| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: GPU Support Forum
Thread: Graphics Card Performance |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 254
|
|
| Author |
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
Just dont expect much help from AMD/ATI or anyone besides a OpenCl programmer and the like. If that is true then why does ATI have far superior driver support for OpenCl than Nvidia?
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
  |
||
|
|
BladeD
Ace Cruncher USA Joined: Nov 17, 2004 Post Count: 28976 Status: Offline Project Badges: 
|
Performance of the new GTX is unexpected low, explanation ? Did you read the reviews when they came out? They were designed that way, so as not to compete with their pro cards! If you want a complex task to be worked on, it tends to be better to use CUDA. I wonder where that idea came from?  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Personally, I have absolutely no problem with them having gaming cards and professional cards segregated. I can't buy an unlocked Xeon to OC and run on a desktop can I?
----------------------------------------It's the smart move. Games dont use DP, so why have it on the card? Every company separates their professional lineup from their consumer one. It's good business. Nano: They have far better OpenCl support because they allow it. NVIDIA uses CUDA. Which is currently, better (read: more mature). But thats because its older, and has had millions of dollars thrown into making it better. I do suspect OpenCl will catch up eventually, but currently it has not. [Edit 2 times, last edit by Former Member at Nov 30, 2012 11:31:36 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The 4xx series and 5xx are better than the 6xx series at crunching at WCG for sure. Shader clock is faster on older models. The app has to be written in a certain way to take advantage of 6xx. |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges:
|
They have far better OpenCl support because they allow it. NVIDIA uses CUDA. Which is currently, better (read: more mature). But thats because its older, and has had millions of dollars thrown into making it better. And what did ATI use before OpenCL?
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
----------------------------------------[Edit 1 times, last edit by nanoprobe at Dec 1, 2012 12:27:00 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
ATI used to use their own equivalent to CUDA called BROOK. They moved away from it a year or two ago (from what i can remember) and only have OpenCL now. There is also OpenMM specifically developed for molecular modeling that was developed by Stanford University that seems to essentially use the same basis as OpenCL, so it works on both AMD and NVIDIA.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
If I'm not mistaken their own proprietary language. But this either a) failed b) was too expensive for them to continue for money returned c) thought open cl would be more appropriate.
Neither of which means anything in regards to the current discussion. This is what I mean when I say more mature http://blog.accelereyes.com/blog/2012/02/17/opencl_vs_cuda_webinar_recap/ |
||
|
|
mmstick
Senior Cruncher Joined: Aug 19, 2010 Post Count: 151 Status: Offline Project Badges: 
|
ATI Stream/CAL/Brook was simply deprecated because it is silly to have more than one GPGPU standard to program for. All funding for that was immediately dropped to help fund OpenCL. Wouldn't it be better if all vendors agreed to use one universal standard so it is easier to program for, instead of forcing programmers to learn multiple different branches that attempt to reach the same solution? It's better to have everyone work together. Thus, AMD hardware has been designed specifically with OpenCL in mind for GPGPU. Once programmers learn how to code the OpenCL way, GPGPU programming will be better in the end.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Agreed. But that time has yet to arrive.
|
||
|
|
BladeD
Ace Cruncher USA Joined: Nov 17, 2004 Post Count: 28976 Status: Offline Project Badges:
|
Agreed. But that time has yet to arrive. It has for anyone that wants their app to run on all GPUs. |
||
|
|
|