Why are projects 9415 and 9414 such low PPD?

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

Moderators: Site Moderators, FAHC Science Team

Joe_H
Site Admin
Posts: 7856
Joined: Tue Apr 21, 2009 4:41 pm
Hardware configuration: Mac Pro 2.8 quad 12 GB smp4
MacBook Pro 2.9 i7 8 GB smp2
Location: W. MA

Re: Why are projects 9415 and 9414 such low PPD?

Post by Joe_H »

The servers know what CLASS of card it is and makes assignments that the card can process. Your "SHOULD know" item would require much additional coding to be done on the AS and WS code before it could be implemented. PG has to weigh the cost in time and programming resources against the gain they might get in WU processing throughput. Doing so would also slow down work on other needed items like Core_22, updating the public bets client so it can become a full release, and other items under development.

In the end what people are complaining about is that these projects do not give as many points as they would like to see on the highest end cards. But they do give more PPD than the next lower rank of cards such as the 1070 and 1060.
Image

iMac 2.8 i7 12 GB smp8, Mac Pro 2.8 quad 12 GB smp6
MacBook Pro 2.9 i7 8 GB smp3
bruce
Posts: 20910
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: Why are projects 9415 and 9414 such low PPD?

Post by bruce »

Kuno wrote:The issue I have with this is the server SHOULD know what card you have from the GPU txt file. It polls the card to see what you have, and bingo. Or does it only poll it to the point where it asks if you have an nvidia or AMD card? But as for your card not getting used to it's full potential, blame windows and it's horrible driver foundation. Once I moved things over to Linux, the GPU usage % issue no longer became an issue. Windows sucks for folding, stop using it.
SHOULD know and DOES know are two different things.

GPUs are screened against the data in GPUs.txt but that information is NOT uploaded to the server. To do that would require a new FAHClient after making changes to several pieces of code on the servers. You are recommending an enhancement -- and probably one that will be considered costly.

The data uploaded by your client is much more limited. In fact, the information uploaded to the servers is designed to answer only one question: Which FAHCore or Cores can be used by this GPU to process Work Units?.

At the present time, there are only two FAHCores for GPUs: a 32-bit and 64-bit variant of FAHCore_21, (the information about whether your OS is 64-bit or only 32-bit is uploaded too.) Since FAHCore_21 is OpenCL compatible, the server doesn't even need to know whether you have a GPU from ATI/AMD or a GPU from nVidia, only that you have a GPU that supports the minimum OpenCL version requirement..
Post Reply