BRP v 1.31 feedback

Alex
Alex
Joined: 1 Mar 05
Posts: 88
Credit: 398734
RAC: 0
Topic 84900

got 4 ati and one cuda.
cuda finished without any problem.
I tried to run 2 wu's on both ati gpu's. Well, one of them is the on-chip apu. This caused the system to crash; bluescreen pointed to an amd-driver problem.
Switching back to one wu per gpu solved the problem. Anyway, not a problem of the application.
3 of five are finished, waiting for validation.

Viking69
Joined: 2 Mar 05
Posts: 19
Credit: 2024604
RAC: 0

BRP v 1.31 feedback

I now have 3 WU's on my only PC that hase a CUDA GPU. One is a 1.28 and 2 are 1.31.

see them here

I see NO progress with these. If I suspend all other work on this PC I do not see that these unit start to process, either as a CPU app or a GPU app. I monitor my cores and GPU temperature, and after I suspended all other owrk all my cores and the GPU go to cool temperatures. This tells me that nothing is going on.

If I look at my profile I only see these WU's and it looks like mone of my other PC's are participating.

Win 7 x64, 16GB Ram, Nvidia GTX550 Ti ver 310.33

Viking69
Joined: 2 Mar 05
Posts: 19
Credit: 2024604
RAC: 0

RE: I now have 3 WU's on my

Message 79395 in response to message 79394

Quote:

I now have 3 WU's on my only PC that hase a CUDA GPU. One is a 1.28 and 2 are 1.31.

see them here

I see NO progress with these. If I suspend all other work on this PC I do not see that these unit start to process, either as a CPU app or a GPU app. I monitor my cores and GPU temperature, and after I suspended all other owrk all my cores and the GPU go to cool temperatures. This tells me that nothing is going on.

If I look at my profile I only see these WU's and it looks like mone of my other PC's are participating.

Win 7 x64, 16GB Ram, Nvidia GTX550 Ti ver 310.33

I noticed that my othe GPU WU's were also having this problem. I restarted BOINC fully and the WU's started processing.

Stephan Goll
Stephan Goll
Joined: 13 Dec 05
Posts: 20
Credit: 1874367
RAC: 0

The scheduler seems to go

The scheduler seems to go wild.

381513 145690 9 Nov 2012 | 18:00:24 UTC 9 Nov 2012 | 19:28:30 UTC Error while computing 1.37 0.05 --- Binary Radio Pulsar Search v1.00 (BRP3cuda32fullCPU)
381169 145518 9 Nov 2012 | 16:24:28 UTC 9 Nov 2012 | 19:28:30 UTC Error while computing 1.37 0.05 --- Binary Radio Pulsar Search v1.00 (BRP3cuda32fullCPU)
380960 145159 9 Nov 2012 | 13:11:17 UTC 9 Nov 2012 | 16:24:28 UTC Error while computing 2.87 0.07 --- Binary Radio Pulsar Search v1.00 (BRP3cuda32fullCPU)
380430 145199 9 Nov 2012 | 10:49:22 UTC 9 Nov 2012 | 13:11:17 UTC Completed and validated 5,796.33 288.62 500.00 Binary Radio Pulsar Search v1.31 (BRP3cuda32nv270)

I got one cuda270 workunit and it was fine. Then I got cuda32full workunits and they all failed.
Stephan

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Joined: 28 Aug 06
Posts: 164
Credit: 1864017
RAC: 0

RE: The scheduler seems to

Message 79397 in response to message 79396

Quote:

The scheduler seems to go wild.

381513 145690 9 Nov 2012 | 18:00:24 UTC 9 Nov 2012 | 19:28:30 UTC Error while computing 1.37 0.05 --- Binary Radio Pulsar Search v1.00 (BRP3cuda32fullCPU)
381169 145518 9 Nov 2012 | 16:24:28 UTC 9 Nov 2012 | 19:28:30 UTC Error while computing 1.37 0.05 --- Binary Radio Pulsar Search v1.00 (BRP3cuda32fullCPU)
380960 145159 9 Nov 2012 | 13:11:17 UTC 9 Nov 2012 | 16:24:28 UTC Error while computing 2.87 0.07 --- Binary Radio Pulsar Search v1.00 (BRP3cuda32fullCPU)
380430 145199 9 Nov 2012 | 10:49:22 UTC 9 Nov 2012 | 13:11:17 UTC Completed and validated 5,796.33 288.62 500.00 Binary Radio Pulsar Search v1.31 (BRP3cuda32nv270)

I got one cuda270 workunit and it was fine. Then I got cuda32full workunits and they all failed.
Stephan

Oh, sorry... that 1.00 BRP3cuda32fullCPU version was supposed to be deprecated (deactivated) on the server side. Ignore those. If you have any of those left, feel free to abort them.

Thanks for reporting this.

Cheers
HBE

Holmis
Holmis
Joined: 4 Jan 05
Posts: 89
Credit: 2104736
RAC: 0

Hi So far the BRP v1.31 has

Hi

So far the BRP v1.31 has ran quite nicely on my host (id 2267).
But I do have two WU that looks a bit interesting from a testing stand point, namely:

WU id: 143172 with 3 inconclusive so far, 2 CUDA and one OpenCL

and

WU id: 147263 with 2 inconclusive, both CUDA

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.