Comments for Parallel Rendering

Comment on pthread_mutex vs atomic operations by eile

eile — Thu, 19 Nov 2009 12:52:03 +0000

I’m by now means a specialist in non-blocking algorithms, but in my understanding a lot of non-blocking data structures rely on atomics for their implementation.

Re. your research – I’m looking forward to see the results!

Comment on pthread_mutex vs atomic operations by r bag

r bag — Thu, 19 Nov 2009 12:42:24 +0000

By the way,We are doing a lot of work on that at INRIA/Alchemy (https://alchemy.futurs.inria.fr/), I cannot give details about the work, but it seems to be very promising
As soon as we get “real” good results (better than atomic operations) I’ll notify you

Comment on pthread_mutex vs atomic operations by r bag

r bag — Thu, 19 Nov 2009 12:33:27 +0000

Of course atomic operations are better than locks

What I mean by not scalable is that they don’t scale well compared to other techniques (like non blocking algorithms). The problem of non blocking algorithms is that they are difficult to write !
By the way I’m also working on atomic operations in GCC And it seems to be the best solution for now

Oncea talks about that in his paper, SPAA 2009 : http://portal.acm.org/citation.cfm?id=1583991.1584050

Comment on pthread_mutex vs atomic operations by eile

eile — Thu, 19 Nov 2009 09:53:19 +0000

It’s a bit broad as a comment. Do you have a citation for your statement?

In fact, atomic ops as used here provide much better scalability then locks.

Comment on pthread_mutex vs atomic operations by r bag

r bag — Thu, 29 Oct 2009 15:13:34 +0000

I have done similar experiments during my master thesis, and the results were also good
but I read somewhere, don’t remeber where exactly, that atomic operations are not scalable !
Do you have any idea about that ?

Comment on Equalizer 0.9 Released! by eile

eile — Wed, 12 Aug 2009 09:31:14 +0000

No, in Equalizer you can use one render thread per GPU. This benchmark was done on a dual-GPU Mac using one Equalizer process:

http://www.equalizergraphics.com/documents/WhitePapers/MultiGPU.pdf

Comment on Equalizer 0.9 Released! by tuan kuranes

tuan kuranes — Wed, 12 Aug 2009 09:28:21 +0000

Thanks for the answer.
So the only way I really can do that in a perfomance wise way is to do as I currently do : two completely separate executable, each “duplicating” GPU resource use, that speaks to each other using IPC…

Comment on Equalizer 0.9 Released! by eile

eile — Tue, 11 Aug 2009 12:08:52 +0000

You don’t need the GPU affinity extension. Your question is answered here, or just check the Equalizer source code:

http://www.equalizergraphics.com/documentation/parallelOpenGLFAQ.html#multigpu

Q: How do OpenGL programs behave with multiple graphics cards on Mac OS X?
A: The OpenGL rendering happens on the card where most of the pixels of the window are located. Areas of the window located on other graphics cards are copied from the main renderer.

Q: How to I address a specific graphics card on Mac OS X (AGL)?
A: OS X 10.4 and earlier: Use the handle obtained by DMGetGDeviceByDisplayID for .
A: OS X 10.5: Use the display mask returned by CGDisplayIDToOpenGLDisplayMask as the value for the AGL_DISPLAY_MASK pixel format attribute.

Comment on Equalizer 0.9 Released! by Equalizer 0.9 Available for Parallel and Scalable OpenGL Applications | The Geeks Of 3D - 3D Tech News

Tue, 11 Aug 2009 11:11:59 +0000

[...] [via] [...]

Comment on Equalizer 0.9 Released! by tuan kuranes

tuan kuranes — Tue, 11 Aug 2009 10:48:05 +0000

Very, very interesting.

I wonder what’s the current status of multiple GPU/windows on MacOsx in Equalizer, as I have hard time selecting GPU or even setting two fullscreen window in the same program. (as MAcOsx doesn’t update it’s driver often enough, the opengl extension about gpu affinity doesn’t exists there…)