Cuda x86 compilers out......octane future speed boost

[mdharrington]

http://www.prnewswire.com/news-releases ... 99879.html

Now that cuda is compiled for x86, I think it would be ideal that octane put this power to use....not sure what a cpu to gpu speed comparison would be, but surely a good i7 would contribute something.

How much work would it be to add this capability to octane?

[radiance]

We are looking into various options currently

Radiance

[t_3]

http://www.prnewswire.com/news-releases/the-portland-group-updates-compilers-to-deliver-nvidia-cuda-architecture-for-x86-platforms-123999879.html

Now that cuda is compiled for x86, I think it would be ideal that octane put this power to use....not sure what a cpu to gpu speed comparison would be, but surely a good i7 would contribute something.

How much work would it be to add this capability to octane?

if you compare luxrender and slg, a fast single cpu reaches about 15-20% of a fast single gpu. the question is, as this is still a cuda compiler, if it dumps the annoying texture limits (ram of course, but number of textures!), or not...

[mdharrington]

so much uncertainty and propaganda out there in gpu vs cpu......
if octane implemented this is would be the only real apples to apples test I have seen...
either way....should decrease render times....perhaps even allow for CPU network rendering along side GPU primary rendering...

Radiance:
Good to hear....look forward to what you guys come up with

[jmfowler]

I would be interested in this, running a Intel 980 hexacore chip with DT.

Of course I also think 1) instancing and 2) switch off primary visibility for objects, is a must before this...