Gheorghe-Teodor Bercea

[ROCm] Improve softmax performance.

41

This patch improves the performance of softmax for 2D tensors by: - using a softmax calculation which eliminates the increase of shared memory usage with the size of the tensor...

Improve performance of reduce sum for 3D shapes

19

Improve performance of reduce sum for 3D shapes.

Fix compilation when enabling indirect function calls.