News for hackers. Stuff that matters.
Re:Facebook Trains ImageNet in 1 Hour by Eridrus on Tuesday June 13, 2017 @07:31AM (#14544039)
Interestingly enough I think the people who gain the most from this paper are cloud providers. Very few orgs will buy 256 GPUs, but with linear scaling, renting them makes a lot of sense.
Re:Facebook Trains ImageNet in 1 Hour by shezi on Tuesday June 13, 2017 @09:34AM (#14544579)
There are "only" 32 GPUs in the cluster, with 8 workers per GPU.
Re:Facebook Trains ImageNet in 1 Hour by dgacmu on Tuesday June 13, 2017 @11:37AM (#14545576)
They had 32 servers in the cluster, each with 8 P100 GPUs. Each GPU was one "worker" in their parlance.

("How to train ResNet-50 in one hour on two million dollars of hardware." :-)