Read more.Fugaku Linpack HPL result of 415petaflops is 2.8x faster than 2nd place IBM Summit.
Read more.Fugaku Linpack HPL result of 415petaflops is 2.8x faster than 2nd place IBM Summit.
oops Fujitsu *simply* forgot to include Epyc 64 core benchmark
When you include a specifically designed chip to speed up the primary purpose (the FP16 AI add on chip) of the supercomputer it 'should' be faster than anything else out there that is using less specialised hardware so it can be more universal in it's usage....
I believe that this post isn't a coincidence that came up just after the previous one
28MW! holy smoke batman, that is some serious energy usage!
There are currently 1 users browsing this thread. (0 members and 1 guests)