Jump to content
3DCoat Forums

crippled OpenGL and CUDA for newer GeForce nVidia cards than 200 series!!!


3dCoatWannabe
 Share

Recommended Posts

  • Advanced Member

I just found out something that is very troubling.

Apparently, nVidia is CRIPPLING the OpenGL of cards > GeForce200.

So that a OpenGL calls on a GTX480 will be ~4 times slower than on a GTX285!!!!

From the reading I'm doing, it appears that nVidia is also cripping CUDA on these cards.

The Quadro has fast OpenGL but crippled CUDA - and the Tesla has full CUDA performance.

I'm including a few links, but a search for '580 opengl crippled' pulls up a lot of them

http://blenderartists.org/forum/showthread.php?236641-Quadro-2000-vs.-GTX580&p=1987116&viewfull=1#post1987116

http://blenderartists.org/forum/showthread.php?217104-CUDA-OpenGL-to-restore-Fermi-viewport-performance-SUSPENDED-FEASIBILITY-DISPROVEN&p=1848257&viewfull=1#post1848257

As my GTX-580 3GB arrives tomorrow, I find this VERY disturbing!!!!

Link to comment
Share on other sites

  • Advanced Member

That's troubling if you are a Lightwave user, but both 3ds Max and 3DC run better on DX by far

Can you choose openGL or DX with AD programs?

I'm really interested in the CUDA acceleration, which in theory should be useful for 3d Coat - as well as a lot of other software I use.

If CUDA hasn't gotten throttled entirely, it might be possible to work around the opengl issue:

http://www.opengl.or...1855#Post291855

Or maybe use one of the directX -> OpenGL frameworks for some of the worst performers - if Andrew has any interest?

http://www.opengl.or...1789#Post291789

But, shame on nVidia for putting out a product that's 4x slower than a 285 as this benchmark shows? As least without documenting it so that purchasers can make an informed choice. I can't spend days of my life trying to become an 'expert' in video cards to make a simple purchase. This really discourages me from purchasing video cards. Imagine if your new CPU were slower than your two year CPU? Intel/AMD wouldn't do that. Ahhh!

Link to comment
Share on other sites

  • Reputable Contributor

I don't know about the others, but you 3ds Max uses DX by default. You can change it to OpenGL, but AD hasn't developed it's software to take advantage of newer OpenGL technologies. They highly recommend the DX, as that is where the focus is at.

I keep asking Andrew about bringing CUDA up to date in 3D Coat. It's not been touched since he first implemented it before v3 was released over 3yrs ago. I've been waiting for this every since he finished the brush engine work some time ago.

Link to comment
Share on other sites

  • Advanced Member

Maybe that's why he hasn't focused on CUDA, due to the throttling?

I don't know? I just had someone post in the ipisoft forum (their mocop software uses CUDA to accelerate) that his GTX-460 card was way faster than his previous GTX-260.

The 580 has 512 cores, the 460 has 288 cores, the 285 has 240 cores, the 260 has 192 cores.

So, maybe even throttled down from what it could be, CUDA is still an improvement over CPU?

Hopefully Andrew will catch that post on a possible CUDA 'fix' for the opengl issue (I think they've only throttled some that are normally used only by 3d apps), and maybe we can see some CUDA/opengl advances in 3d-coat?

But, I'm sure Andrew is more knowledgeable than any of us on CUDA/opengl issues. I'd love to hear his feedback on the nVidia situation?

Link to comment
Share on other sites

  • Reputable Contributor

I don't understand what you mean CUDA is being throttle back in the same reference to OpenGL. They are separate entities. I knew there were some issues with OpenGL and NVidia for almost a year now, but I have heard nothing about CUDA being affected.

Link to comment
Share on other sites

  • Advanced Member

I don't have reliable info for CUDA or benchmarks.

I saw one post who said that CUDA was crippled for GeForce and Quadro, but fully enabled for Tesla.

The post from the user on ipiSoft seemed to indicate CUDA was considerably speeded up, at least for his 460.

The 'real world' benchmarks with CUDA accelerated Premiere seem to indicate that CUDA performance has stayed the same, or that the test may not be a reliable indicator of performance.

http://www.studio1pr...emiereCS5-2.htm

I think there's a difference in hardware also, they may have taken out some chip real estate that processed double floating point in CUDA, so .. it would depend on what you were doing with it.

Hopefully, CUDA gets faster with more cores. I wish I had some benchmarks for CUDA.

** EDIT *** this benchmark seems to indicate CUDA is getting faster in newer GTX cards!

http://kernelnine.com/?p=218

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...