Re:Show HN: High-performance ahead-of-time compiler for Mach...
by andrew-wja on Friday December 14, 2018 @05:14AM
(#18679777)
The short answer is that there are dozens of ways to use that GEMM to do convolution (convolution algorithms), and there are NUM_ALGORITHMS * NUM_LAYERS way to implement the network.
Our toolkit figures out which of those arrangements are the fast ones!
Our toolkit figures out which of those arrangements are the fast ones!