crossfire or sli?

Message boards : Number crunching : crossfire or sli?
Message board moderation

To post messages, you must log in.

AuthorMessage
merle van osdol

Send message
Joined: 23 Oct 02
Posts: 809
Credit: 1,980,117
RAC: 0
United States
Message 1572421 - Posted: 15 Sep 2014, 22:09:46 UTC
Last modified: 15 Sep 2014, 22:10:19 UTC

Does using crossfire or sli do anything worthwhile for lunatics AP or MB?
Compared of course to 2 separate gpu's.
ID: 1572421 · Report as offensive
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1572429 - Posted: 15 Sep 2014, 22:46:33 UTC - in response to Message 1572421.  

Nope. At least I've never seen any advantage to doing it and from what others have said, it doesn't seem to make the least bit of difference.
ID: 1572429 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1572441 - Posted: 15 Sep 2014, 23:15:20 UTC

Those features are used to make your collection of GPUs work as one for rendering purposes. Where the slight lag between the devices talking to one another isn't very noticeable.
In the case of BOINC & SETI@home you are preforming a separate function on each GPU. Having a single task processed across several GPUs may cause a massive slow down in processing while data moves across the PCIe. Instead of internally on the GPU.

In the past I think you actually needed to disable those features in order for BOINC to use all of the GPUs. However with newer drivers I think it no longer matters.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1572441 · Report as offensive
merle van osdol

Send message
Joined: 23 Oct 02
Posts: 809
Credit: 1,980,117
RAC: 0
United States
Message 1572449 - Posted: 15 Sep 2014, 23:32:26 UTC - in response to Message 1572441.  

Thanks for the info. Just wanted to make certain I wasn't missing some opportunity.
ID: 1572449 · Report as offensive
MarkJ Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 08
Posts: 1139
Credit: 80,854,192
RAC: 5
Australia
Message 1572631 - Posted: 16 Sep 2014, 12:01:32 UTC
Last modified: 16 Sep 2014, 12:06:35 UTC

I believe CUDA 6 adds support for using cards in SLI mode (ie treating them as a single larger GPU). It may take some coding and I don't think anyone has produced a CUDA 6 Multi Beam app (yet). Jason would probably be the person to ask.

XT Libraries: Automatic Scaling to Multiple GPUs

CUDA 6 introduces XT Library interfaces which provide automatic scaling of cuBLAS level 3 and 2D/3D cuFFT routines to 2 or more GPUs. This means that if you have one or more dual-GPU accelerator cards in your workstation or cluster node, you can automatically take advantage of them for intensive FFTs and matrix-matrix multiplication. cuBLAS XT also enables multiplication of matrices that are too large to fit in the memory of a single GPU, because it operates directly on matrices allocated in CPU memory, tiling the matrix and overlapping computation with memory transfers. The result is linear scaling of very large GEMM operations to multiple GPUs, as the following figure shows.

BOINC blog
ID: 1572631 · Report as offensive

Message boards : Number crunching : crossfire or sli?


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.