• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

News "Aurora’s Troubles Move Frontier into Pole Exascale Position" - HPCwire

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

moinmoin

Platinum Member
Jun 1, 2017
2,772
3,673
136
It seems that also in raw performance, since it takes 9,000 nodes for Mi200 and PVC to get to 1.5 exaflop, but 16,000 for A100 (according to the estimates in the article). That's almost 2:1 ratio.
The article also indirectly mentions that it's AMD's approach to MCM that made this possible: Technically it's not one CPU but 8 chiplets, and not four GPUs but 8 let's call them "GPUlets". Nvidia has no product (yet) that fits the same amount of performance in that space.
 

jpiniero

Lifer
Oct 1, 2010
10,109
2,366
136
If Argonne ends up cancelling Aurora completely and goes with AMD+nVidia instead, I imagine they would wait for Hopper.
 

jpiniero

Lifer
Oct 1, 2010
10,109
2,366
136
Well, Frontier is available now.
Given that they spent the money on the A100 cluster, they must consider Aurora and Frontier two different projects. Guess it's not out of the question they would just expand Frontier if Argonne ends up cancelling Aurora.
 
Thread starter Similar threads Forum Replies Date
solidsnake1298 CPUs and Overclocking 3

ASK THE COMMUNITY