nemesis wrote:
> Sabrina Kilian escreveu:
>> nemesis wrote:
>> Right, they bring a dramatic speed increase to other systems and
>> programs, so you THINK they will here as well. Prove it, or wait.
> 
> I would think pov-ray's triangle handling to be similar to that of other
> raytracers but if that's not the case, I'll just wait for your word on
> it, sweetie.

How do you suggest that the ray-tracing part separates the object that
it intersects? Without a complete overhaul, it will have to support all
of the objects that POV-Ray supports. Otherwise, you have to decide at
the time the ray hits an object whether you process that on the GPU or
CPU. Then, what do you gain by offloading just the triangle and sphere
code to the GPU?

>> Depends on the card, again. If the card is only offering double or
>> triple the FLOPs of the FPU/CPU, then the speed loss by faking it in
>> software will not be better. Wide range of hardware, remember?
> 
> Wide range of hardware has never prevented povray running on low-end
> hardware or super-duper minicomputers.  Users never complained of the
> vastly different speeds.

Super computers are not in the same hands as higher end video cards. The
range of cards that run GPGPU type code is pretty staggering compared to
the range of CPUs on the market right now. You trust the compilers and
alpha libraries to work properly, I do not.

> So you'll be targetting CUDA?  It's not cross-platform as OpenCL,
> although much more mature for now.

I will be shooting for either one, depending on which I like the syntax
of better. And which one is better documented. CUDA is mature, I have an
available card and several people with different cards available to test
on. OpenCL would be better for the long term, but a quick look over some
specs suggested that OpenCL might not be able to take advantage of the
shared system memory.

More memory, less chance to see the memory bound errors on my alpha
code. And then reducing that available resource in chucks will let me
start seeing where I can make other improvements.

>> If you happen to have a computer with a high end GPU that I can run
>> comparison benchmarks on, great. Otherwise, my development time will be
>> limited to how often I can ship code off to friends and get benchmarks
>> and profiles.
> 
> sadly, I'm already an old wig with a Q6600 and a cheap nvidia card (9400
> GT if I remember correctly).

Perfect, that card has even fewer stream processors than mine; You just
volunteered.

Post a reply to this message