difficulty got adjusted again) but roughly
455 2011-01-13 04:06:20 <Cusipzzz> some blocks were really long today, like 45 mins+
456 2011-01-13 04:06:36 <kiba> around 7200 bitcoins
457 2011-01-13 04:06:54 <nevezen> so what's the current difficulty at the moment? is there a scale?
458 2011-01-13 04:07:01 <nanotube> ;;bc,stats
459 2011-01-13 04:07:03 <gribble> Current Blocks: 102392 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 423 blocks | Next Difficulty In About: 2 days, 12 hours, 51 minutes, and 54 seconds | Next Difficulty Estimate: 18046.79146339
460 2011-01-13 04:07:06 <nanotube> nevezen: --^
461 2011-01-13 04:07:09 <nevezen> ahh
462 2011-01-13 04:07:53 <nevezen> will the difficulty just continue increasing to infinity?
463 2011-01-13 04:08:17 <nanotube> it will continue adjusting to keep block gen to about 1 per 10 min.
464 2011-01-13 04:08:18 <newsham> difficulty is adjusted to target a specific block rate.
465 2011-01-13 04:08:34 <newsham> so if a bunch of miners go offline, it could come down
466 2011-01-13 04:08:44 <newsham> but more likely it will grow as bitcoin popularity grows
467 2011-01-13 04:08:56 <nevezen> oh, so it's not based on the number of current existing blocks
468 2011-01-13 04:09:22 <nanotube> no
469 2011-01-13 04:09:23 <tcatm> http://tcatm.github.com/bitcoin-js-remote/ updated js-remote with QR and bitcoin: URI support
470 2011-01-13 04:09:24 <newsham> its based on how fast people have been mining recently
471 2011-01-13 04:09:28 <nanotube> ;;bc,wiki difficulty
472 2011-01-13 04:09:29 <gribble> https://en.bitcoin.it/wiki/Difficulty | Jan 3, 2011 ... Difficulty is a measure of how difficult it is to find a new block compared to ... Current difficulty, as output by BitCoin's getDifficulty. ...
473 2011-01-13 04:09:53 <nevezen> more time in the wiki then for me :)
474 2011-01-13 04:13:18 <nanotube> tcatm: ooh nice. :)
475 2011-01-13 04:15:07 <hacim> i'm guessing most miners are using 5970's but since everyone who had 5870's switched to 5970's when they came out, those are much cheaper, if half the h/sec
476 2011-01-13 04:15:39 <nanotube> well, you also have to consider hpW
477 2011-01-13 04:15:44 <nevezen> is there a maximum amount of USD bitcoin can hold?
478 2011-01-13 04:15:55 <nanotube> nevezen: bitcoin doesn't hold any usd
479 2011-01-13 04:16:02 <hacim> nanotube: hpW is how much juice its going to cost?
480 2011-01-13 04:16:02 <nanotube> any more than an oz of gold holds usd
481 2011-01-13 04:16:12 <hacim> tcatm: thats pretty sweet!
482 2011-01-13 04:16:12 <nanotube> hacim: yea, hash per watt
483 2011-01-13 04:16:16 <nevezen> I mean if you wanted to spend 10 million dollars, converted to btc
484 2011-01-13 04:16:46 <nanotube> ah, well you can look at current market price to get an idea ,,bc,mtgox
485 2011-01-13 04:16:47 <gribble> {"ticker":{"high":0.405,"low":0.3176,"vol":40439,"buy":0.34,"sell":0.414,"last":0.405}}
486 2011-01-13 04:16:48 <Cusipzzz> it would be difficult to buy that many
487 2011-01-13 04:17:02 <nanotube> nevezen: but you couldn't really buy many at that price before driving up price
488 2011-01-13 04:17:03 <hacim> hum, I'm surprised the hpW aren't on http://174.143.149.98/wiki/doku.php?id=bitcoin_miners
489 2011-01-13 04:17:11 <nevezen> it was just an example
490 2011-01-13 04:17:14 <gribble> https://en.bitcoin.it/wiki/Mining_Hardware_Comparison | Jan 3, 2011 ... Mining Hardware Comparison. From Bitcoin. Jump to: navigation, search. This page is a stub, you can help by expanding it. ...
491 2011-01-13 04:17:14 <nanotube> ;;bc,wiki mining hardware comparison
492 2011-01-13 04:17:18 <nanotube> hacim: --^ they are here
493 2011-01-13 04:17:29 <hacim> oh thats a better page
494 2011-01-13 04:17:36 <x6763> nevezen: http://www.bitcoin.org/sites/default/files/bitcoin.pdf - Satoshi's (bitcoin creator) whitepaper
495 2011-01-13 04:18:04 <hacim> slightly more Mhash/W
496 2011-01-13 04:18:27 <hacim> err, slightly less
497 2011-01-13 04:20:22 <nanotube> look at your electric bill. ;)
498 2011-01-13 04:20:45 <hacim> looks like 19.98cents/kWH
499 2011-01-13 04:20:55 <newsham> see also http://aws.amazon.com/free/
500 2011-01-13 04:21:32 <hacim> so if a 5870 can do 1.668Mhash/Watt and that is my cost of electricity
501 2011-01-13 04:21:41 <hacim> hmm someone must have made a calculator for this already
502 2011-01-13 04:21:55 <hacim> newsham: are people really using amazon for mining?
503 2011-01-13 04:22:48 <newsham> dunno
504 2011-01-13 04:25:12 <nanotube> hacim: not that i know of... but you can calculate for yourself. running it for 24 hours will be ,,(math calc 24*188/1000) kw which will cost you ,,(math calc 24*188/10000*.1998) usd
505 2011-01-13 04:25:42 <gribble> 4.512
506 2011-01-13 04:25:52 <gribble> 0.09014976
507 2011-01-13 04:26:11 <nanotube> er whoops, extra zero in second one. .9014 it will cost per day
508 2011-01-13 04:26:29 <hacim> what is the 188?
509 2011-01-13 04:26:54 <nanotube> 188 watts that the card uses
510 2011-01-13 04:27:17 <nanotube> (see table)
511 2011-01-13 04:27:36 <hacim> oh right, duh
512 2011-01-13 04:27:48 <hacim> ,,(math calc 24*294/10000*.01998)
513 2011-01-13 04:28:02 <gribble> 0.014097888
514 2011-01-13 04:28:39 <grubles> which card are we talking about?
