1)
Message boards :
Number crunching :
Running SETI@home on an nVidia Fermi GPU
(Message 1023789)
Posted 8 Aug 2010 by Codeman05 Post: Hey thanks for the heads up Jason, not sure how I missed that lol Wow...so far looks to be a sizable boost on my box...nice... :) On the bright side, did also manage to get multiple WUs running on the 460s :D Guessing I had some bad units before, because I cant figure out what I did for the life of me... thanks again guys Hey! You're a Squire!, when you get comfortable you might like to try the x32f build in Beta testing. |
2)
Message boards :
Number crunching :
Running SETI@home on an nVidia Fermi GPU
(Message 1023773)
Posted 8 Aug 2010 by Codeman05 Post: Hi Josef, Yes I've been logging with GPU-z and the clock speeds are staying at 100%. However they usually however around 60-70% load...which I'm not sure if that is normal or not. I will try putting in the flops values for the card and see what that does. Appreciate the help |
3)
Message boards :
Number crunching :
Running SETI@home on an nVidia Fermi GPU
(Message 1023772)
Posted 8 Aug 2010 by Codeman05 Post: I have a batch that hasn't uploaded yet (network issues). Looks like they are finishing up in the 22-25min range for the most part. With some 12 mins as you noted |
4)
Message boards :
Number crunching :
Running SETI@home on an nVidia Fermi GPU
(Message 1023762)
Posted 8 Aug 2010 by Codeman05 Post: Hey guys, I just put together a seti box with 2x GTX460s...the performance is pretty poor however. I understand Fermi support isn't real solid at present, but my results seem worse than what I'm seeing browsing around. System Info Intel i7 920 2x GTX460 Windows 7-64bit BOINC 6.11.4 Lunatics UI 0.36 (MB only) My Fermi's are currently spending ~22-30 minutes per workunit...each. According to GPUz, they are running at speed with ~75% GPU load. I've tried running multiple WUs on the GPUs, and that yeilds really awefull results. GPU usage drops to ~40%, and complete about 0.25% in 5 minutes. Below is my app_info that I cobbled together from this thread. Any assistance would be great...or maybe this is to be expected (VLARs?) ?? Thank you <app_info> <app> <name>setiathome_enhanced</name> </app> <file_info> <name>AK_v8b_win_x64_SSSE3x.exe</name> <executable/> </file_info> <file_info> <name>setiathome_6.10_windows_intelx86__cuda_fermi.exe</name> <executable/> </file_info> <file_info> <name>cufft32_30_14.dll</name> <executable/> </file_info> <file_info> <name>cudart32_30_14.dll</name> <executable/> </file_info> <file_info> <name>libfftw3f-3-1-1a_upx.dll</name> <executable/> </file_info> <app_version> <app_name>setiathome_enhanced</app_name> <version_num>603</version_num> <platform>windows_intelx86</platform> <file_ref> <file_name>AK_v8b_win_x64_SSSE3x.exe</file_name> <main_program/> </file_ref> </app_version> <app_version> <app_name>setiathome_enhanced</app_name> <version_num>603</version_num> <platform>windows_x86_64</platform> <file_ref> <file_name>AK_v8b_win_x64_SSSE3x.exe</file_name> <main_program/> </file_ref> </app_version> <app_version> <app_name>setiathome_enhanced</app_name> <version_num>610</version_num> <platform>windows_intelx86</platform> <avg_ncpus>0.200000</avg_ncpus> <max_ncpus>0.200000</max_ncpus> <plan_class>cuda_fermi</plan_class> <file_ref> <file_name>setiathome_6.10_windows_intelx86__cuda_fermi.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>cudart32_30_14.dll</file_name> </file_ref> <file_ref> <file_name>cufft32_30_14.dll</file_name> </file_ref> <file_ref> <file_name>libfftw3f-3-1-1a_upx.dll</file_name> </file_ref> <coproc> <type>CUDA</type> <count>1</count> <----- Changed to 0.5 when attempted to run 2x WUs/gpu </coproc> </app_version> </app_info> |
5)
Message boards :
Number crunching :
2nd GPU stuck at idle clock speeds??
