Speaking of cheap gpgpu - did someone over here benchmark the Geforce Titan? It's not having ECC i assume, yet it is unlimited in double precision, running at 1/3 speed and just $999 and neary 850Mhz.
should be lot of flops and double precision flops. On Mar 6, 2013, at 10:00 AM, Brice Goglin wrote: > Le 06/03/2013 07:00, Mark Hahn a écrit : >> and drop it into most any gen1/2/3 PCIe x16 slot and it'll work >> (assuming >> I provide the right power and cooling, of course.) >>> The issue here is that because we offer 8GB of memory on the >>> cards, some >>> BIOSes are unable to map all of it through the PCI either due to >>> bugs or >>> failure to support so much memory. This is not the only people >>> suffering >> interesting. but it seems like there are quite a few cards out there >> with 4-6GB (admittedly, mostly higher-end workstation/gp-gpu cards.) >> is this issue a bigger deal for Phi than the Nvidia family? >> is it more critical for using Phi in offload mode? > > Those non-Phi cards have a lot of internal memory but they only > export a > small part of it directly to the host (look at lspci -vvv, you'll see > only about 200 MB for your big NVIDIA Teslas). This small part can be > mapped, everything else needs DMA. > > Phi lets you map everything (at least in the kernel, not sure if their > driver lets you map to userspace too) but the BIOS has more work to > setup such big mappings. > > Brice > > _______________________________________________ > 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 _______________________________________________ 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
