https://www.hpcwire.com/2017/06/29/reinders-avx-512-may-hidden-gem-intel-xeon-scalable-processors
the article almost seemed machine-generated to me; did it actually say anything that we didn't already know (basically, that avx-512 in skylake/purley can perform double the vector rate of avx(-256)?
quote from the article; GPUs aren?t designed to operate on non-FP data.
yes, strange indeed. in fact, if the article is about vectorization, one could easily argue that the GPU model for vectorization (since surely a SM has to be considered a vector core) is in some ways working out better than the accretion of SIMD hacks in x86_64... _______________________________________________ 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