(Message 1020360)
Posted 28 Jul 2010 by Codeman05 Post: Just for reference, I ended up going with an older NV driver that I'm using on another rig, Forceware 197.45. That seems to have done it. Before I was getting driver crashes every 2-3 hours, so far I'm 14 hours with zero errors. So far so good. |
6)
Message boards :
Number crunching :
2nd GPU stuck at idle clock speeds??
(Message 1020202)
Posted 27 Jul 2010 by Codeman05 Post: I found on my BFG GTX 260 any overclock on the memory over 1200 would I've just been running stock so far at 590/1000/1296 (core/mem/shader). So far the older driver's seem to be helping, we''ll see how it does overnigiht though. |
7)
Message boards :
Number crunching :
2nd GPU stuck at idle clock speeds??
(Message 1020173)
Posted 27 Jul 2010 by Codeman05 Post: Are you getting a driver crash sometime in the night? Mine did that until I put a rubber spacer between my cards to let more air in. Actually...yes I am. I stopped by the house at lunch and started looking through the Event Viewer in Admin Tools. When this happens I am getting the following error: "Display driver nvlddmkm stopped responding and has successfully recovered." The WU's being processed then crash, showing as errored out. The GPU's then start on the next WU's, but with the low clock speeds and slowness. Closing Boinc and restarting it doesn't work, I have to reboot the box. Then it's fine until the driver crashes again. Temps look good though, I'm under 70C. Think I'll try and older NV driver first and see if that helps any. Thanks again for the input guys :) |
8)
Message boards :
Number crunching :
2nd GPU stuck at idle clock speeds??
(Message 1020124)
Posted 27 Jul 2010 by Codeman05 Post: Thanks T.A, that's good info. I am running an older version of Precision and the problem your describing sounds exactly like what I'm seeing. I'll give the new version a shot as well. Thanks again |
9)
Message boards :
Number crunching :
2nd GPU stuck at idle clock speeds??
(Message 1020106)
Posted 27 Jul 2010 by Codeman05 Post: 1) Gigabyte X58 GA-X58-UD5 2) Forceware 258.96 3) EVGA Precision 4) W7 64 | Boinc 6.10.18 | both cards are EVGA GTX260 216s | using gpuz to watch load Upon further research into this it looks like the second gpu will *usually* start out fine and then at some point overnight (or a few hours) drop to idle clock speeds, but still be processing WUs (judging by the memory/gpu load). I noticed the #2 card was set to be the PhysX card in the NV control panel, I just disabled that for the heck of it to see if that has any affect. Thanks for the input |
10)
Message boards :
Number crunching :
2nd GPU stuck at idle clock speeds??
(Message 1020104)
Posted 27 Jul 2010 by Codeman05 Post: I don't have a dummy plug on the second GPU, I thought that only helped when the second GPU wasn't being seen by BOINC?? I have some extras though so I can give that a shot. No custom cc_config. I just rebooted and now both cards are running at full speed again. Seems like the 2nd card randomly will downclock to idle speeds (still under full load and crunching). I'll keep an eye on it now that I'm aware of the issue and see when it drops off. |
11)
Message boards :
Number crunching :
2nd GPU stuck at idle clock speeds??
