biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

cleanup at::Tag::needs_fixed_stride_order (#28974)

Browse Source 
Signed-off-by: Boyuan Feng <boyuan@meta.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
This commit is contained in:
Boyuan Feng
2025-11-20 02:51:36 -08:00
committed by GitHub
parent 322cb02872
commit a903d59ffa
2 changed files with 20 additions and 51 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
csrc/cpu/torch_bindings.cpp
View File

@@ -172,7 +172,6 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
// Quantization
#if defined(__AVX512F__) || (defined(__aarch64__) && !defined(__APPLE__)) || \
defined(__powerpc64__)
at::Tag stride_tag = at::Tag::needs_fixed_stride_order;
// Helper function to release oneDNN handlers
ops.def("release_dnnl_matmul_handler(int handler) -> ()",
&release_dnnl_matmul_handler);
@@ -208,15 +207,13 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
// Compute int8 quantized tensor for given scaling factor.
ops.def(
"static_scaled_int8_quant(Tensor! out, Tensor input, Tensor scale,"
"Tensor? azp) -> ()",
{stride_tag});
"Tensor? azp) -> ()");
ops.impl("static_scaled_int8_quant", torch::kCPU, &static_scaled_int8_quant);
// Compute int8 quantized tensor and scaling factor
ops.def(
"dynamic_scaled_int8_quant(Tensor! out, Tensor input, Tensor! scale, "
"Tensor!? azp) -> ()",
{stride_tag});
"Tensor!? azp) -> ()");
ops.impl("dynamic_scaled_int8_quant", torch::kCPU,
&dynamic_scaled_int8_quant);
#endif