WARNING: This website is obsolete! Please follow this link to get to the new Albert@Home website!
Posts by Infusioned |
21)
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.23/1.24 (OpenCL) feedback thread
(Message 112002)
Posted 2 May 2012 by Infusioned Post: I see your host has now a mix of old and new WUs ? I poked through my history and all my wu's have 20110421 in them. I started aborting batches to try and get some new ones, but no dice so far. Unless I am mistaken, the 20110421 is the datestamp for when the data was recorded? Or is that the datestamp from when it was split? I have the day off tomorrow so I will abort/babysit Boinc to try and get some newer ones. |
22)
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.23/1.24 (OpenCL) feedback thread
(Message 111997)
Posted 1 May 2012 by Infusioned Post: I'm not sure why, but I've thrown 3 error recently: http://albert.phys.uwm.edu/workunit.php?wuid=67888 http://albert.phys.uwm.edu/workunit.php?wuid=66586 http://albert.phys.uwm.edu/workunit.php?wuid=66147 edit: Upon examination, all the wu's that errored start with this: <core_client_version>7.0.26</core_client_version> <![CDATA[ <message> Incorrect function. (0x1) - exit code 1 (0x1) </message> |
23)
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.23/1.24 (OpenCL) feedback thread
(Message 111993)
Posted 29 Apr 2012 by Infusioned Post: One more thing: while the workunit mentioned above was sent only recently, it was generated already on the 23rd of April, so it is still one of the "tweaked" workunits. Once the newly generated workunits are reached out, we should see a reduced memory usage and some modest performance increase. I was afraid of that. However, I didn't know how to decipher what date p2030.20110421.G41.06+00.53.N.b6s0g0.00000_3728_0 ... Never mind, I just realized that 20110421 means April 21, 2011. |
24)
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.23/1.24 (OpenCL) feedback thread
(Message 111988)
Posted 29 Apr 2012 by Infusioned Post: Also, is there a way to make them thumbnails in my post and when you click them they link to larger images (just to not annoy people with really large images)? |
25)
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.23/1.24 (OpenCL) feedback thread
(Message 111987)
Posted 29 Apr 2012 by Infusioned Post:
March? April? Or May of last year? wu 4/29/2012 9:29:59 AM | Albert@Home | Starting task p2030.20110421.G41.06+00.53.N.b6s0g0.00000_3728_0 using einsteinbinary_BRP4 version 123 (atiOpenCL) in slot 0 |
26)
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.23/1.24 (OpenCL) feedback thread
(Message 111985)
Posted 28 Apr 2012 by Infusioned Post: GPU load is steady at 20-21%, and CPU load literally bounces: 5%,15%,6%,14%,4%,17%, etc. with 17 being the highest I've seen. |
27)
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.22 feedback thread
(Message 111961)
Posted 20 Apr 2012 by Infusioned Post: Here is a WU from Seti@Home Beta's OpenCL application: http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=3973426 I am 58327 and someone with a GTX 590 GPU, Intel 2600k CPU, Cuda OpenCL client is 56759. My CPU seconds are 1463 and theirs are 2061. My GPU seconds are 3244 and theirs are 2198. My CPU time is actually lower (75%) of the 2600k, but my GPU time is ~150% of the GTX 590 (which again, is curious, given the GFLOP numbers). My conclusion from all this is then, that the Albert AMD OpenCL application isn't as quite as optimized as the Albert CUDA application. Can anyone confirm/deny? |
28)
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.22 feedback thread
(Message 111960)
Posted 16 Apr 2012 by Infusioned Post: This implies that slow CPUs will slow down OpenCL workunits much more than they slow down CUDA workunits. Understood. However, that's why I checked Anandtech's benchmarks to see just how much faster the 2600k was than my cpu. The benchmarks do not reflect a 66% performance difference so there is something else going on. Also, unless I read the charts wrong, comparing the GFLOPS between the two video cards, theoretically the 6950 should smoke the 550Ti in SP output (2253 vs. 691.2). So, back to my original question, is the OpenCL app that unoptimized compared to the CUDA app? |
