I'm sure some of you have seen this already but if you haven't, an update on GPU 3 progress.
http://folding.typepad.com/news/2010/07/index.html
http://folding.typepad.com/news/2010/07/index.html
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Navigation
Install the app
How to install the app on iOS
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
More options
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
GPU 3 update
- Thread starter Nathan_P
- Start date
Saw it yes. I guess the high hopes many people had for the OpenCL-based GPU client have been dashed yet again. It's just not mature enough yet to giver the performance CUDA gives, apparently. TBH, I think that there'll remain a CUDA client for the foreseeable future just because it's so efficient compared to other solutions.
sirmonkey1985
[H]ard|DCer of the Month - July 2010
- Joined
- Sep 13, 2008
- Messages
- 22,412
Saw it yes. I guess the high hopes many people had for the OpenCL-based GPU client have been dashed yet again. It's just not mature enough yet to giver the performance CUDA gives, apparently. TBH, I think that there'll remain a CUDA client for the foreseeable future just because it's so efficient compared to other solutions.
its not that its so efficient.. its just that it takes a lot less work to use the cuda client then to design an openCL client from the ground up while the cuda client still works perfectly fine.. so the opencl client just tends to sit at the bottom of the list of things to get done.. though the openCL client onces the issues are ironed out should end the cuda client..
its not that its so efficient.. its just that it takes a lot less work to use the cuda client then to design an openCL client from the ground up while the cuda client still works perfectly fine.. so the opencl client just tends to sit at the bottom of the list of things to get done.. though the openCL client onces the issues are ironed out should end the cuda client..
Maybe so, but from what I have heard around the internet it seems that OpenCL implementations in drivers are still iffy, especially for nVidia, and of course OpenCL is trying to barge into a field which CUDA has dominated virtually unopposed for the last years.
Now, I do realize that Stanford & F@H doesn't have infinite resources to work on things like this, and if I knew more about OpenCL I might be able to assist the OpenCL client project. I still don't think, however, that the OpenCL client will so easily end the reign of CUDA with F@H. In other fields I don't see a huge uptake of OpenCL either.