3 ms·

The TFLOPS numbers are not directly comparable. The TPUs use reduced precision in some areas, whereas I am guessing the Titan V numbers are based on single prec

The TFLOPS numbers are not directly comparable. The TPUs use reduced precision in some areas, whereas I am guessing the Titan V numbers are based on single precision operations.

cobookman 9y ago
If all you need is reduced precision, then it’s fair to compare the two. I’d also assume memory bandwidth matters just as much as TFLOPS for ml workloads.
modeless 9y ago
Titan V numbers are also reduced precision (16 bits), using their tensor cores.