515 2011-01-13 04:28:41 <nanotube> hacim: /1000 not /10000
516 2011-01-13 04:28:50 <hacim> ,,(math calc 24*294/1000*.1998)
517 2011-01-13 04:29:07 <gribble> 1.4097888
518 2011-01-13 04:29:09 <nanotube> grubles: ati 5870
519 2011-01-13 04:29:11 <hacim> err no, thats off again, that should be 14cents/day
520 2011-01-13 04:29:13 <grubles> ah, i have a 5970
521 2011-01-13 04:29:47 <nanotube> grubles: 5970 is 294W stock. so make your calculations accordingly. :)
522 2011-01-13 04:30:16 <nanotube> hacim: no, it's 1.4 usd per day.
523 2011-01-13 04:30:27 <nanotube> to run the 5970 24/7 at your cost of power.
524 2011-01-13 04:30:44 <nanotube> ;;bc,calc 600000
525 2011-01-13 04:30:47 <gribble> The average time to generate a block at 600000 Khps, given current difficulty of 16307.48285682 , is 1 day, 8 hours, 25 minutes, and 33 seconds
526 2011-01-13 04:30:57 <grubles> mines oc'd to 885mhz core with 1.162v so it has to be more than that
527 2011-01-13 04:31:07 <nanotube> hacim: and you'll gen a block about once every 1.5 days, which is worth about 20 usd. so heftily profitable.
528 2011-01-13 04:31:15 <nanotube> grubles: yea probably.
529 2011-01-13 04:31:21 <hacim> actually, i was wrong about my cost of power, its ,,(math calc 24*188/1000*.109362)
530 2011-01-13 04:31:22 <gribble> 0.493441344
531 2011-01-13 04:31:31 <nanotube> nice
532 2011-01-13 04:31:34 <hacim> perday for the 5870
533 2011-01-13 04:32:00 <gribble> 0.771658272
534 2011-01-13 04:32:00 <hacim> and ,,(math calc 24*294/1000*.109362) for the 5970
535 2011-01-13 04:32:09 <grubles> +60000 k/hash/s with the overclock
536 2011-01-13 04:33:02 <hacim> nanotube: a block about once every 1.5 days for the 5870 or the 5970?
537 2011-01-13 04:33:04 <grubles> totaling ~660 m/hash/s
538 2011-01-13 04:34:53 <grubles> i think you misread, its 60k
539 2011-01-13 04:35:25 <nanotube> hacim: it's 600mhps for the 5970, about half that for the 5870
540 2011-01-13 04:35:30 <nanotube> (assuming a bit of overclocking)
541 2011-01-13 04:35:38 <nanotube> haha yea
542 2011-01-13 04:35:51 <grubles> a stock 5970 i think is 500 m/hash/s
543 2011-01-13 04:36:03 <hacim> table says 535.06
544 2011-01-13 04:36:14 <grubles> yeah sounds right
545 2011-01-13 04:36:56 <nanotube> yep
546 2011-01-13 04:37:07 <nanotube> that's stock. people routinely oc it to around 600 though
547 2011-01-13 04:37:21 <nanotube> but ... ,,(bc,calc 535000)
548 2011-01-13 04:37:22 <gribble> The average time to generate a block at 535000 Khps, given current difficulty of 16307.48285682 , is 1 day, 12 hours, 21 minutes, and 56 seconds
549 2011-01-13 04:37:28 <nanotube> still about 1.5 days. :)
550 2011-01-13 04:37:42 <hacim> ,,(bc,calc 313.65)
551 2011-01-13 04:37:43 <gribble> The average time to generate a block at 313.65 Khps, given current difficulty of 16307.48285682 , is 7 years, 4 weeks, 1 day, 13 hours, 36 minutes, and 9 seconds
552 2011-01-13 04:37:49 <hacim> oops
553 2011-01-13 04:37:53 <grubles> lol
554 2011-01-13 04:38:00 <hacim> ,,(bc,calc 313650)
555 2011-01-13 04:38:01 <gribble> The average time to generate a block at 313650 Khps, given current difficulty of 16307.48285682 , is 2 days, 14 hours, 1 minute, and 46 seconds
556 2011-01-13 04:38:12 <hacim> wow, basically 2x
557 2011-01-13 04:38:31 <grubles> with my crossfire 5770s thats how long it would take
558 2011-01-13 04:38:32 <hacim> 2x the time, 2x the Mhash/sec, 2x the money
559 2011-01-13 04:39:06 <hacim> grubles: isn't the 5770's ,,(bc,calc 156830) ?
560 2011-01-13 04:39:09 <gribble> The average time to generate a block at 156830 Khps, given current difficulty of 16307.48285682 , is 5 days, 4 hours, 3 minutes, and 18 seconds
561 2011-01-13 04:39:42 <grubles> basically know that a 5970 has 3200 stream processors, half that for the 5870, half again for the 5770, so 1600 for the 5870 and 800 for the 5770.
562 2011-01-13 04:40:06 <grubles> i had 2 in crossfire :p
563 2011-01-13 04:40:15 <hacim> ah
564 2011-01-13 04:40:22 <hacim> i wonder what people are able to overclock their 5770's at
565 2011-01-13 04:40:31 <hacim> err, 5870s
566 2011-01-13 04:40:57 <grubles> not sure, i've heard 2 5870s in crossfire overclock and perform better than the 5970 though
567 2011-01-13 04:41:12 <nanotube> but also eat more power
568 2011-01-13 04:41:31 <grubles> but i got my 5970 in order to save room for another one to have quadfire
569 2011-01-13 04:41:34 <grubles> indeed
570 2011-01-13 04:42:23 <hacim> what is crossfire/quadfire?
571 2011-01-13 04:42:57 <grubles> having more than one graphics card run in parallel
572 2011-01-13 04:43:03 <lfm> hacim ati multi-gpu system
573 2011-01-13 04:43:09 <hacim> ah
574 2011-01-13 04:43:17 <grubles> you can have 2,3, or 4 cards
575 2011-01-13 04:43:26 <hacim> so i'm guessing everyone is waiting for the next gen card to come out
576 2011-01-13 04:43:37 <LobsterMan> does anyone have any experience overclocking nvidia cards?
