@glennklockwood.com What do you make of Satoshi Matsuoka’s latest paper and the later HPCWire article? arxiv.org/pdf/2606.06510
arxiv.org
Max Hawkins
@hawkinsmax.bsky.social
CSE PhD student at Georgia Tech working to beneficially integrate computing at scale into society (or at least that's the goal).
@glennklockwood.com What do you make of Satoshi Matsuoka’s latest paper and the later HPCWire article? arxiv.org/pdf/2606.06510
arxiv.org
Flop/s are dead; long live the flop. My work proposing a fundamental computing performance framework based on information theory is now on the arxiv (arxiv.org/pdf/2508.05621). In it, Rich Vuduc and I suggest we evaluate computing like communication - with Shannon's channel capacity. #hpc
arxiv.org
Looks like #NVIDIA nerfed Blackwell Ultra #FP64 performance to make room for low-precision FP formats. Each B300 GPU now has ~1.25 TFLOPs of dense FP64 vector and tensor core performance (a ~90% decrease from the B200). Not exactly unexpected, but drastic. Source: www.nvidia.com/en-us/data-c...
NVIDIA HGX Platform
Enabling advanced AI and HPC in every data center.
nvidia.com