3 ms·
I checked MLPerf website, and it looks like A100 is outperforming TPUv3, and is also more capable (there does not seem to be a working implementation of RL for
by lostmsu 6y ago
I checked MLPerf website, and it looks like A100 is outperforming TPUv3, and is also more capable (there does not seem to be a working implementation of RL for Go on TPU).
To be fair, TPUv4 is not out yet, and it might catch up using the latest processes (7nm TSMC or 8nm Samsung).
https://mlperf.org/training-results-0-7 https://mlperf.org/training-results-0-7