I am not convinced that mic will still be optimized for all types of graphic oriented operations. Again your tossing a crap load of threads and power at a problem that a more optimized smaller chip might be better at.
Side note. We did a vray,Houdini , and C4d test using a quad 2.6ghz 16 core AMD (64cores) vs a dual 6 core 3.46ghz intel with hyper threading (24 threads) and it blew the socks off AMD. There are a lot of factors like code optimization. But at a certain point too many threads becomes inefficient and the raw clock speed of intel showed its teeth.
It all remains to be seen. That said my team is meeting with Nathan next Wed to see what this is really all about and I will be happy to share what we find out. Especially if we get to test the hardware.
Sent from my Sprint iPhone
On Mar 23, 2012, at 8:35 AM, Roland Bavington rolandb@sgi.com wrote:
The problem with all these things is the time it takes to port/write code for each iteration of hardware. Once Intel gets going with Many Integrated Core (MIC) http://www.intel.com/content/www/us/en/architecture-and-technology/many-integrated-core/intel-many-integrated-core-architecture.html I would have thought the market for build to order hardware to accelerate graphics will fall apart. MIC uses x86 codes and it is possible to port multi-threaded code directly to the new technology with a re-compile rather than a re-write.
-----Original Message----- From: Anthony Hoit [mailto:ahoit@thegraphicfilmcompany.com] Sent: 22 March 2012 18:46 To: studiosysadmins-discuss@studiosysadmins.com Subject: Re: [SSA-Discuss] Fwd: Progeniq's Accelerates Animation and VFX Rendering Speeds by 400%-800%
This is a little more information that Nathan sent me today about the tech.
-----------------------
Thought I'd send you some links on the technology we use, the chips are called Field Programmable Gate Arrays (FPGAs), they're made by this company called Xilinx up in the SF bay area; this is one such FPGA: http://www.xilinx.com/products/silicon-devices/fpga/artix-7/index.htm
And there are new generation FPGAs that have ARM cores built in, directly connected to the FPGA 'fabric' http://www.xilinx.com/products/silicon-devices/epp/zynq-7000/index.htm
We've been using several generations of FPGAs already over the years (our very early prototype Renderboost used some older generation Xilinx Virtex series FPGAs), so the new generation stuff coming out such as the ARM-FPGA hybrid chips are very interesting for us, and we'd most definitely have our next generation Renderboost leveraging these.
-----------------------
-anthony
On 3/22/2012 4:22 AM, Julian Firminger wrote: > Was about to mention exactly this. I too have a Pure, somewhere, in a > box, probably trying to mate with the BF444 SD Greed that got retired > at the same time. > > Surely massively parallel GPU's, on their current rate of improvement, > will knock any ASIC out of the water fairly quickly as things develop.
> Or are they saying they can keep apace by continual architectural > development updates on the FPGA's?? THAT could be interesting. > > Julian > >> -----Original Message----- >> From: studiosysadmins-discuss-bounces@studiosysadmins.com >> [mailto:studiosysadmins-discuss-bounces@studiosysadmins.com] On >> Behalf Of Jez Tucker >> Sent: Wednesday, 21 March 2012 10:57 PM >> To: discuss@studiosysadmins.com >> Subject: Re: [SSA-Discuss] Fwd: Progeniq's Accelerates Animation and >> VFX Rendering Speeds by 400%-800% >> >> Curious. >> >> This [perhaps not exactly this, details are sketchy] has been done >> before. >> >> Things that pop to mind are the ART Renderdrive / Pure and the >> original RenderMan RM-1 machine. >> In both instances, CPUs surpassed the performance of the hardware >> card very quickly. >> Indeed, I have 6x ART Pure cards collecting dust in a box. Anyone >> want one or an RD 2500? No? Thought not. >> >> So bar the quick very short-term gain, what's the point? >> How long before the cards are passed by general compute? >> >> Or are I am missing something...? >> >> --- >> Jez Tucker >> Senior SysAdmin >> Rushes >> www.rushes.co.uk >> >> >> From: studiosysadmins-discuss-bounces@studiosysadmins.com >> [studiosysadmins-discuss-bounces@studiosysadmins.com] on behalf of >> Anthony Hoit [ahoit@thegraphicfilmcompany.com] >> Sent: 21 March 2012 20:58 >> To: studiosysadmins-discuss@studiosysadmins.com >> Subject: Re: [SSA-Discuss] Fwd: Progeniq's Accelerates Animation and >> VFX Rendering Speeds by 400%-800% >> >> So I met with Darren Nathan. Got the low down. They have built a >> board that handles ray-tracing. He said that he had been working with >> Chaos Group for a while now and will have V-Ray support very soon. He >> said he will give me some test cards so I can get some numbers when >> that happens. He also mentioned Arnold support in the works and >> Renderman was already working. Now I'm kind of excited to get my >> hands on one of these guys to see what it can do. >> >> -anthony >> >> Rushes Postproduction Limited, 66 Old Compton Street, London W1D 4UH >> tel: +44 (0)20 7437 8676 >> web: http://www.rushes.co.uk >> The information contained in this e-mail is confidential and may be >> subject to legal privilege. If you are not the intended recipient, >> you must not use, copy, distribute or disclose the e-mail or any part >> of its contents or take any action in reliance on it. If you have >> received this e-mail in error, please e-mail the sender by replying >> to this message. All reasonable precautions have been taken to ensure >> no viruses are present in this e-mail. Rushes Postproduction Limited >> cannot accept responsibility for loss or damage arising from the use >> of this e-mail or attachments and recommend that you subject these to >> your virus checking procedures prior to use. > >