Re:TPU Deep Dive
by monocasa on Sunday June 22, 2025 @10:32AM
(#44347246)
The actual cost of the part (within reason) doesn't matter all that much for a hyperscaler. The real cost is in the perf/watt, which an FPGA is around an order of magnitude worse for the same RTL.
[ Parent
| on HN ]
There are no comments on this comment yet.