> A system made up of 6,144 TPU v5ps reached the GPT-3 training checkpoint in 11.77 minutes, placing a distant second to an 11,616-Nvidia H100 system, which accomplished the task in about 3.44 minutes. That top TPU system was only about 25 seconds faster than an H100 computer half its size.
I thought TPUs was supposed to be faster than GPUs
> A system made up of 6,144 TPU v5ps reached the GPT-3 training checkpoint in 11.77 minutes, placing a distant second to an 11,616-Nvidia H100 system, which accomplished the task in about 3.44 minutes. That top TPU system was only about 25 seconds faster than an H100 computer half its size.
I thought TPUs was supposed to be faster than GPUs