[ROCm] torch.sum optimization by increasing min_values_per_thread (#1591

)

Follow-up to pytorch#135397.
AMD gpus perform better with fewer thread blocks.
So increase the min_values_per_thread as well. 
This helped improved
[CvT](https://github.com/facebookresearch/FAMBench/tree/main/benchmarks/cvt)
benchmark performance on MI300X

Co-author: @carlobertolli

aten/src/ATen/native/cuda/Reduce.cuh

@@ -1091,14 +1091,16 @@ ReduceConfig setReduceConfig(const TensorIterator& iter){
     config.output_mult[0] = config.split_output(block_width);
   }
 
+#ifdef USE_ROCM
+  // AMD gpus perform better with fewer thread blocks
+  constexpr int min_values_per_thread = 128;
+  constexpr int max_values_per_thread = 1024;
+#else
   constexpr int min_values_per_thread = 16;
-#ifndef USE_ROCM
   constexpr int max_values_per_thread = 256;
-#else
-  constexpr int max_values_per_thread = 1024;
 #endif
 
-  if (config.values_per_thread() >= block_height * 16 || config.values_per_thread() >= max_values_per_thread) {
+  if (config.values_per_thread() >= block_height * min_values_per_thread || config.values_per_thread() >= max_values_per_thread) {
     // Divide the input across warps in a thread-block, if that leaves at least
     // 16 elements to be summed by each thread. This will require inter-warp
     // reduction using shared memory.