clang format

src/targets/cpu/lowering.cpp

@@ -561,26 +561,26 @@ struct cpu_softmax
             par_for(batch_shape.elements(), [&](auto i) {
                 auto idx = compute_batch_indices(i, batch_shape);
 
-                for (size_t j = 0; j < n_dims; ++j)
+                for(size_t j = 0; j < n_dims; ++j)
                 {
                     idx[op.axis] = j;
                     batch_max[i] = std::max(batch_max[i], input(idx.begin(), idx.end()));
                 }
 
-                for (size_t j = 0; j < n_dims; ++j)
+                for(size_t j = 0; j < n_dims; ++j)
                 {
                     idx[op.axis]  = j;
                     size_t index  = output_shape.index(idx);
                     output[index] = std::exp(input[index] - batch_max[i]);
                 }
 
-                for (size_t j = 0; j < n_dims; ++j)
+                for(size_t j = 0; j < n_dims; ++j)
                 {
                     idx[op.axis] = j;
                     batch_sum[i] += output(idx.begin(), idx.end());
                 }
 
-                for (size_t j = 0; j < n_dims; ++j)
+                for(size_t j = 0; j < n_dims; ++j)
                 {
                     idx[op.axis] = j;
                     output(idx.begin(), idx.end()) /= batch_sum[i];
@@ -638,20 +638,20 @@ struct cpu_logsoftmax
 
             par_for(batch_shape.elements(), [&](auto i) {
                 auto idx = compute_batch_indices(i, batch_shape);
-                for (size_t j = 0; j < n_dims; ++j)
+                for(size_t j = 0; j < n_dims; ++j)
                 {
                     idx[op.axis] = j;
                     batch_max[i] = std::max(batch_max[i], input(idx.begin(), idx.end()));
                 }
 
-                for (size_t j = 0; j < n_dims; ++j)
+                for(size_t j = 0; j < n_dims; ++j)
                 {
                     idx[op.axis]  = j;
                     size_t index  = output_shape.index(idx);
                     output[index] = input[index] - batch_max[i];
                 }
 
-                for (size_t j = 0; j < n_dims; ++j)
+                for(size_t j = 0; j < n_dims; ++j)
                 {
                     idx[op.axis] = j;
                     batch_sum[i] += std::exp(output(idx.begin(), idx.end()));
@@ -659,7 +659,7 @@ struct cpu_logsoftmax
 
                 batch_sum[i] = std::log(batch_sum[i]);
 
-                for (size_t j = 0; j < n_dims; ++j)
+                for(size_t j = 0; j < n_dims; ++j)
                 {
                     idx[op.axis] = j;
                     output(idx.begin(), idx.end()) -= batch_sum[i];

src/targets/gpu/device/logsoftmax.cpp

@@ -35,7 +35,7 @@ argument logsoftmax(hipStream_t stream,
             // the current optimization
             const size_t max_block_size = 1024;
             size_t block_size           = 1;
-            while (block_size < max_block_size and block_size < n_dim)
+            while(block_size < max_block_size and block_size < n_dim)
             {
                 block_size *= 2;
             }
@@ -57,7 +57,7 @@ argument logsoftmax(hipStream_t stream,
                 lds_data[block_size] = input_ptr[0];
                 for(size_t i = thr_idx; i < thread_num; i += block_size)
                 {
-                    if (i < n_dims)
+                    if(i < n_dims)
                     {
                         data_idx[axis]    = i;
                         lds_data[thr_idx] = input_ptr[desc_data.linear(data_idx)];
@@ -97,10 +97,11 @@ argument logsoftmax(hipStream_t stream,
                 item_num                 = n_dims;
                 for(size_t i = thr_idx; i < thread_num; i += block_size)
                 {
-                    if (i < n_dims)
+                    if(i < n_dims)
                     {
                         data_idx[axis] = i;
-                        lds_data[thr_idx]    = input_ptr[desc_data.linear(data_idx)] - lds_data[block_size];
+                        lds_data[thr_idx] =
+                            input_ptr[desc_data.linear(data_idx)] - lds_data[block_size];
                         lds_data[thr_idx] = ::exp(to_hip_type(lds_data[thr_idx]));
                     }
 

src/targets/gpu/device/softmax.cpp

@@ -33,7 +33,7 @@ argument softmax(hipStream_t stream,
             // use one block for items in one batch.
             const size_t max_block_size = 1024;
             size_t block_size           = 1;
-            while (block_size < max_block_size and block_size < n_dims)
+            while(block_size < max_block_size and block_size < n_dims)
             {
                 block_size *= 2;
             }
@@ -56,7 +56,7 @@ argument softmax(hipStream_t stream,
                 lds_data[block_size + 1] = 0;
                 for(size_t i = thr_idx; i < thread_num; i += block_size)
                 {
-                    if (i < n_dims)
+                    if(i < n_dims)
                     {
                         data_idx[axis]    = i;
                         lds_data[thr_idx] = input_ptr[desc_data.linear(data_idx)];
@@ -95,10 +95,11 @@ argument softmax(hipStream_t stream,
                 item_num = n_dims;
                 for(size_t i = thr_idx; i < thread_num; i += block_size)
                 {
-                    if (i < n_dims)
+                    if(i < n_dims)
                     {
                         data_idx[axis] = i;
-                        lds_data[thr_idx]    = input_ptr[desc_data.linear(data_idx)] - lds_data[block_size];
+                        lds_data[thr_idx] =
+                            input_ptr[desc_data.linear(data_idx)] - lds_data[block_size];
                         lds_data[thr_idx] = ::exp(to_hip_type(lds_data[thr_idx]));
                     }