29)
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.22 feedback thread
(Message 111958)
Posted 15 Apr 2012 by Infusioned Post: I have been away for a bit due to my motherboard dying, and when I got back up and running with the rebuild I was waiting for 7.0.25 to go live so I could run Milkway with Albert without using a beta version for a live project. So, poking through my WU times, I am hovering at ~ 5900 GPU seconds, and ~ 3,200 CPU seconds per WU. AMD Phenom II x4 975 (couldn't find an 1100T for non-ripoff prices and am waiting for Piledriver [not happy with Bulldozer]) AMD HD 6950 8G DDR3 1600 Win 7 x64 Boinc 7.0.25 This WU vs. i2600k Sandybridge/550Ti shows the 2600k coming in at 1/3 the time of my cpu. However, Anandtech Bench does not show the 2600k as 66% faster. Also, wikipedia shows the AMD HD6950 SP GFLOPS at 2253 and the NVIDIA GTX 55Ti SP GFLOPS at 691.2, but the 550Ti time is 2/3 of mine. So, my question is, what gives? Is the OpenCL app that unoptimized compared to the CUDA app? |
30)
Message boards :
Problems and Bug Reports :
Validate error
(Message 111671)
Posted 10 Jan 2012 by Infusioned Post: 2h/WU and i have 4 Validate error. For what it is worth I am running my 5870 with driver version 1.4.1546 [Catalyst 11.9] and haven't had any problems. http://albert.phys.uwm.edu/show_host_detail.php?hostid=1380 |
31)
Message boards :
Problems and Bug Reports :
NVIDIA validation issues (GTX 295, 470, 580, 590, et. al)
(Message 111639)
Posted 1 Jan 2012 by Infusioned Post: Poking through my work units, it seems various NVIDIA cards are having trouble validating: (I am running an AMD 5870 (atiOpenCL) on Win 7 x64, SP1) Validates: WU 24389 atiOpenCL + BRP3SSE WU 24563 atiOpenCL + BRP3cuda32 (GTX 550 Ti) WU 24633 atiOpenCL + BRP3cuda32OSX (Quadro 4000) WU 24650 atiOpenCL + BRP3cuda32 (GTX 570) WU 24702 atiOpenCL + BRP3cuda32 (GeForce 8700M GT) WU 24715 atiOpenCL + BRP3cuda32 (GeForce 9600 GT) WU 24735 atiOpenCL + BRP3cuda32 (GeForce GTX 580) WU 24739 atiOpenCL + BRP3cuda32 (GeForce GTX 260) Here's where things get weird... Does Not Validate: WU 20316 atiOpenCL + (GTX 570) + (GTX 295) The GTX 295 did not validate with either the atiOpenCL or the GTX 570, but the atiOpenCL & GTX 570 validated together. Digging though the GTX 295 host history shows that they have a history of hit and miss validation errors with both NVIDIA & ATi. WU 20274 atiOpenCL + (AMD 5800 [Linux]) + (AMD 6250/6310) + (GTX 580) + (AMD A8-3850 APU) The Linux platform AMD 5800 & the AMD 6250/6310 errored out while computing. The A8-3850 APU did not validate with GTX 580, but did validate with my 5870. Digging through the AMD A8-3850 APU host history, I found two other workunits where the GTX 580 does not validate with it: WU 21874 WU 24782 WU 23559 atiOpenCL + (AMD 5800 [Linux]) + (GTX 580) + (GTX 470) The (BRP3cuda32) GTX 580 did not validate against the (BRP3cuda32nv270) GTX 470. The the unit was then sent to Bikeman (AMD 5800 [Linux]) and myself. The two atiOpenCL app's validated against each other. WU 23524 atiOpenCL + (GTX 580) + (GTX 580) I validated against a GTX 580, but the other GTX 580 did not validate. As I finished typing this I noticed that the GTX 580 that kept invalidating was the same host http://albert.phys.uwm.edu/show_host_detail.php?hostid=1210 and he has 193 invalid results. But it is still strange that in WU 23559 that the GTX 470 did not validate against the AMD 5800 and it took another atiOpenCL to validate. That would point to the atiOpenCL app, however, I have validated against plenty of NVIDIA cards (shown above) previously. |
32)
Message boards :
Problems and Bug Reports :
AMD ATI Radeon HD 4700/4800 (RV740/RV770)
(Message 111628)