577 2011-01-13 04:44:00 <grubles> i think even the 6970 wont be as good as the 5970
578 2011-01-13 04:44:13 <hacim> what about the 6990 amd?
579 2011-01-13 04:44:14 <nanotube> LobsterMan: probably same general rules apply: be careful, monitor temps, etc.
580 2011-01-13 04:44:22 <grubles> in terms of bitcoin mining
581 2011-01-13 04:44:29 <LobsterMan> my temps are fine, but my driver crashed if i took my clock too high
582 2011-01-13 04:44:30 <hacim> http://www.fudzilla.com/graphics/item/21487-amd-to-have-hd-6990-dual-cards-in-february?utm_source=twitterfeed&utm_medium=twitter
583 2011-01-13 04:44:30 <LobsterMan> lol
584 2011-01-13 04:44:45 <grubles> samething happens with ati cards
585 2011-01-13 04:44:47 <nevezen> I use rivatuner..
586 2011-01-13 04:44:50 <nanotube> hacim: i think none of the new cards will be better mining wise... that's what i've been hearing
587 2011-01-13 04:44:56 <grubles> probably means your voltage is too low
588 2011-01-13 04:44:56 <LobsterMan> i'm using rivatuner too
589 2011-01-13 04:44:58 <hacim> interesting
590 2011-01-13 04:45:07 <hacim> nanotube: what are you running?
591 2011-01-13 04:45:11 <nevezen> but I max out my fanspeed..
592 2011-01-13 04:45:14 <LobsterMan> i've got 2 gtx275's running stock @ 684mhz, i took them to 760mhz and my driver crashed
593 2011-01-13 04:45:27 <nanotube> hacim: mining wise - nothing. my comp is a laptop.
594 2011-01-13 04:45:45 <nanotube> LobsterMan: so... up the voltage if you can... or just don't take them up so high. :)
595 2011-01-13 04:45:49 <grubles> lobsterman, have you increased the voltage for your cards
596 2011-01-13 04:46:08 <LobsterMan> can i do that in rivatuner? or would i need to do that in the bios?
597 2011-01-13 04:47:00 <nevezen> I think rivatuner can do it..
598 2011-01-13 04:47:25 <LobsterMan> my temps are fine, even at 760MHz (around 82), but i guess something broke because my display went black for a few secs and then i got a warning that the driver crashed but it had recovered, then my cards went into some low power mode until i rebooted...
599 2011-01-13 04:47:40 <grubles> lobsterman, did you just set it 760 or gradually get to that speed
600 2011-01-13 04:48:26 <nevezen> is 100'C too high for an 8800GT?
601 2011-01-13 04:48:27 <LobsterMan> i went from 684 to like 715, then 730, then 760
602 2011-01-13 04:48:47 <LobsterMan> "Display driver nvlddmkm stopped responding and has successfully recovered."
603 2011-01-13 04:49:10 <grubles> yeah thats what happens usually when your overclock is unsuccessful
604 2011-01-13 04:49:14 <lfm> i think 100 c is like the limit for most any chip
605 2011-01-13 04:49:19 <grubles> or you get the blue screen
606 2011-01-13 04:49:31 <LobsterMan> i'm running at 733 now and all seems ok
607 2011-01-13 04:49:39 <LobsterMan> like i said my temps were fine at 760 so i thought it was ok
608 2011-01-13 04:49:41 <grubles> yeah 760 might just be too high
609 2011-01-13 04:49:56 <grubles> if you havent increased the volts
610 2011-01-13 04:50:18 <hacim> nanotube: was wondering because I'm only running linux, and not sure if I can overclock without windows
611 2011-01-13 04:50:30 <hacim> i'm only on a laptop myself, but I have a old shuttle pc sitting here off...
612 2011-01-13 04:50:38 <nanotube> hacim: ah well, i'm only running linux too. but i don't really have any tips for you overclocking-wise. :)
613 2011-01-13 04:50:48 <lfm> hacim how old?
614 2011-01-13 04:50:53 <grubles> hacim, what graphics card do you have
615 2011-01-13 04:50:57 <nanotube> nevezen: from what i hear... yes 100c is too high. you should aim for the 70s
616 2011-01-13 04:51:24 <LobsterMan> nvidia cards run kind of hot
617 2011-01-13 04:51:31 <hacim> grubles: i have an ancient card (geforce fx5700), wont even work for a gpu miner... but I am looking at getting a 5870 for about $150
618 2011-01-13 04:51:36 <nevezen> mine is currently at 92, stock speed but the fan is on 100%
619 2011-01-13 04:51:48 <hacim> lfm: the shuttle is a bit old, i'm not sure if it would even take the card
620 2011-01-13 04:51:59 <nanotube> hacim: ooh, 150, that's way less than half the price. may make sense to get two of those rather than one 5970
621 2011-01-13 04:52:06 <nanotube> (or even one of those)
622 2011-01-13 04:52:07 <nevezen> I think the minimum requirement for CUDA is an 8800GT
623 2011-01-13 04:52:13 <nevezen> or openCL
624 2011-01-13 04:52:32 <lfm> hacim ya if its one of those real small shuttle boxes it might be too small for many cards
625 2011-01-13 04:52:40 <hacim> thats what i am worried about
626 2011-01-13 04:52:46 <hacim> nanotube: yeah, not a bad price... but only one avail
627 2011-01-13 04:52:58 <hacim> lfm: it is one of the small ones unfortunately
628 2011-01-13 04:53:01 <LobsterMan> my card is one of those "oc" cards that came stock with a slightly higher clock speed, i guess that should be taken into account too
629 2011-01-13 04:53:15 <grubles> a 5870 might be too big for a shuttle case
630 2011-01-13 04:53:31 <nanotube> unless you run open-case. :)
631 2011-01-13 04:54:01 <grubles> you should hear how loud my 5970 is with an open air case
632 2011-01-13 04:54:20 <nanotube> noise smoise. :)
633 2011-01-13 04:54:29 <nanotube> stick it in the basement where it's nice and cool, and forget about it.
634 2011-01-13 04:54:37 <hacim> if you have a basement :)
635 2011-01-13 04:54:49 <grubles> lol thats where it is right now
636 2011-01-13 04:54:59 <grubles> its like a little blow dryer
637 2011-01-13 04:55:10 <LobsterMan> will poclbm make use of the shader clock as well as the core clock?
