lightseq
lightseq copied to clipboard
Performance measurement with Pytorch/MLIR, Tensorflow/XLA
How about performance measurement with Pytorch/MLIR, Tensorflow/XLA? These have fusing operations to run faster on GPU. We need to compare with MLIR or XLA to get the real comparison on the performance.
https://github.com/llvm/torch-mlir
We haven't tried mlir, but compared with xla with 1.3x speedup in our in-house translation env. If you have more results, welcome to share with us.