Update embedding_forward_quantized_cpu_template.cpp to use initialized output memory instead of uninitialized (#5054)

Summary:

X-link: facebookresearch/FBGEMM#2064

We observe, if the memory of output is uninitialized, the output may be garbage.

This is because certain memory is untouched. The proposed fix is a quick workaround, but it will be more efficient to directly fill the untouched memory with zero.

Reviewed By: sryap

Differential Revision: D85447298

Author: Hang Qu

fbgemm_gpu/codegen/inference/embedding_forward_quantized_cpu_template.cpp

@@ -209,7 +209,11 @@ Tensor int_nbit_split_embedding{{ "_nobag" if nobag else "" }}_codegen_forward_{
     if (o_dtype == SparseType::INT8) {
       total_adjusted_D += T * kINT8QparamsBytes;
     }
-    output = at::empty({B, total_adjusted_D}, dev_weights.options().dtype(getScalarType(o_dtype)).pinned_memory(pinned_memory));
+    if (!output_is_int4 && !output_is_int8) {
+      output = at::zeros({B, total_adjusted_D}, dev_weights.options().dtype(getScalarType(o_dtype)).pinned_memory(pinned_memory));
+    } else {
+      output = at::empty({B, total_adjusted_D}, dev_weights.options().dtype(getScalarType(o_dtype)).pinned_memory(pinned_memory));
+    }
     {% else %}
     constexpr int kINT8QparamsBytes = 4; // no bag int8 output aligns with fbgemm weights storage size and layout
     constexpr int kINT4QparamsElems = 8; // scale + bias takes 4 bytes which are 8 int4 elements
@@ -219,8 +223,11 @@ Tensor int_nbit_split_embedding{{ "_nobag" if nobag else "" }}_codegen_forward_{
     } else if (o_dtype == SparseType::INT4) {
       adjusted_D += kINT4QparamsElems;
     }
-    output = at::empty({total_L, adjusted_D}, dev_weights.options().dtype(getScalarType(o_dtype)).pinned_memory(pinned_memory));
-
+    if (!output_is_int4 && !output_is_int8) {
+      output = at::zeros({total_L, adjusted_D}, dev_weights.options().dtype(getScalarType(o_dtype)).pinned_memory(pinned_memory));
+    } else {
+      output = at::empty({total_L, adjusted_D}, dev_weights.options().dtype(getScalarType(o_dtype)).pinned_memory(pinned_memory));
+    }
     {% endif %}