638 2011-01-13 04:55:24 <nevezen> heh, mine is kinda' loud too
639 2011-01-13 04:55:57 <grubles> lobsterman, i believe the memory clock doesnt affect the hashes per second
640 2011-01-13 04:56:05 <LobsterMan> yeah that it doesn't seem to
641 2011-01-13 04:56:14 <LobsterMan> but i can adjust both the core clock and shader clock with rivatuner
642 2011-01-13 04:56:29 <grubles> right
643 2011-01-13 04:56:34 <hacim> hm noise could be a problem
644 2011-01-13 04:56:35 <grubles> oh i see
645 2011-01-13 04:56:53 <grubles> is there a way to link the clocks to be the same
646 2011-01-13 04:56:58 <lfm> tuning for bitcoins may make it slopw for graphics
647 2011-01-13 04:57:05 <LobsterMan> yeah a little checkbox for "link clocks"
648 2011-01-13 04:57:15 <LobsterMan> only the core clock seems to affect hashrate as best i can see...
649 2011-01-13 04:57:26 <grubles> hmm ok
650 2011-01-13 04:57:37 <grubles> i cant even set the shader clock with my card so i didnt know
651 2011-01-13 04:57:46 <grubles> it must be linked i guess
652 2011-01-13 05:00:27 <grubles> lobsterman, do you know about -v and -w flags for the py miner
653 2011-01-13 05:00:35 <LobsterMan> yeah
654 2011-01-13 05:00:46 <grubles> did they help at all
655 2011-01-13 05:00:54 <LobsterMan> -v seems to make the miner crash for me, but i set -w 256 and i get a slight increase in hashrate
656 2011-01-13 05:01:14 <hacim> $349 is not too bad for a 5970
657 2011-01-13 05:01:37 <grubles> where is that happening
658 2011-01-13 05:01:41 <nanotube> hacim: where are you getting that? last i saw 5970s were going for 500+
659 2011-01-13 05:01:55 <hacim> i found one on a local craigslist
660 2011-01-13 05:02:07 <LobsterMan> grubles the miner crashes after like 10s if i try to use -v at all
661 2011-01-13 05:02:12 <grubles> i would definitely take that offer up if i had the cash
662 2011-01-13 05:02:27 <nanotube> hacim: sounds like a great offer - assuming card is in working order. :)
663 2011-01-13 05:02:46 <nanotube> i'd get it while you have the chance.
664 2011-01-13 05:03:26 <grubles> lobsterman, interesting, i had a gtx 260 and im pretty sure -v worked on it
665 2011-01-13 05:03:40 <grubles> lobsterman, are your drivers updated
666 2011-01-13 05:03:46 <hacim> nanotube: heh, i dont even know if I have the hardware to run it, much less a place to put it where it wont be annoyingly loud
667 2011-01-13 05:03:51 <nanotube> hehe
668 2011-01-13 05:03:54 <LobsterMan> i have the most recent official, non beta nvidia drivers
669 2011-01-13 05:04:02 <nanotube> hmm, also there's one on ebay too: http://cgi.ebay.com/Sapphire-Radeon-HD-5970-ATI-2GB-GDDR5-Video-Card-/170590174729?pt=PCC_Video_TV_Cards&hash=item27b7f77e09
670 2011-01-13 05:04:33 <grubles> not bad at all
671 2011-01-13 05:04:48 <hacim> yeah I was just looking at that one too
672 2011-01-13 05:04:58 <hacim> its a little cheaper than the one I was talking about
673 2011-01-13 05:05:12 <hacim> now all you are going to jump on it :P
674 2011-01-13 05:05:27 <grubles> lol
675 2011-01-13 05:06:04 <grubles> not i
676 2011-01-13 05:06:20 <nanotube> well not me - i have an old com in a good case, but it'd need a new mobo+cpu+ram+psu combo to take the 5970. which would have me out at least another 300
677 2011-01-13 05:07:02 <hacim> yeah, i'm guessing that might be my situation too
678 2011-01-13 05:07:17 <grubles> yeah im pretty sure id need a new psu for another
679 2011-01-13 05:07:19 <hacim> probably those don't go in agp slots
680 2011-01-13 05:07:41 <nanotube> yea these are pci-e
681 2011-01-13 05:08:08 <nanotube> hehe, that card's got more ram than my main comp...
682 2011-01-13 05:08:18 <hacim> yeah me too :P
683 2011-01-13 05:08:35 <nevezen> man, I thought 512MB video ram was a monster
684 2011-01-13 05:08:42 <nevezen> but now we're up to 2GB?
685 2011-01-13 05:09:01 <nanotube> ya... crazy times
686 2011-01-13 05:09:10 <grubles> and terabytes of storage
687 2011-01-13 05:09:13 <nanotube> bitcoins are up to .4 usd... video ram is up to 2gb...
688 2011-01-13 05:09:53 <nevezen> I suppose 2GB isn't the max we've seen..
689 2011-01-13 05:10:15 <hacim> yeah, i dont have pci-e
690 2011-01-13 05:10:21 <grubles> there is a 4gb 5970 its $1000 though
691 2011-01-13 05:10:25 <nevezen> ouch
692 2011-01-13 05:10:27 <grubles> the asus one
693 2011-01-13 05:10:35 <grubles> yeah
694 2011-01-13 05:10:39 <hacim> thats pricey
695 2011-01-13 05:11:08 <nanotube> and the extra ram is useless for mining...
696 2011-01-13 05:11:17 <grubles> yeah
697 2011-01-13 05:11:23 <nevezen> I'm curious.. I don't suppose people are running bitcoin clients on tesla clusters, are they?
698 2011-01-13 05:11:37 <grubles> possibly
699 2011-01-13 05:11:42 <nanotube> they could... but teslas aren't as good as atis
700 2011-01-13 05:12:14 <nevezen> I'm behind like about 2 years on components.. :(
701 2011-01-13 05:12:19 <grubles> how would someone cluster say 4 computers each with 4 5970s?
702 2011-01-13 05:12:24 <nevezen> I know nothing about ATI cards
703 2011-01-13 05:12:33 <grubles> with a pool?
704 2011-01-13 05:12:54 <nanotube> grubles: same as you'd cluster 2 computers with 1 5970 each.
