• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

OpenCL/CUDA and SLI

geoffry

Senior member
I did a quick GOOG search but couldn't find anything conclusive.

Does anyone here know how well CUDA apps scale on SLI? From what I found it would be somewhat easy to get decent scaling as CUDA offers good control of the GPUs. Also, it looks like Badaboom works fast on a 9800GX2 but I'm uncertain if it uses both GPUs or just 1.

And is there any information out on whether OpenCL scales decently off 2 NVDA cards?

Thanks in advance.
 
They don't. The CUDA programmer has to manually implement it if they want to scale their CUDA kernel to multiple GPUs. Many CUDA apps sometimes BSOD with SLI enabled, as well (including both of the BOINC ones in my case🙁).
 
Most CUDA apps won't run on SLI. Take F@H for instance. You must disable SLI, and then run separate instances, each set with an affinity for each GPU.
 
Back
Top