Message boards :
Number crunching :
Update for GPU AP (both NV and ATi) to rev560
Message board moderation
Author | Message |
---|---|
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Here the link to GPU AP rev560 with improved response times on BOINC exit/suspend request. http://dl.dropbox.com/u/60381958/AP6_r560_GPU.7z Don't forget to change executable file name in your app_info files. |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
I am on it. Both here and at Beta. I got my first Mixed Beta task completed. First 97% done with .r555 and the last part with .r560 http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=10732282 Just hope it will validate. It will ;) Is V6 V6.01 V6.02 compatible??? I saw the ap tapes are started again. Lets hope I catch 1 or 2 tasks. Is it ok to run this on Beta as well? //TRuEQ TRuEQ & TuVaLu |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Sure it compatible, for what else I would post it? ;) |
Horacio Send message Joined: 14 Jan 00 Posts: 536 Credit: 75,967,266 RAC: 0 |
Is this version intended to be used by anybody or is it only for SetiBeta? If it's public, can we have an example of the Nvidia app_info for this new app version? Is the issue with CPU usage on new NV drivers solved? Are the .cl and .dll files needed for NV? mmm... Can I have an easter egg? :D |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
Is this version intended to be used by anybody or is it only for SetiBeta? Copy the .exe .cl and the .dll files to your c:\ProgramData\Boinc\projects\setiathome.berkeley.edu folder. It works on both Beta and here on Main. I think this is White and with a yellow core. GPU only app_info.xml <app_info> <app> <name>astropulse_v6</name> </app> <file_info> <name>AP6_win_x86_SSE3_OpenCL_NV_r560.exe</name> <executable/> </file_info> <app_version> <app_name>astropulse_v6</app_name> <version_num>601</version_num> <avg_ncpus>0.04</avg_ncpus> <max_ncpus>0.2</max_ncpus> <plan_class>CUDA</plan_class> <cmdline>-instances_per_device 2 -unroll 10 -ffa_block 8192 -ffa_block_fetch 4096</cmdline> <coproc> <type>CUDA</type> <count>0.5</count> </coproc> <file_ref> <file_name>AP6_win_x86_SSE3_OpenCL_NV_r560.exe</file_name> <main_program/> </file_ref> <flops>30987654321</flops> </app_version> </app_info> TRuEQ & TuVaLu |
Horacio Send message Joined: 14 Jan 00 Posts: 536 Credit: 75,967,266 RAC: 0 |
Copy the .exe .cl and the .dll files to your c:\ProgramData\Boinc\projects\setiathome.berkeley.edu Thanks! I've installed it and everything seems ok (but, I had to see what happens when it gets work). The remaining doubt is about the Nvidia drivers, should I keep using the old 266.58 that was needed by the NV-r521 to not clog the CPUs? (In fact, 266.58 is Ok and I have no reason to upgrade it, but one of my hosts has 2 560Ti's and the older driver for them is 266.66 which is said that has some issues...) |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
Copy the .exe .cl and the .dll files to your c:\ProgramData\Boinc\projects\setiathome.berkeley.edu I don't have the perfect answer for your question. But here at lunatics they have a thread about drivers that is pretty big. http://lunatics.kwsn.net/12-gpu-crunching/latest-nvidiadriver-and-cudaversion.0.html Here is a link if you want to read some. //TRuEQ TRuEQ & TuVaLu |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
The remaining doubt is about the Nvidia drivers, should I keep using the old 266.58 that was needed by the NV-r521 to not clog the CPUs? 270.xx and up suffer from excess CPU usage. AFAIK it still not fixed in latest drivers. I use 263.xx for crunching with GTX250 and it goes very well. |
Wiggo Send message Joined: 24 Jan 00 Posts: 34744 Credit: 261,360,520 RAC: 489 |
The remaining doubt is about the Nvidia drivers, should I keep using the old 266.58 that was needed by the NV-r521 to not clog the CPUs? So in other words, more testing/programing needs to be done with later drivers to get things right is what you are saying then? Cheers. |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
The remaining doubt is about the Nvidia drivers, should I keep using the old 266.58 that was needed by the NV-r521 to not clog the CPUs? I'd say if your current drivers are working fine. Don't upgrade to anything newer here then 269.xx NVIDIA still have some work to fix the CPU-Issue on their new drivers. I'm thinking of downgrading my 296.10 to 266.58. //TRuEQ TRuEQ & TuVaLu |