705 2011-01-13 05:13:11 <nanotube> just run miner processes on each, and gave it fetch work from the pool
706 2011-01-13 05:13:20 <grubles> in my experience the pooled miners dont have as high of hashes/s than the py miner for instance
707 2011-01-13 05:13:24 <lfm> grubles miners let you specify the ip address for the server
708 2011-01-13 05:13:43 <nanotube> you can run the py miner and just give it the pool address. the miner is exactly the same.
709 2011-01-13 05:13:49 <gribble> No fancy GPU farm, and don't want to wait for months for a block gen? Join the mining pool! http://mining.bitcoin.cz/
710 2011-01-13 05:13:49 <nanotube> see the ,,pool page for details.
711 2011-01-13 05:14:08 <nanotube> at least, that pool. puddinpop's pool does use separate software
712 2011-01-13 05:14:22 <grubles> im talking about the client software
713 2011-01-13 05:14:44 <nanotube> yes, same miner client.
714 2011-01-13 05:14:50 <lfm> grubles yes the bitcoin client is a server to miners
715 2011-01-13 05:15:02 <grubles> with the rpc miner i had less hashes/s than the py miner
716 2011-01-13 05:15:27 <nanotube> what's the 'rpc miner' ?
717 2011-01-13 05:15:50 <nevezen> from the point of joining that pool, when did you start generating coins?
718 2011-01-13 05:15:54 <grubles> http://www.bitcoin.org/smf/index.php?topic=2444.0 here
719 2011-01-13 05:16:07 <lfm> they're all rpc miners but I think hhe means pudnpop's miners
720 2011-01-13 05:16:23 <grubles> yes his
721 2011-01-13 05:16:25 <nanotube> ah those are puddinpop's miners
722 2011-01-13 05:16:33 <nanotube> but you can use the py miner with the slush pool
723 2011-01-13 05:16:43 <nanotube> see that pool page
724 2011-01-13 05:17:00 <grubles> yep
725 2011-01-13 05:17:16 <hacim> i wonder what mhash/sec could be achieved with amazon
726 2011-01-13 05:17:16 <lfm> ifaik all the miners use the same protocol and can be used with any of the sevrers, bitcind or a pool
727 2011-01-13 05:17:19 <nanotube> just saying... if you have a 5970 you don't really need to use the pool. :)
728 2011-01-13 05:17:35 <grubles> lol indeed
729 2011-01-13 05:17:47 <nanotube> lfm: unless something changed in puddinpop's miner... i think his pool server does something special... ?
730 2011-01-13 05:18:02 <nanotube> hacim: probably not enough to pay for itself.
731 2011-01-13 05:18:08 <lfm> nanotube its still compatible afaik
732 2011-01-13 05:18:20 <grubles> so with two seperate computers, you would just set bitcoin.exe -server on one and connect all the py miners to it?
733 2011-01-13 05:18:28 <nanotube> grubles: yes
734 2011-01-13 05:18:39 <lfm> grubles exactly
735 2011-01-13 05:18:48 <grubles> i see
736 2011-01-13 05:19:13 <nanotube> lfm: mm i guess something did change, then. heh
737 2011-01-13 05:19:19 <hacim> nanotube: yeah, i was just wondering if anyone has tried it and what hash/sec they get. curious about what it ends up costing
738 2011-01-13 05:19:49 <nanotube> hacim: someone did... but not at these prices. :)
739 2011-01-13 05:20:29 <lfm> nanotube oh ya, pudinpop has a whole new set of miners from his original one, the new ones work with stadard getwork calls
740 2011-01-13 05:21:02 <grubles> hacim, you could have 2 top of the line 1000 dollar cpus and they wouldnt get half the hashes/s as a 100 dollar gpu
741 2011-01-13 05:21:04 <hacim> oh, search doesn't work unless you are logged in
742 2011-01-13 05:21:10 <nanotube> lfm: ah ic
743 2011-01-13 05:21:39 <hacim> grubles: amazon has gpu clusters
744 2011-01-13 05:21:57 <nevezen> what does google apps have? :)
745 2011-01-13 05:22:05 <grubles> oh, i didnt know that :D
746 2011-01-13 05:22:09 <hacim> http://aws.amazon.com/ec2/hpc-applications/
747 2011-01-13 05:22:09 <xelister> hacim:  amazon builds clusters on shitty nvidias
748 2011-01-13 05:22:32 <hacim> 2 x NVIDIA Tesla
749 2011-01-13 05:23:10 <hacim> not even listed on the chart
750 2011-01-13 05:26:25 <nanotube> hacim: yea ask ArtForzZz what those can do, hps wize
751 2011-01-13 05:26:28 <nanotube> wise
752 2011-01-13 05:26:39 <grubles> seems like those have only 448 stream processors
753 2011-01-13 05:27:00 <lfm> ecc memory
754 2011-01-13 05:27:53 <grubles> gtx 275 have 240
755 2011-01-13 05:27:53 <nanotube> so... maybe like 90mhps?
756 2011-01-13 05:28:40 <grubles> nvidia are better at floating point operations which bitcoin does none of i believe
757 2011-01-13 05:28:46 <nanotube> right
758 2011-01-13 05:28:56 <hacim> i'm trying to figure out how much it would cost to run a gpu miner on an amazon instance
759 2011-01-13 05:29:09 <xelister> hacim: it is not proffitable I can bet
760 2011-01-13 05:29:12 <nanotube> hacim: they don't seem to have the prices on that right up front anywhere.
761 2011-01-13 05:29:17 <nanotube> xelister: even at .4/btc ?
762 2011-01-13 05:29:25 <grubles> i think i saw it was ~$.04 an hour
763 2011-01-13 05:29:36 <xelister> hm
764 2011-01-13 05:29:38 <nanotube> mmm ok
765 2011-01-13 05:29:55 <xelister> ok but how much dicking around with that?
766 2011-01-13 05:29:58 <grubles> or a couple grand a year
767 2011-01-13 05:30:10 <nanotube> so .04 an hour, you get two teslas at 90mhps each, you also get some number of 'cpu compute units' whatever those are, say another 10 mhps
768 2011-01-13 05:30:19 <nanotube> so 190 mhps total.
769 2011-01-13 05:30:21 <xelister> you just get ssh login with X to account on linux where you can use OpenCL and all is installed?
