Quote:
Originally Posted by Stuart C
I would assume the quad core CPUs would be more of an advantage if you were to use the machine as a server on the network you submit jobs to. It would still need a fair amount of processing power to manage these jobs, and that wouldn't necessarily be something that would make efficient use of a massively parallel processing system (which is what GPUs tend to be).
Also, most versions of Windows and Linux are not optimised for massively parallel systems.
Also, you may need a fair amount of processing horsepower to send and recieve data to/from the tesla system. After all, if you need the kind of processing power these things offer, you will be dealing with terabytes (possibly many terabytes) of data.
|
if you look on the nvidia spec 1 quad core cpu per Tesla GPU with 240 cores which is the highest end of the Fermi's IIRC? Also if you look at the systems being sold they appear in the main to ship with either 1 or 2 xeon quad core cpus and not 4