clang format

src/targets/gpu/device/logsoftmax.cpp

@@ -54,8 +54,9 @@ void logsoftmax(hipStream_t stream, const argument& result, const argument& arg,
 
                 auto item_num = (remaining_item_num > block_size) ? block_size : remaining_item_num;
                 // reduce_max(lds_data, block_size, thr_idx, item_num, max_block_size);
-                block_reduce<type, max_op<type>>(lds_data, max_op<type>{}, block_size, thr_idx, item_num, max_block_size);
-                
+                block_reduce<type, max_op<type>>(
+                    lds_data, max_op<type>{}, block_size, thr_idx, item_num, max_block_size);
+
                 remaining_item_num -= block_size;
             }
 
@@ -77,7 +78,8 @@ void logsoftmax(hipStream_t stream, const argument& result, const argument& arg,
 
                 auto item_num = (remaining_item_num > block_size) ? block_size : remaining_item_num;
                 // reduce_sum(lds_data, block_size, thr_idx, item_num, max_block_size);
-                block_reduce<type, sum_op<type>>(lds_data, sum_op<type>{}, block_size, thr_idx, item_num, max_block_size);
+                block_reduce<type, sum_op<type>>(
+                    lds_data, sum_op<type>{}, block_size, thr_idx, item_num, max_block_size);
 
                 remaining_item_num -= block_size;
             }

src/targets/gpu/include/migraphx/gpu/device/reduce_opers.hpp

@@ -30,11 +30,11 @@ struct sum_op
 
 template <class T, class Op>
 inline __device__ void block_reduce(T* data_ptr,
-                                  Op op,
-                                  std::size_t block_size,
-                                  std::size_t thr_idx,
-                                  std::size_t item_num,
-                                  std::size_t max_index)
+                                    Op op,
+                                    std::size_t block_size,
+                                    std::size_t thr_idx,
+                                    std::size_t item_num,
+                                    std::size_t max_index)
 {
     while(true)
     {
@@ -42,7 +42,7 @@ inline __device__ void block_reduce(T* data_ptr,
         auto size   = item_num / 2;
         for(std::size_t i = thr_idx; i < size; i += block_size)
         {
-            //data_ptr[i] = ::max(to_hip_type(data_ptr[i]), to_hip_type(data_ptr[i + stride]));
+            // data_ptr[i] = ::max(to_hip_type(data_ptr[i]), to_hip_type(data_ptr[i + stride]));
             data_ptr[i] = op(data_ptr[i], data_ptr[i + stride]);
         }
         __syncthreads();
@@ -62,7 +62,6 @@ inline __device__ void block_reduce(T* data_ptr,
     __syncthreads();
 }
 
-
 template <class T>
 inline __device__ void reduce_max(T* data_ptr,
                                   std::size_t block_size,