(Message 1020099)
Posted 27 Jul 2010 by Codeman05 Post: Hi guys, I recently popped a second GPU (GTX260) in one of my crunchers and have noticed that this second GPU...while crunching...will remain at stock clock speeds and thus is rather slow. For example, the 1st GTX260 is running at 650mhz core, the 2nd is running at only 400Mhz and is almost twice as slow. I've tried with SLI enabled, SLI disabled, overclocking etc... and can't seem to get the #2 card to throttle up. I'm running BOINC 6.10.18 with the latest Lunatics Unified Installer and the Nvidia 258.96 drivers (W7 64). Any suggestions? Thanks in advance. |
12)
Message boards :
Technical News :
Fill 'er up (Jun 23 2010)
(Message 1007653)
Posted 24 Jun 2010 by Codeman05 Post: Thank you for the update Matt and for all the hard work you and the team have put in! |
13)
Message boards :
Technical News :
Maybe we're fixed? (March 16, 2010)
(Message 979946)
Posted 17 Mar 2010 by Codeman05 Post: woohoo, cleared out a few thousand completed WU's overnight! Haven't received any new units yet, but I'd say that's a large step in the right direction :) Thanks for the update Eric and all of the work you guys have been putting in. |
14)
Message boards :
Technical News :
The more things change, the more they stay the same.
(Message 979051)
Posted 15 Mar 2010 by Codeman05 Post: You guys are lucky, I can't even get downloads. Most of my crunchers have a large cache and still running, but a couple are sucking fumes. Here's hoping for a quick fix... |
15)
Message boards :
Number crunching :
NVidia GT240 - any good for crunching & expected RAC?
(Message 974481)
Posted 27 Feb 2010 by Codeman05 Post: wow, that made a hell of a difference, good call. Seems to be running cuda about 4x faster |
16)
Message boards :
Number crunching :
NVidia GT240 - any good for crunching & expected RAC?
(Message 974479)
Posted 27 Feb 2010 by Codeman05 Post: thanks, I'll give that a shot :) |
17)
Message boards :
Number crunching :
NVidia GT240 - any good for crunching & expected RAC?
(Message 974475)
Posted 27 Feb 2010 by Codeman05 Post: Andy can you clarify what you did? Did you install the CUDA via the Unified installed and just override with the Cuda 2.3 drivers? I seem to be having the same issue you were experiencing a couple posts back with Seti tossing all of my cuda WUs |
18)
Message boards :
Number crunching :
NVidia GT240 - any good for crunching & expected RAC?
(Message 974373)
Posted 27 Feb 2010 by Codeman05 Post: I just picked up a pair of these with the DDR5 @ frys for $80 a piece...couldn't resist giving them a shot. It's going to be a little while until the RAC stabilizes, but so far it looks like they are knocking out a WU in about 35mins a piece (196.21 w/ Lunatics Unified 0.2) That about inline with what you guys are seeing? |
19)
Message boards :
Number crunching :
Approx. MB Unit completion times on i7?
(Message 972156)
Posted 20 Feb 2010 by Codeman05 Post: Hey guys, Finally back in the crunching after most of my farm went down a few months ago. Hopefully everything will be back up this week. However, I have a few questions. The first question involved the replacement i7 920 cpu that I'm running. It is running the Lunatics Unified Installer 0.2. CPU is at 4.0GHz (HT off). My MB units appear to be crunching in about 60-70 minutes on average, per core, on MB units. Does this sound about right? For some reason, I thought my old i7 was faster (~30-40m)...but it has been a while. I understand that the crunch times will fluctuate...I'm just looking to see if I'm in the ballpark. Lastly,the last time I read a thread on HT vs HT off...the consensus seemed to be that it didn't really matter, the performance impact in the end was marginal. Is that still the case? Thanks for your input! |
20)
Message boards :
Number crunching :
Hyperthreading i7 and CUDA
(Message 899611)
Posted 26 May 2009 by Codeman05 Post: thank you for all of your work (and patience) John. Definitely looking forward to your conclusion. I currently have an i7 with 3 GPU's that I have run for a month with HT on and this past month with HT disabled. I won't get into the details as my "experiment" was in no way performed with the level of detail that yours was....but it does seem that my results do mimic yours. I did see a slight RAC/DAC increase with HT disabled. CUDA units processed about 33% faster and APs just over 50% faster, resulting in a slightly higher net output. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.