@Prentice: > Certainly looking forward to running some benchmarks on these systems myself > when I can.
Apparently you can get access to these machines immediately in Azure... https://twitter.com/hpcnotes/status/1457755481544577039 5:02 PM · Nov 8, 2021·Twitter Web App Andrew Jones @hpcnotes Like the sound of the high performance #AMD EPYC #Milan-X processors from today’s @LisaSu keynote? As far as I know, the only place you can actually get access to use Milan-X today is via Microsoft #Azure #HPC HBv3 You can sign up at http://aka.ms/MilanXPreview -- Jim James Cownie <jcow...@gmail.com> Mob: +44 780 637 7146 > On 8 Nov 2021, at 20:45, Prentice Bisbal via Beowulf <beowulf@beowulf.org> > wrote: > > Did anyone else catch the AMD Datacenter Premier Keynote this morning? AMD > made some pretty impressive claims, like their MI200 GPUs are 4.5x faster > than NVIDIA A100s. and one of these GPUs will be ~4x faster than an entire > Summit node. Their Milan (Zen4) CPUs will have 96 up to 96 cores, and their > Bergamo "cloud native" CPUs will have up to 128 cores. > > It was live streamed on YouTube this morning, and is available for watching > at the link below, which is how I watched it: > > https://youtu.be/ECHhuvuiNzs > > I think it's definitely worth a watch to see what's coming, and to see AMD's > performance claims. Certainly looking forward to running some benchmarks on > these systems myself when I can. > > -- > Prentice > > _______________________________________________ > Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing > To change your subscription (digest mode or unsubscribe) visit > https://beowulf.org/cgi-bin/mailman/listinfo/beowulf _______________________________________________ Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing To change your subscription (digest mode or unsubscribe) visit https://beowulf.org/cgi-bin/mailman/listinfo/beowulf
