Posted 09/30/2009 at 07:41:59pm

It looks like they're supporting SIMD and SIMT according to their speaker from Cray, which is what I believe they have previously supported. 

Posted 07/31/2008 at 01:10:05am

What would truly be interesting to see is the hardware that the interviewees were using before being prestented with the Vista-running notebook. If you're moving from a P3 to a 2.2ghz T7200, we're talking quite a difference, of course it's going to be faster. I can also only speculate on how much hand holding was done during the interview, that is to say, did they just let the individuals rome around or did they direct them?

Posted 07/22/2008 at 08:41:13pm

According to [H] the partners are certainly not jumpingship.

http://www.hardocp.com/news.html?news=MzM4MTcsLCxoZW50aHVzaWFzdCwsLDE= 

" We picked up the phone and called both EVGA and XFX. Word is that Charlie’s story linked above is, “Bogus” or "BS," depending on who you listen to."

However, it is a given that if Intel's larabee GPU is a serious contendor, they will most certainly end up targetting NVIDIA's key partners.

Posted 07/07/2008 at 03:43:13pm

Can you fit a 40cm long 3x120mm fan radiator inside that thing?

Posted 03/24/2008 at 12:38:12am

No, there isn't a bottle neck due to the smaller manufacturing process. What really matters is the speed of the proc itself. There is a huge difference between GPU and CPU procs mainly due to how they handle different things. GPU's are specially designed with massive parallel processing in mind while CPU's typically don't have to deal with those same sort of tasks (less threads and what not). Also, the manufacturing size itself doesn't denote anything, efficiency and speed of the proc both matter by far more. However, the smaller sizes tend to consume less power at idle. So, if one were to stick a core 2 duo or even a core 2 quad into a graphics card one would probably experience a slowdown in performance from the card because neither of those processors are designed with graphics work in mind. The reason why NVIDIA uses a larger process is because they have a harder time changing process sizes due to overhead. It just recently changed to 65nm and it will probably be a while before we see graphics cards for NVIDIA in the 45nm range. However, because AMD is both manufacturing CPUs and GPUs it is easier for them to change sizes as they can share equipment between their former ATI branch and their AMD CPU manufacturing.

Posted 10/23/2007 at 03:19:21am

Really? I think that it was right of MPC to ding them for putting something on a system that knowingly hampers its performance, especially considering this comps intended audience. Also, these reviews are serious business; they should know that MPC doesn't base scores off of shininess, but rather functionality and performance. I think that they sent one in with Radeons to show off its tweaked nforce chip set.