Turning Data Centers into ‘AI Factories’: Nvidia Introduces Hopper, H100 GPU

Grace Hopper's nanosecond video is always a fun watch:

* https://www.youtube.com/watch?v=9eyFDBPk4Yw

Hm, why is NVidia only talking about FP16 in regards to the new transformer blocks. Isn't BFloat16 better in most cases?