3 ms·
The Volta and higher architecture share ALUs with the floating point units, so nothing is really going to waste. Same with the tensor cores.
by shaklee3 7y ago
The Volta and higher architecture share ALUs with the floating point units, so nothing is really going to waste. Same with the tensor cores.