Hehe no, but it was not full speed either.
Best regards,
Vlado
Just a little question…i think that wasn’t clearly answered before: Are we limited to the RAM that the PHI brings with it, or can it use all available system memory?
Yea I +1 on ram question how is it at the end?
Also WOHOOOOOOOOOOOOOOO YEA BABEEEEEEE LOVE UUUUUU !! 
You are limited to the RAM on the card. Currently 6 GB are usable out of 8 GB (2 GB are reserved for the Xeon Phi OS). However, the card can access files on the main HDD, so things like tiled EXR files, .vrmesh files and so on should work fine thus reducing the memory requirements compared to a GPU.
Best regards,
Vlado
Im wondering how in present state phi is better then geforce card strictly for rendering argument?
It can run the full production V-Ray. Do you need another argument?
Best regards,
Vlado
Epic Epic Epic.
Vlado can you list moththerboards that you got PHI to work with? - I hear they are very picky.
Also which one is it ? 3100 or 5100?
wow, thats big!
wow, finally!
Another question tho… will PHI work with this
http://www.fusionio.com/products/iofx/
Its around 1650GB of ram… could it utilize this by any chance?
Thanks, bye.
I have no idea; it would be best to ask those guys directly.
Best regards,
Vlado
surely, but my understanding is that its not just plug and play, so you do need to write something (not sure what) to make it work with vray as it has its own os right? So not really knowing how much work will go into that, its still time and effort to make it work and obviously with that you will introduce bugs which will need to be resolved etc etc, at present phy has 60 1 ghz cores, so that is why I am wondering, the cost and time weighs in favor of gpu?
For the GPU, we had to rewrite everything from the ground up, and we still have work to do. (This is also in part why it’s so much faster than a CPU - it is a way simpler renderer.) Every feature that we want to support on the GPU, we need to code separate and try to make sure, as much as possible, that it produces a similar result to the production renderer. Whatever we have to do for the Xeon Phi, the efforts involved will be much much much less.
Best regards,
Vlado
Lets just hope intel doesn’t drop this design after few years of trial and error.
Even if they do, we would have only spent a small amount of time on it. It would be a much bigger letdown if nVidia dropped CUDA, which I sincerely hope won’t happen anytime soon
Best regards,
Vlado
FusionIO is NOT RAM…it’s an SSD cache used e.g. in fileservers…nothing else…1.4GB/s are pretty slow compared to your systems RAM
Heya
Yea its like SSD but a tad faster… SSD goes at 500mb/s PCIE SSD goes at 900mb/s if we lucky… this goes quite a bit faster so it should serve as an best option if there would be a need to use PHI with large data sets… if it would work.
The best option would be to use the system RAM no? I’m not sure how the FusionIO thing would even connect to the Phi card…
Best regards,
Vlado
Is this using OpenCL mode?
I saw this news.
http://www.khronos.org/news/permalink/intel-releases-sdk-with-conformant-opencl-1.2-support-for-intel-xeon-phi-co
We’ve not tried running V-Ray RT GPU on it yet.
We’ll do next week probably and we’ll share our results then.