TOP500 uses FP64 performance in ranking. nVidia's 1 exaflop claim is the ~4 petaflops of FP8/INT8 * 256. FP64 performance of modern nvidia gpu's is actually f

TOP500 uses FP64 performance in ranking. nVidia's 1 exaflop claim is the ~4 petaflops of FP8/INT8 * 256. FP64 performance of modern nvidia gpu's is actually far, far less. The ratio to FP32 isn't even 2:1 anymore (not since Pascal I think) since they realize most machine learning is done with FP32 or less. 64 bit (or 'double precision') is still king in the HPC world though, as it is what you will find in large numerical solutions in fields like computation fluid dynamics, nuclear physics, etc.