Re:Correcting Intel's Deep Learning Benchmark Mistakes
by fnord123 on Wednesday August 17, 2016 @05:52AM
(#12303457)
Omnipath also sucks compared to Infiniband. How are they making so many inroads into HPC with these offerings? I mean, aside from their dominant-for-good-reason CPUs.
[ Parent
| on HN ]
Re:Correcting Intel's Deep Learning Benchmark Mistakes
by gnufx on Wednesday August 17, 2016 @09:56AM
(#12304691)
Please share the data, particularly for the built-in interfaces on ~70 cores which are supposed to be available this year (given the blanket statement). Omni-path seems to be worrying Mellanox, judging by a recent visit.