770 2011-01-13 05:30:23 <gribble> The average time to generate a block at 190000 Khps, given current difficulty of 16307.48285682 , is 4 days, 6 hours, 23 minutes, and 52 seconds
771 2011-01-13 05:30:23 <nanotube> ;;bc,calc 190000
772 2011-01-13 05:30:30 <gribble> The average time to generate a block at 90000 Khps, given current difficulty of 16307.48285682 , is 1 week, 2 days, 0 hours, 10 minutes, and 23 seconds
773 2011-01-13 05:30:30 <lfm> ;;bc,calc 90000
774 2011-01-13 05:30:56 <xelister> 150 coins... 45 usd
775 2011-01-13 05:30:59 <nanotube> so in about 4 days, you gen a block on average, which gives you 20 usd at current prices
776 2011-01-13 05:31:08 <nanotube> (to be generous)
777 2011-01-13 05:31:17 <gribble> 3.84
778 2011-01-13 05:31:17 <nanotube> ;;math calc 4*24*.04
779 2011-01-13 05:31:25 <xelister> so 90M or 190M ?
780 2011-01-13 05:31:26 <lfm> 90 mhash/s or 190 mhash/s??
781 2011-01-13 05:31:28 <nanotube> and cost to run on amazon is 3.84 usd for the 4 days
782 2011-01-13 05:31:37 <nanotube> lfm: 90 per one tesla. there are two teslas in one unit
783 2011-01-13 05:31:46 <lfm> oh ok
784 2011-01-13 05:31:47 <nanotube> and i added another 10mhps for the cpu compute units included
785 2011-01-13 05:32:01 <nanotube> (just ballpark, cuz i have no idea what their 'compute units' are...
786 2011-01-13 05:32:04 <hacim> so it seems profitable
787 2011-01-13 05:32:07 <nanotube> so it seems that it is profitable
788 2011-01-13 05:32:11 <nanotube> unless we're missing something.
789 2011-01-13 05:32:28 <hacim> i've been meaning to try amazon, maybe I should give it a whirl
790 2011-01-13 05:32:38 <nanotube> you get 20 usd per 3.84 usd investment... with no up front costs of hardware
791 2011-01-13 05:32:57 <nanotube> hacim: give it a try and let us know. :) specifically what their cost per hour turns out to be, and what the hps performance is
792 2011-01-13 05:32:58 <nanotube> heh
793 2011-01-13 05:32:59 <hacim> if we didn't miss something
794 2011-01-13 05:33:04 <grubles> every 4 days though
795 2011-01-13 05:33:08 <nanotube> yea, maybe it's not .04 per hour.
796 2011-01-13 05:33:21 <nanotube> grubles: 4 days at .04 per hour is only 3.84
797 2011-01-13 05:33:36 <nanotube> and you can join a pool, to smooth out your generation :)
798 2011-01-13 05:34:02 <gribble> {"ticker":{"high":0.405,"low":0.3176,"vol":40283,"buy":0.3401,"sell":0.414,"last":0.405}}
799 2011-01-13 05:34:02 <nanotube> well, i was rather generous at .4 usd/btc. the highest bid is only ,,bc,mtgox
800 2011-01-13 05:34:03 <xelister> can you join a pool and send fake hashes to "smooth out" your generation ;) ?
801 2011-01-13 05:34:07 <nanotube> .34, not .4
802 2011-01-13 05:34:20 <nanotube> xelister: no... that's been discussed at length.
803 2011-01-13 05:34:26 <xelister> wait, actually what the clients send, some easier hashes?
804 2011-01-13 05:34:33 <nanotube> pools are pretty good at weeding out fakes.
805 2011-01-13 05:34:34 <xelister> like >>4 the difficulty ?
806 2011-01-13 05:34:49 <lfm> hi is 0.405
807 2011-01-13 05:34:50 <nanotube> yea, slush pool asks for diff=1 hashes as shares
808 2011-01-13 05:35:25 <grubles> i suppose it is do-able with the rising price of bitcoin, but a block every 4 days is not going to net you a lot of coins
809 2011-01-13 05:35:33 <hacim> cluster GPU instances are actually $2.10/hr
810 2011-01-13 05:35:37 <grubles> and then the difficulty will probably rise which means you pay more for your cluster time
811 2011-01-13 05:35:41 <xelister> nanotube: so what is keeping users from sending easy hashes as agreed, but if they hit a winnig hard hash then just keep it? is it because they do not own the data needed to use the nonce?
812 2011-01-13 05:36:10 <hacim> thats $201.60 for four days
813 2011-01-13 05:36:39 <lfm> xelister right
814 2011-01-13 05:36:44 <xelister> so you get 20 USD for 200 USD invested
815 2011-01-13 05:36:48 <hacim> http://aws.amazon.com/ec2/#pricing
816 2011-01-13 05:37:24 <annodomini> xelister: Right. The hash is the hash of the whole block, including the nonce. The whole block includes the coin minting transaction. You can't change the block after the fact; that invalidates the hash.
817 2011-01-13 05:37:58 <nanotube> xelister: yes the hash contains the pool's key in the 50btc gen tx. keeping the block is useless to the client.
818 2011-01-13 05:38:37 <nanotube> hacim: haha, 2.10 per hour
819 2011-01-13 05:38:49 <nanotube> a couple orders of magnitude off on the cost there we are hehe
820 2011-01-13 05:38:50 <nevezen> not worth the investment then?
821 2011-01-13 05:38:52 <hacim> i thought that didn't make sense
822 2011-01-13 05:39:02 <gribble> 201.6
823 2011-01-13 05:39:02 <nanotube> ;;math calc 24*4*2.10
824 2011-01-13 05:39:15 <nanotube> yea
825 2011-01-13 05:39:36 <nanotube> i guess bitcoins need to go up in price to about 4 usd/btc in order for it to be profitable
826 2011-01-13 05:39:50 <nanotube> (while difficulty stays the same)
827 2011-01-13 05:40:05 <nevezen> this faintly reminds me back when I was running dnet
828 2011-01-13 05:41:04 <nanotube> hacim: well, it was a nice thought. hehe
829 2011-01-13 05:41:34 <LobsterMan> hmm
830 2011-01-13 05:41:47 <LobsterMan> is 'PCIE spread spectrum' likely to have any significant impact?
