On Thu, Sep 13, 2012 at 08:48:37AM -0600, Orion Poplawski wrote: > It seems the question before us is core scaling versus raw MHz. I've seen > many codes saturate the memory bandwidth before filling all of the cores and > it seems you have a trade off between MHz and cores these days at a given > price point. But perhaps that doesn't even matter if you're memory bandwidth > limited anyways.
Apropos memory bottleneck, did everyone catch that Haswell has stacked memory? AMD has pretty much to do the same with their APUs, so OpenCL seems to have a pretty bright future in HPC. _______________________________________________ Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
