onnxruntime/js/web/lib/wasm/jsep/webgpu/ops
Jiajia Qin 785e2b1eae
[js/webgpu] Optimize softmax by vector (#18153)
### Description
This PR enables `softmax` outputs max supported components instead of
scalar for each thread.

Softmax with input[0]: [12,4096,4096] becomes 47.86 ms from 55.11 ms
2023-10-30 16:05:35 -07:00
..
3rd-party
argminmax.ts
bias-add.ts
bias-split-gelu.ts
binary-op.ts
common.ts
concat.ts
conv-grouped.ts
conv-transpose.ts
conv.ts
einsum.ts
expand.ts
fuse-utils.ts
gather-elements.ts
gather.ts
gemm.ts
instance-norm.ts
layer-norm.ts
matmul.ts
pad.ts
pool.ts
range.ts
reduce.ts
resize.ts
skip-layer-norm.ts
slice.ts
softmax.ts [js/webgpu] Optimize softmax by vector (#18153) 2023-10-30 16:05:35 -07:00
split.ts
tile.ts
transpose.ts
unary-op.ts
where.ts