Nvidia has a substantial lead over rivals Intel, Advanced Micro Devices, and IBM when it comes to peddling graphics co-processors, and it wants to keep that lead and extend it, if possible. That means doing boring old stuff that server and operating system makers have to do, such as lining up application software vendors so they …
OK, I'll bite....
It seems to me that using those GPU cycles is a really groovy way to crunch. But....
Comparing apples & pears is always interesting/difficult/wasteoftime/essential/etc.
The pears are mildly well described:- _128 Opteron cores on its Cray XT3 super_, _1,024 Power cores on its BlueGene/L super_.
The lemon, sorry apple ( 'tho I doubt it ), are somewhat more nebulous:- _personal supercomputer with two Tesla C1060 GPU co-processors_.
Of course, I am an old fart who has been locked in a cage for a few millennium, so what do I know? Help me out here folks. A _personal supercomputer_ (that well known er, appliance?) consists of what approximately? Running what approximately? At what cost approximately?
Not that I want to get picky or anything.
Paris, 'cos the angle is in the dangle.
help for old farts
read it and weep
What - no Crysis jokes?
"Minimum specs for Tesla PSC.....1200W PSU" - ouch. On the bright side you'd save on heating bills. Although at ~£1k per unit and you needing 3-4 of them the capital costs (as reasonable as they are) probably outweigh any savings you might make.
Still, if I could convince the lass... Stick a PS3 emulator on it and it's a slam dunk.