Doesn't anybody use TPUs from Google? Given the heterogeneous nature of GPUs, RAM, tensor cores, etc. it would be nice to have a direct comparison of, say, num

Doesn't anybody use TPUs from Google? Given the heterogeneous nature of GPUs, RAM, tensor cores, etc. it would be nice to have a direct comparison of, say, number of teraflops-hour per dollar, or something like that.