[GitHub] [incubator-mxnet] ZhennanQin commented on a change in pull request #16623: [mkldnn-v1.0]rm int8 sum workaround

[GitBox <gi...@apache.org>]

ZhennanQin commented on a change in pull request #16623: [mkldnn-v1.0]rm int8 sum workaround
URL: https://github.com/apache/incubator-mxnet/pull/16623#discussion_r339374290
 
 

 ##########
 File path: src/operator/quantization/mkldnn/mkldnn_quantized_elemwise_add.cc
 ##########
 @@ -161,8 +161,14 @@ static void MKLDNNQuantizedElemwiseAddForward(const nnvm::NodeAttrs& attrs, cons
   std::vector<mkldnn::memory::desc> in_desc;
   in_desc.push_back(dataA_mem->get_desc());
   in_desc.push_back(dataB_mem->get_desc());
-  auto output_desc = dataA_mem->get_desc();
-  output_desc.data.data_type = static_cast<mkldnn_data_type_t>(output_data_type);
+  size_t i_ndim = in_data[quantized_elemwise_add_enum::kDataA].shape().ndim();
+  mkldnn::memory::dims i_dims = mkldnn::memory::dims(i_ndim);
+  for (size_t i = 0; i < i_ndim; i++) {
+    i_dims[i] = static_cast<int>(in_data[quantized_elemwise_add_enum::kDataA].shape()[i]);
 
 Review comment:
   Too complex. Change to
   ```
   const auto in_shape = in_data[quantized_elemwise_add_enum::kDataA].shape();
   mkldnn::memory::dims i_dims(in_shape.begin(), in_shape.end());
   ```

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services