831 2011-01-13 05:41:50 <LobsterMan> it was auto in my bios
832 2011-01-13 05:41:51 <nanotube> maybe if they were offering 2x5970 on the gpu clusters instead of the teslas.
833 2011-01-13 05:41:53 <LobsterMan> but i just set it to disabled
834 2011-01-13 05:43:37 <hacim> nanotube: yeah, i was all ready to give them a shot too :)
835 2011-01-13 05:43:46 <grubles> it says something about making it easier to overclock in my mobo's manual
836 2011-01-13 05:43:53 <hacim> there is no way I'm dropping $210 for four days of computing :)
837 2011-01-13 05:43:58 <grubles> overclock a gpu i mean
838 2011-01-13 05:44:23 <nevezen> yes, disable spread spectrum when overclocking (the cpu)
839 2011-01-13 05:45:41 <hacim> looks like someone tried this: http://www.bitcoin.org/smf/index.php?topic=1334.msg26185#msg26185
840 2011-01-13 05:45:47 <hacim> 53481 khash/s
841 2011-01-13 05:46:17 <xelister> hacim: I wonder who buys such clusters thou
842 2011-01-13 05:46:26 <hacim> seriously
843 2011-01-13 05:48:08 <nevezen> scientific folks?
844 2011-01-13 05:48:33 <nevezen> run their simulations during the day. let it run bitcoin during the wee hours of night
845 2011-01-13 05:48:39 <nanotube> hacim: well, for the both of them, 112 mhps.
846 2011-01-13 05:48:46 <nanotube> but still, quite a bit worse than we thought. :)
847 2011-01-13 05:49:22 <nanotube> they do have a 'prepaid' cluster version, where you pay 5000 usd or so for a whole year.
848 2011-01-13 05:49:31 <nanotube> at that rate, though... it's cheaper to just make your own. :)
849 2011-01-13 05:50:17 <hacim> yeah, for real
850 2011-01-13 05:50:47 <grubles> and you cant even play games on it :p
851 2011-01-13 05:50:59 <nanotube> heh
852 2011-01-13 05:51:43 <xelister> or warm your house
853 2011-01-13 05:52:05 <hacim> FPGA doesn't help bitcoin?
854 2011-01-13 05:52:28 <nanotube> there's been discussion of that as well. they cost a lot.
855 2011-01-13 05:52:44 <nanotube> art did design and order a bunch of asics, though...
856 2011-01-13 05:52:46 <xelister> hacim: ArtForzZz is building his own processor thing
857 2011-01-13 05:52:51 <nanotube> we're all waiting to see how that turns out.
858 2011-01-13 05:52:55 <nevezen> or you can hack your own
859 2011-01-13 05:52:59 <nevezen> http://nsa.unaligned.org/
860 2011-01-13 05:53:40 <grubles> nevezen, interesting....
861 2011-01-13 05:57:31 <hacim> where can I find info about ArtForzZz's efforts?
862 2011-01-13 05:57:51 <mrb_> NSA@home is old... the guy claims about 3.3 billion MD5 hashes/sec, which matches the performance of a single HD 5830 ($170)
863 2011-01-13 05:58:43 <mrb_> and with a higher power consumption (240W) than a 5830 (175W)
864 2011-01-13 06:00:22 <mrb_> current FPGAs offer maybe a bit better perf/W but worse perf/$ than GPUs on SHA256 workloads (Bitcoin)
865 2011-01-13 06:00:37 <mrb_> 22:28 < grubles> nvidia are better at floating point operations which bitcoin does none of i believe
866 2011-01-13 06:01:34 <mrb_> not exactly true. Nvidia has better double-precision fp perf, ATI GPUs kick the ass of Nvidia GPUs on single precision fp workloads just like on integer workloads
867 2011-01-13 06:02:45 <mrb_> because Fermi can do 1 dp fp op every 2 cycles, whereas ATI can do 1 dp fp op only on 1 of the 4 ALUs (69xx) or 1 of the 5 ALUs (68xx/5xxx)
868 2011-01-13 06:06:41 <lfm> and 57xx and lower cant do any dp floating ops
869 2011-01-13 06:09:12 <grubles> i stand corrected :)
870 2011-01-13 06:10:58 <lfm> figuring out nvidia dp float support is at least as screwy
871 2011-01-13 06:11:44 <lfm> basiclly 260 and later have dp but not 260m or any of the |m| (mobile) models
872 2011-01-13 06:12:06 <mrb_> yep, 57xx can't do dp. but 58xx and 59xx can of course. so grubles is still incorrect :)
873 2011-01-13 06:16:04 <xelister> mrb_: yeah once my buddy said that
874 2011-01-13 06:16:12 <xelister> later I found out he took lots of blue pills
875 2011-01-13 06:18:43 <nevezen> it's still a neat way to crunch numbers
876 2011-01-13 06:19:23 <lfm> i think the whole opencl/cuda stuff is kinda messy myself
877 2011-01-13 06:21:07 <xelister> opencl?
878 2011-01-13 06:23:36 <lfm> gpu architechture is still in it infancy and gpu software support (OpenCL compiler etc) is worse
879 2011-01-13 06:26:47 <nevezen> more khash/s in the next few years with the same hardware? :p
880 2011-01-13 06:27:23 <lfm> naw they'll keep barging ahead with both hardware and software
881 2011-01-13 06:27:23 <sipa> i doubt it
882 2011-01-13 06:27:29 <sipa> maybe a few %
883 2011-01-13 06:28:16 <nevezen> I'm sure we'll find new uses for old hardware..
884 2011-01-13 06:28:37 <lfm> the gpu will get closer integration with the cpu sorta like the fpu (8087 line) did
885 2011-01-13 06:29:03 <nevezen> they tried that with pinetrail I think
886 2011-01-13 06:29:08 <nevezen> in netbooks
887 2011-01-13 06:29:24 <lfm> just a first attempt, its not the last
888 2011-01-13 06:32:01 <LobsterMan> heh i think i broke something <_<
889 2011-01-13 06:32:30 <LobsterMan> every time i try to start furmark now....my drivers crash, even with all oc'ing turned off
890 2011-01-13 06:32:34 <LobsterMan> http://www.ozone3d.net/benchmarks/fur/
891 2011-01-13 06:32:56 <lfm> do a full power down?
