I haven´t got any BRP3 wu the last days, but if I remember correctly the priority *is* changed from the lowest level (S5GC) to the second lowest level (BRP3).
I know! But if Boinc runs two CPU tasks on my dual core together with one CUDA task, it only gets enough CPU-time when the priority is set to realtime. On every other priority-level is won´t go over 3% CPU-time and 6% GPU-load.
I can also only get the BRP3 Cuda app to 65% GPU usage if i have no CPU apps running,
if i have one core crunching eithier S5 HF search #1 or CPU Astropulse, then GPU utilisation drops to ~40%,
if both cores are crunching Einstein or Astropulse, GPU utilisation drops to 8%,
the BRP3 Cuda app is fine when paired with Raistmer's OpenCL ATI Astropulse app,
Can you look into testing apps with different thread priorities, Both Jason Gee's and Raistmer's Seti Multibeam and Astropulse apps have adjusted thread priorities,
this is done to stop the GPU getting starved of work:
Jason's x32f Cuda app:
setiathome_CUDA: Found 1 CUDA device(s):
Device 1: GeForce 9800 GTX/9800 GTX+, 495 MiB, regsPerBlock 8192
computeCap 1.1, multiProcs 16
clockRate = 1900000
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce 9800 GTX/9800 GTX+ is okay
SETI@home using CUDA accelerated device GeForce 9800 GTX/9800 GTX+ Priority of process raised successfully
Priority of worker thread raised successfully
Raistmer's OpenCL ATI Astropulse app:
FFA thread block override value:6144
FFA thread fetchblock override value:1536
Running on device number: 0 Priority of worker thread raised successfully
Priority of process adjusted successfully
OpenCL platform detected: NVIDIA Corporation
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns 0 device
Used slot is 0; AstroPulse v. 5.06
Non-graphics FFTW USE_CONVERSION_OPT
Windows x86 rev 456, 5.06 match, by Raistmer with support of Lunatics.kwsn.net team. SSE2
OpenCL version by Raistmer
oclFFT fix for ATI GPUs by Urs Echternacht
ffa threshold mod, by Joe Segur.
static fftw lib, built by Jason G.
SSE3 dechirping by JDWhale
Raistmer's OpenCL ATI Multibeam app:
Number of period iterations for PulseFind setted to:2
Number of app instances per device setted to:1
Running on device number: 0 Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: NVIDIA Corporation
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns 0 device, slots 0 to 0 (including) will be checked
Used slot is 0; Info : Building Program (clBuildProgram):main kernels: OK code 0
I'm not sure about the issue with the priorities as I'm currently only running one cpu task on a quad.
To the project Devs, great work with the app!
The performance of the BRP compared to the ABP app is a step in the right direction. Though on my GTX260-216 55nm it's only using about 60% of the GPU processing power according to GPU-z while using 10% of a CPU core (q9300 @ 3.0Ghz). For v 1.04 very nice! Now lets just cross our fingers and hope things work out and we get to see 95-98% usage soon! :D
My system received the first BRP3 WU I have seen this morning. The first one processed in 2061 seconds, which is considerably quicker than the previous generation CUDA work unit which took around 2.36 hours via the same GPU. GPU load on one GPU (580) was 35-45% and 60-65% on the other GPU (8800).
This is the kind of performance that would be expected from previous tests performed on the new app.
I noticed you are using driver 262.99 . Is anybody experiencing the problem described here with this driver version?
CU
HBE
Performance has been consistent so far after processing a dozen or so tasks. I have my BOINC client set to 50% CPU usage via my quad-core i7. With this setup, four CPU threads run and then two more GPU threads are able to run at the same time with the overall CPU load around 60% and the GPU load the same as what I posted before.
I'd rerun the driver install program, sometimes it backs out of installing the latest driver, or windows could have reverted it to an earlier driver,
but try restarting windows first.
I'd rerun the driver install program, sometimes it backs out of installing the latest driver, or windows could have reverted it to an earlier driver,
but try restarting windows first.
RE: I haven´t got any BRP3
)
I know! But if Boinc runs two CPU tasks on my dual core together with one CUDA task, it only gets enough CPU-time when the priority is set to realtime. On every other priority-level is won´t go over 3% CPU-time and 6% GPU-load.
