After few stop/restart ( hibernate, shutdown and also lost power ) the last time the WU calculate for about 12 hours with 39,162.21 second CPU time and 3,614.39 for calculate the result seem not good !?!
I try another one !
Your result failed to validate against one computed by a HD 6900 series one, but probably it's not your result that is off, but the HD 6900's. We now made a modified app for HD6900s to improve their validation rate. I'm confident that when the 3rd result comes in, your result will validate against it.
But still, the runtime is far too long. I would use a tool like GPU-Z to check whether the card is running at the spec'ed clock rates and using the full PCIe 2.0 bandwidth. It should be more than twice as fast according to my gut feeling.
Hi, there should be an error with the batch because i run another job since 14:36hours and no break point in view ! even 0 % done. I will just discard these 22 job.
Hi, there should be an error with the batch because i run another job since 14:36hours and no break point in view ! even 0 % done. I will just discard these 22 job.
I don't think it was a batch problem, because every workunit I checked had been successfully completed by a quorum partner ('wingmate').
Hi, there should be an error with the batch because i run another job since 14:36hours and no break point in view ! even 0 % done. I will just discard these 22 job.
What was the GPU load in GPU-Z like? any ATI driver restarts? If there was a driver restart on start there is unlikely to be any progress until you get the app to exit, and restart.
(you can post the GPU-Z load img by clicking the camera symbol in the top right, then by clicking 'upload to free image hosting' then posting the VBB link)
It was the next tab along (sensors) i wanted while you were running a 1.24 atiOpenCL task.
Edit: That Wu is running fine, it has GPU load, and the temp is up, if you find a Wu that has no GPU load, and is probably hogging a CPU core, then it's likely you have had a driver restart,
By suspending GPU usage, waiting for it to exit (watch task manager), then resuming GPU usage might get it running properly,
1:39-1:59 per WU, running 2
)
1:39-1:59 per WU, running 2 WUs concurrently on a stock HD5850
After few stop/restart (
)
After few stop/restart ( hibernate, shutdown and also lost power ) the last time the WU calculate for about 12 hours with 39,162.21 second CPU time and 3,614.39 for calculate the result seem not good !?!
I try another one !
Hi! Your result failed to
)
Hi!
Your result failed to validate against one computed by a HD 6900 series one, but probably it's not your result that is off, but the HD 6900's. We now made a modified app for HD6900s to improve their validation rate. I'm confident that when the 3rd result comes in, your result will validate against it.
But still, the runtime is far too long. I would use a tool like GPU-Z to check whether the card is running at the spec'ed clock rates and using the full PCIe 2.0 bandwidth. It should be more than twice as fast according to my gut feeling.
Cheers
HB
I have pick up the GPU-Z and
)
I have pick up the GPU-Z and upload the result there it is http://www.techpowerup.com/gpuz/2vpra/
waiting for your comment.
Hmm...I would have expected a
)
Hmm...I would have expected a higher memory throughput for this card. Also the test indicates PCIe 1.1 (!) , I would have expected PCIe 2.0.
Does anyone here have the same model and can run a comparison test?
Cherrs
HB
Hi, there should be an error
)
Hi, there should be an error with the batch because i run another job since 14:36hours and no break point in view ! even 0 % done. I will just discard these 22 job.
RE: Hi, there should be an
)
I don't think it was a batch problem, because every workunit I checked had been successfully completed by a quorum partner ('wingmate').
RE: Hi, there should be an
)
What was the GPU load in GPU-Z like? any ATI driver restarts? If there was a driver restart on start there is unlikely to be any progress until you get the app to exit, and restart.
(you can post the GPU-Z load img by clicking the camera symbol in the top right, then by clicking 'upload to free image hosting' then posting the VBB link)
Claggy
Here is what you want
)
Here is what you want !
and :
I already cancel the 21 job left, sorry !
It was the next tab along
)
It was the next tab along (sensors) i wanted while you were running a 1.24 atiOpenCL task.
Edit: That Wu is running fine, it has GPU load, and the temp is up, if you find a Wu that has no GPU load, and is probably hogging a CPU core, then it's likely you have had a driver restart,
By suspending GPU usage, waiting for it to exit (watch task manager), then resuming GPU usage might get it running properly,
Claggy