Jamie Send message Joined: 5 Apr 06 Posts: 162 Credit: 9,867,955 RAC: 0 |
The remaining doubt is about the Nvidia drivers, should I keep using the old 266.58 that was needed by the NV-r521 to not clog the CPUs? Can't say what needs to be done to resolve the high CPU usage issues, but the highest verison I've been able to go to is 267.24 (beta) without getting any issues |
Wiggo Send message Joined: 24 Jan 00 Posts: 34744 Credit: 261,360,520 RAC: 489 |
I'll settle with just doing them only on my CPU's only until someone works out what the problem is and gets it 100% fixed before I even think about letting my GPU's have a go at them. [edit] But that's what volunteer developers/testers are for isn't it? [/edit] Cheers. |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
I'll settle with just doing them only on my CPU's only until someone works out what the problem is and gets it 100% fixed before I even think about letting my GPU's have a go at them. I didn't know we tested stuff for NVIDIA here.... Are we? //TRuEQ TRuEQ & TuVaLu |
JohnDK Send message Joined: 28 May 00 Posts: 1222 Credit: 451,243,443 RAC: 1,127 |
I'm using 268,36 on my almost new notebook and have just finished the first AP, all went fine and the CPU use was rather low, mostly under 5%. The Stderr output contains some stuff that I wonder about, is there something I should do or is that all normal info? http://setiathome.berkeley.edu/result.php?resultid=2387164861 |
red-ray Send message Joined: 24 Jun 99 Posts: 308 Credit: 9,029,848 RAC: 0 |
The remaining doubt is about the Nvidia drivers, should I keep using the old 266.58 that was needed by the NV-r521 to not clog the CPUs? What about using a GTX 680 when the only drivers are 301.10? How high is "excessive CPU usage" as a % of a CPU on a 2.66GHz Core 2 Quad please? |
Mike Send message Joined: 17 Feb 01 Posts: 34258 Credit: 79,922,639 RAC: 80 |
I'm using 268,36 on my almost new notebook and have just finished the first AP, all went fine and the CPU use was rather low, mostly under 5%. Looks good. Normal info. With each crime and every kindness we birth our future. |
Mike Send message Joined: 17 Feb 01 Posts: 34258 Credit: 79,922,639 RAC: 80 |
The remaining doubt is about the Nvidia drivers, should I keep using the old 266.58 that was needed by the NV-r521 to not clog the CPUs? It will use a full core. Thats why its called 100% CPU bug. It took AMD almost 6 month to fix this bug so you will have to see how long it takes for nvidia to resolve this. With each crime and every kindness we birth our future. |
Wiggo Send message Joined: 24 Jan 00 Posts: 34744 Credit: 261,360,520 RAC: 489 |
I'll settle with just doing them only on my CPU's only until someone works out what the problem is and gets it 100% fixed before I even think about letting my GPU's have a go at them. Did you read the title of this thread? And where have you been otherwise lately? Sorry but the answer yes and has been for quite some while now as both are Open CL based apps. ;) Cheers. |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
JohnDK wrote: I'm using 268,36 on my almost new notebook and have just finished the first AP, all went fine and the CPU use was rather low, mostly under 5%. As Mike said, that's normal output for the app. However, it does indicate that BOINC asked the app to exit 4 times. It did so cleanly and restarted from a checkpoint when BOINC told it to. 3 of those exit/restart cycles used the same checkpoint (at 24.32% progress), IOW BOINC did not let it run long enough to produce another checkpoint. That might suggest tuning some BOINC settings to be sure that kind of sequence doesn't happen too often. You could look in the BOINC messages around 49 minutes after the task first started to see why BOINC thought it needed to shut the task down. AP processing is done in 111 large chunks, the OpenCL apps checkpoint at the beginning of each, so an exit/restart can cost 0.9009 percent progress at most. On average it would be half that, so the 4 exit/restarts during that task may have increased runtime about 1.8%. Definitely no reason to panic. Joe |
JohnDK Send message Joined: 28 May 00 Posts: 1222 Credit: 451,243,443 RAC: 1,127 |
JohnDK wrote:I'm using 268,36 on my almost new notebook and have just finished the first AP, all went fine and the CPU use was rather low, mostly under 5%. OK thanks. I think the restarts was done by myself :) |
©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.