I can also only get the BRP3
)
I can also only get the BRP3 Cuda app to 65% GPU usage if i have no CPU apps running,
if i have one core crunching eithier S5 HF search #1 or CPU Astropulse, then GPU utilisation drops to ~40%,
if both cores are crunching Einstein or Astropulse, GPU utilisation drops to 8%,
the BRP3 Cuda app is fine when paired with Raistmer's OpenCL ATI Astropulse app,
Can you look into testing apps with different thread priorities, Both Jason Gee's and Raistmer's Seti Multibeam and Astropulse apps have adjusted thread priorities,
this is done to stop the GPU getting starved of work:
Jason's x32f Cuda app:
setiathome_CUDA: Found 1 CUDA device(s):
Device 1: GeForce 9800 GTX/9800 GTX+, 495 MiB, regsPerBlock 8192
computeCap 1.1, multiProcs 16
clockRate = 1900000
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce 9800 GTX/9800 GTX+ is okay
SETI@home using CUDA accelerated device GeForce 9800 GTX/9800 GTX+
Priority of process raised successfully
Priority of worker thread raised successfully
Raistmer's OpenCL ATI Astropulse app:
FFA thread block override value:6144
FFA thread fetchblock override value:1536
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully
OpenCL platform detected: NVIDIA Corporation
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns 0 device
Used slot is 0; AstroPulse v. 5.06
Non-graphics FFTW USE_CONVERSION_OPT
Windows x86 rev 456, 5.06 match, by Raistmer with support of Lunatics.kwsn.net team. SSE2
OpenCL version by Raistmer
oclFFT fix for ATI GPUs by Urs Echternacht
ffa threshold mod, by Joe Segur.
static fftw lib, built by Jason G.
SSE3 dechirping by JDWhale
Raistmer's OpenCL ATI Multibeam app:
Number of period iterations for PulseFind setted to:2
Number of app instances per device setted to:1
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: NVIDIA Corporation
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns 0 device, slots 0 to 0 (including) will be checked
Used slot is 0; Info : Building Program (clBuildProgram):main kernels: OK code 0
Claggy
I'm not sure about the issue
)
I'm not sure about the issue with the priorities as I'm currently only running one cpu task on a quad.
To the project Devs, great work with the app!
The performance of the BRP compared to the ABP app is a step in the right direction. Though on my GTX260-216 55nm it's only using about 60% of the GPU processing power according to GPU-z while using 10% of a CPU core (q9300 @ 3.0Ghz). For v 1.04 very nice! Now lets just cross our fingers and hope things work out and we get to see 95-98% usage soon! :D
RE: RE: My system
)
Performance has been consistent so far after processing a dozen or so tasks. I have my BOINC client set to 50% CPU usage via my quad-core i7. With this setup, four CPU threads run and then two more GPU threads are able to run at the same time with the overall CPU load around 60% and the GPU load the same as what I posted before.
Today I got a new message in
)
Today I got a new message in my "message" tab :
23/12/2010 22:43:38 Einstein@Home Message from server: NVIDIA display driver 26000 or later needed
My NVDIA GPU is a GT230 (a bit weak, I know) but the driver version is 260.99.
So what ?
Thx for your answer (I didn't find the answer elsewhere)
Boinc is reporting your
)
Boinc is reporting your driver as: NVIDIA GeForce GT 230 (1508MB) driver: 25936
Claggy
Must I restart BOINC ? Or
)
Must I restart BOINC ? Or launch a benchmark ?
My driver is really 260.99 ...
I'd rerun the driver install
)
I'd rerun the driver install program, sometimes it backs out of installing the latest driver, or windows could have reverted it to an earlier driver,
but try restarting windows first.
Claggy
Exit BOINC fully and restart
)
Exit BOINC fully and restart it. Don't just exit BOINC Manager and restart that, as the underlying client will perfectly run without the GUI.
BOINC Manager->File->Exit, check "Stop running science applications when exiting manager"->OK. Then restart everything through Start->Programs->BOINC->BOINC Manager.
RE: I'd rerun the driver
)
Windows did it but I re-installed vr 260.99