50% GPU Utilization

If you think it might be a driver problem, see viewforum.php?f=79

Moderators: Site Moderators, FAHC Science Team

Re: 50% GPU Utilization

Postby PantherX » Tue Apr 07, 2020 3:47 am

Under the Sensors tab, look at GPU Load. Mine hovers between 98% and 100% on average while folding.
ETA:
Now ↞ Very Soon ↔ Soon ↔ Soon-ish ↔ Not Soon ↠ End Of Time

Welcome To The F@H Support Forum Ӂ Troubleshooting Bad WUs Ӂ Troubleshooting Server Connectivity Issues
User avatar
PantherX
Site Moderator
 
Posts: 6584
Joined: Wed Dec 23, 2009 10:33 am
Location: Land Of The Long White Cloud

Re: 50% GPU Utilization

Postby bruce » Tue Apr 07, 2020 7:00 am

r0k0 wrote:i have gpu-z but in the sensor tab, what do i look for ?


I generally look at two sensors. GPULoad and PerfCapReason

You do have to consider the size of the protein being folded, however. For large proteins, performance is limited by the number of shaders. For small proteins, GPULoad will be limited by the Memory Controller Load and possibly by the Bus Interface Load. You can't make use of a large number of shaders if you can't get the data fast enough to keep them all busy.
bruce
 
Posts: 19836
Joined: Thu Nov 29, 2007 11:13 pm
Location: So. Cal.

Re: 50% GPU Utilization

Postby r0k0 » Tue Apr 07, 2020 10:16 am

ahhh haha thanks guys
Main - Intel 10700K / Asus Z490-E / 32GBGB DDR4-3200 / 2080Ti
Folding - HP Z220 Xeon E3 1245 V2 / 16GB DDR3-1600 / GTX 1060 6GB
Folding - HP Z420 Xeon E5 2650 / 16GB DDR3-1600 ECC / GTX 1070Ti
Folding - HP Z800 2X Xeon X5675 / 48GB DDR3-1333 ECC / RX480
r0k0
 
Posts: 67
Joined: Thu Mar 19, 2020 6:01 pm
Location: Gatineau, QC

Re: 50% GPU Utilization

Postby Rel25917 » Wed Apr 08, 2020 6:20 pm

GPU load is the one to look at.
Rel25917
 
Posts: 299
Joined: Wed Aug 15, 2012 3:31 am

Re: 50% GPU Utilization

Postby saustin » Wed Jul 08, 2020 9:43 am

bruce wrote:
r0k0 wrote:i have gpu-z but in the sensor tab, what do i look for ?


I generally look at two sensors. GPULoad and PerfCapReason

You do have to consider the size of the protein being folded, however. For large proteins, performance is limited by the number of shaders. For small proteins, GPULoad will be limited by the Memory Controller Load and possibly by the Bus Interface Load. You can't make use of a large number of shaders if you can't get the data fast enough to keep them all busy.


I am new at this, but I was wondering how do you tell the size of the protein? I have been trying to figure out the very erratic PPD numbers I get from my 4 identical 2080 Supers. I know this plays some part in it, but I think there is more going on than just that. I have two Identical machines with 2 2080's each all 4 the exact same card as well. My PPD can go from 2.9M est to 999k EST on any given WU, all while GPU Z tells me that 1 card in each machine never sees above a 68-69% workload on the GPU. The other is at 90-98%. It's driving me batty trying to figure it out. I have tried different drivers, I even swapped motherboards (brands as well) just to see. The lower Estimated PPD WU always coincided with the low utilization card as well.

Any ideas?
saustin
 
Posts: 1
Joined: Wed Jul 08, 2020 9:32 am

Re: 50% GPU Utilization

Postby Joe_H » Wed Jul 08, 2020 2:11 pm

saustin wrote:I am new at this, but I was wondering how do you tell the size of the protein?


The size of the simulation is the number of atoms is listed in the Project Summary - https://apps.foldingathome.org/psummary - for every project. Currently that ranges from under 10,000 to over 400,000. That is not directly the "size of the protein as it includes the solvent around the protein, but is useful in a relative manner.
Image

iMac 2.8 i7 12 GB smp8, Mac Pro 2.8 quad 12 GB smp6
MacBook Pro 2.9 i7 8 GB smp3
Joe_H
Site Admin
 
Posts: 6535
Joined: Tue Apr 21, 2009 5:41 pm
Location: W. MA

Previous

Return to GPU Projects and FahCores

Who is online

Users browsing this forum: No registered users and 1 guest

cron