892 2011-01-13 06:33:00 <LobsterMan> several times
893 2011-01-13 06:33:15 <LobsterMan> maybe something got corrupted somewhere along the line
894 2011-01-13 06:33:39 <LobsterMan> it seems like it crashes when furmark tries to read the clock speeds
895 2011-01-13 06:33:46 <lfm> well overcloacking can do permanent damage
896 2011-01-13 06:33:51 <nathan7> fur?
897 2011-01-13 06:34:07 <LobsterMan> it starts up for a second, then when it would normally show me the clock speeds the program crashes, then my drivers crash
898 2011-01-13 06:34:11 <LobsterMan> (see my link for furmark.....)
899 2011-01-13 06:34:54 <LobsterMan> idk....rivatuner is working fine, and everest reports the correct clock speeds, and my system is still stable
900 2011-01-13 06:36:12 <lfm> try some game or opengl screensaver?
901 2011-01-13 06:36:22 <LobsterMan> i've tried counterstrike and it works fine
902 2011-01-13 06:36:33 <LobsterMan> only furmark is crashing which is strange
903 2011-01-13 06:36:36 <lfm> ok furmark is broken
904 2011-01-13 06:37:09 <LobsterMan> actually i have one more thing i think i can try....
905 2011-01-13 06:37:41 <lfm> geez that 4 way in your link seems to have no cooling for the middle cards!
906 2011-01-13 06:37:59 <LobsterMan> wish me luck
907 2011-01-13 06:38:08 <lfm> bye
908 2011-01-13 06:39:01 <mrb_> I doubt high-end GPUs will ever be integrated in CPUs
909 2011-01-13 06:40:33 <lfm> well they mioght tie em in closer than pcie bus like on the hyperlinks or something
910 2011-01-13 06:41:04 <mrb_> graphics manufacturer are already constrained by the 300W TDP envelope of a double-width card. putting this on a single CPU socket would be an even smaller envelope
911 2011-01-13 06:42:00 <lfm> smaller feature sizes will help power/heat limits
912 2011-01-13 06:42:13 <mrb_> high-end Intel CPUs already have integrated PCIe lanes... how much 'closer' can you get?
913 2011-01-13 06:42:44 <xelister> if the balls arent touching, it's not compact
914 2011-01-13 06:42:52 <lfm> hyperlinks like those used between cpu chips
915 2011-01-13 06:44:00 <lfm> for better shareing of memory at least
916 2011-01-13 06:44:24 <lfm> access to each other's memory
917 2011-01-13 06:45:17 <mrb_> it's the Jevons paradox: make the die more power-efficient, and the manufacturers will simply find a way to put the 'saved' energy to new use, eg. packing more ALUs on their dies
918 2011-01-13 06:45:56 <mrb_> and get back right to the previous power consumption levels
919 2011-01-13 06:46:06 <lfm> could also be related to the peter principle
920 2011-01-13 06:46:12 <mrb_> that's why all GPUs have plateau'd at 300W over the past 3 years
921 2011-01-13 06:46:25 <xelister> it was common PSU limit?
922 2011-01-13 06:46:40 <mrb_> despite constant improvement of the perf/W
923 2011-01-13 06:46:46 <LobsterMan> meh i think furmark just sucks
924 2011-01-13 06:46:47 <LobsterMan> lol
925 2011-01-13 06:46:54 <mrb_> 300W is the max authorized by the PCIe specs
926 2011-01-13 06:46:59 <mrb_> for double-width cards
927 2011-01-13 06:47:03 <lfm> every watt has to be cooled too
928 2011-01-13 06:47:27 <mrb_> 75W from the slot, 75W from the 6-pin power connector, and 150W from the 8-pin one
929 2011-01-13 06:48:32 <lfm> so cant they use two 8 pins
930 2011-01-13 06:49:04 <sipa> wow, 75W from the PCIe slot?
931 2011-01-13 06:49:07 <mrb_> electrically speaking they could, but they would be violating the specs
932 2011-01-13 06:49:44 <LobsterMan> ;;bc,stats
933 2011-01-13 06:49:47 <gribble> Current Blocks: 102416 | Current Difficulty: 16307.48285682 | Next Difficulty At Block: 102815 | Next Difficulty In: 399 blocks | Next Difficulty In About: 2 days, 9 hours, 18 minutes, and 3 seconds | Next Difficulty Estimate: 18111.47884977
934 2011-01-13 06:49:47 <mrb_> manufacturers do this in-house though for their GTX 595 prototypes (2 x 8-pin)
935 2011-01-13 06:50:02 <lfm> gross amount of power anyway. they need to do more with less
936 2011-01-13 06:50:35 <LobsterMan> ;;bc,calc 109100
937 2011-01-13 06:50:38 <gribble> The average time to generate a block at 109100 Khps, given current difficulty of 16307.48285682 , is 1 week, 0 days, 10 hours, 19 minutes, and 40 seconds
938 2011-01-13 06:50:49 <mrb_> ArtForz says the GTX 595 can't be dual-GF110 because of the excessive power consumption but I think it will. Nvidia will simply reduce the voltage and clocks until it its under 300W
939 2011-01-13 06:50:56 <lfm> part of the reason the P4 sucked was it took too much power
940 2011-01-13 06:51:00 <mrb_> s/its/fits/
941 2011-01-13 06:52:34 <lfm> lobsterman so its running now?
942 2011-01-13 06:55:18 <mrb_> same thing for CPUs. they have plateau'd at ~130W because that's 5.4 Amp on each wire of the 4-pin ATX12V connector which is dedicated to the CPU. You can do up to 6, 7 Amps without the wires heating too much but it requires 16 gauge which most PSUs don't have
943 2011-01-13 07:01:45 <lfm> and intel got its arse kicked when amd was able to make the athlon 64 more efficient than the p4. big customers went amd in droves
944 2011-01-13 07:03:55 <mrb_> yep.
945 2011-01-13 07:14:55 <LobsterMan> lfm seems to be working
946 2011-01-13 07:15:11 <lfm> wtg
947 2011-01-13 07:15:38 <LobsterMan> from 684->720MHz and stable
948 2011-01-13 07:17:17 <lfm> Average interval last 100 blocks: 7.66 min
949 2011-01-13 07:17:42 <lfm> someone is poring on the steam