[Jakub Vomacka]

firepro is different from common radeon cards mostly by drivers and ECC memory. so gpu chips for new mac pro could be pretty cheap, wonder what ram they use for their custom gpus in mac pro.

Radeon 7970 uses the same core and it is very interesting how it shows up in OsX 10.9 https://forums.macrumors.com/showthread.php?t=1594669

7970 is pretty cheap, $400 for whole card. imagine they just get the chips, in large quantity…

[Jakub Vomacka]

according to various info found on the net about gpu specifications and bandwidth, i think this is the card used: https://www.amd.com/US/PRODUCTS/WORKSTATION/GRAPHICS/ATI-FIREPRO-3D/W9000/Pages/w9000.aspx#3

here is some more info: https://www.anandtech.com/show/6137/the-amd-firepro-w9000-w8000-review-part-1/2

basically its workstation version of 7970 with better specs and large amount of ram

[Jakub Vomacka]

beware – when rendering in mpegstreamclip you will lose your timecode. all clips will start at 0.

sometimes that might be a problem

[Jakub Vomacka]

“… it is a pretty standard accelerator card for HPC and generic number crunching. It looks like a cluster of x86 servers that are on a TCP/IP network, and takes almost no programming expertise to port to. Optimization is very similar to a standard x86 CPU, and you can use the Intel tools that are very well established in the HPC world.”

basicly takes x86 code and after recompilation in intel software its able to run on many more parallel x86 CPU cores

more to read: https://www.theregister.co.uk/2012/11/12/intel_xeon_phi_coprocessor_launch/

https://newsroom.intel.com/community/intel_newsroom/blog/2012/11/12/intel-delivers-new-architecture-for-discovery-with-intel-xeon-phi-coprocessors?cid=rss-258152-c1-278335
https://download.intel.com/newsroom/kits/xeon/phi/pdfs/Intel-Xeon-Phi_Factsheet.pdf
https://download.intel.com/newsroom/kits/xeon/phi/pdfs/Intel-Xeon-Phi-Coprocessor_ProductBrief.pdf

[Jakub Vomacka]

i tried to replicate the test here on my hardware, going with CUDA for Premiere. i have hackintosh:
i7 920 – 3.8Ghz
6 GB RAM
Nvidia 560 Ti

test made on 10.8.2
FCPX 10.0.6
PPro CS6 v6.0.0
1 min of ProresHQ 1080p25

Gaussian Blur Export ProresHQ:
PPro – 0:49 (export through AME 0:41)
FCPX – 0:22

to try something else, apart from Prores Codec
Gaussian Blur Export H264 (960x540p25 – “better quality” settings in fcpx):
PPro – 1:00
FCPX – 1:06

One thing to note here can be the quality of encodes, with blur there is not much to visually compare.
Personally i think (tried many tests for various apps) that CUDA downscaling in PPro (especially for interlaced footage) cannot be matched in quality in any other NLE. Its pretty fast too! FCPX comes very blurry compared to PPro on this task. Therefore, CUDA gives “sometimes” access to better algorithms for certain tasks, so possibly a better quality.

PS: I tried activating OpenCL on my Nvidia card in Premiere and that gave me 4:50 on ProresHQ export. Might be different on AMD cards though, because Nvidia doesnt like OpenCL (still pretty new on their platform) as much as AMD.