World's Fastest AI Supercomputer 'Perlmutter' is Online

First came the news in the summer of Japanese supercomputer'Fugaku' dethroning IBM's Summit as the World's most powerful Supercomputer -- scoring High-Performance Linpack (HPL) of 415.5 petaflops, beating the now second-placed Summit system by a factor of 2.8x. One company that has made some huge strides in the supercomputing & AI space is the chip giant Nvidia. I covered these developments in one of my pieces, late last year --which included its Next-gen AI chip dubbed as A100, AI Platform NVIDIA Maxine & the building of the UK's most powerful Armband Supercomputer. The collaboration between Nvidia and the U.S National Energy Research Scientific Computing Center (NERSC) has just brought online what's being dubbed as the world's fastest supercomputer for AI workloads. Dubbed as'Perlmutter' -- it is named after astrophysicist Saul Perlmutter with a processing capacity of 6,144 NVIDIA A100 Tensor Core GPUs and will be tasked with stitching together the largest ever 3D map of the visible universe, among other projects.