[rothron]

While interesting, it's comparing 5000+ Tensor Processing Units against 64 CPU threads. I suspect this isn't a fair comparison by watts spent.

[bazzargh]

No - 5000 TPUs were used for training, but the 100-game matches used just 4 TPUs (bottom of p4 in the paper)

[sls]

from the paper:

> Training proceeded [...] using 5,000 first-generation TPUs

> We evaluated the fully trained instances of AlphaZero against Stockfish, Elmo and the previous version of AlphaGo Zero [...]. AlphaZero and the previous AlphaGo Zero used a single machine with 4 TPUs.

So the 5000 TPUs processing power or energy consumption should be compared with those spent in devising Stockfish, not running it.