Posted 27 Dec 2011 by Infusioned Post: I was running a 4870 and had all the same issues Ageless had. Then, about 3-4 weeks ago, I found a killer deal on a 5870 and have since upgraded. I didn't even bother to put the 4870 in tandem with the 5870. Next on the agenda is a new motherboard, RAM (possibly CPU, waiting on Piledriver) and another 5870. The point is, I can't see ever going back to the 4xxx series. Like Ageless said, maybe it's altogether simpler to deprecate support for 4xxx ? |
33)
Message boards :
Problems and Bug Reports :
going to have to pass on this project
(Message 111627)
Posted 27 Dec 2011 by Infusioned Post: I'm guessing the admins don't quite get what our little underground (for lack of a better term) group of beta testers do. Whenever a new project pops up, word gets passed around through or team forums, IRC channels, etc. There's some hardcore guys with enough cores and gpus to put a serious strain on a new project. I certainly can understand your motivations, however, I would argue that a beta project is not the best choice for that. A beta project is a (sometimes) long bumpy road where we sign up to purposefully stub our toes and faces on any jagged corners the software has in the hopes of discovering bugs/inefficiencies to help fix them. The end goal is to help the devs to develop a solid release everyone can then use. Credit is the last of anyone's concern. I'd say you are in the wrong place. Also, not to be a d!ck, but I'd say credit grubbing is contradictory to beta testing entirely. |
34)
Message boards :
News :
Sending work
(Message 111490)
Posted 5 Dec 2011 by Infusioned Post: It doesn't mean that NVidia cards don't silently generate overflows in general. Also, the second post I quoted details how even the CUDA app was not validating against a CPU due to the order of calculations and the validator needed tweaking to regard them as weakly similar. |
35)
Message boards :
News :
Sending work
(Message 111487)
Posted 4 Dec 2011 by Infusioned Post: This is from the Seti@Home Beta message boards (developing an ATi OpenCL App): http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=1867 Please, update your hosts with this builds. ATi and NV builds got speed increase, HD5-version added, NV version added. Thanks, I downloaded the WU and confirm the difference. Maybe this is part of the issue regarding validation? |
36)
Message boards :
News :
Sending work
(Message 111416)
Posted 26 Nov 2011 by Infusioned Post: CAL ATI Radeon HD 4700/4800 (RV740/RV770) (1024MB) driver: 0.1 It seems Ageless and I have the same video card. Thus far, I haven't been able to validate a single WU: http://albert.phys.uwm.edu/workunit.php?wuid=12534 http://albert.phys.uwm.edu/workunit.php?wuid=12549 http://albert.phys.uwm.edu/workunit.php?wuid=12550 http://albert.phys.uwm.edu/workunit.php?wuid=12551 http://albert.phys.uwm.edu/workunit.php?wuid=12294 Also (as mentioned above) it looks like there are problems with the nVidia cards validating with themselves (wu 12294): The GTX 260 validates with the GTX 590, but not with the GTX 580. |
37)
Message boards :
Wish List :
Low credit for such a risky project
(Message 111387)
Posted 22 Nov 2011 by Infusioned Post: Well, this thready is already plenty long, so I guess my .02 won't hurt. I never understood the deal with credit/cobblestones anyway. I thought the 1 credit per 1 WU from the original Seti@home was just fine. /shrug In any case, psychologically I think 0 is a little nebulous for some people. I think as long as there is some indicator of contribution, most are happy. As far as alpha testing goes, I'm just here to help the OpenCL along so I can use my ATi Gpu's and could care less about credit in this setting. |
38)
Message boards :
Problems and Bug Reports :
Upload/Download server
(Message 111311)
Posted 19 Nov 2011 by Infusioned Post: I think the hamster died, getting .37 kBps trying to download the binaries. Otherwise I most fail and I am set to retry. |