clang format

src/include/migraphx/shape.hpp

@@ -227,7 +227,6 @@ struct shape
 
     private:
     std::shared_ptr<const shape_impl> impl;
-
 };
 
 void migraphx_to_value(value& v, const shape& s);

src/targets/gpu/device/mul_add.cpp

@@ -28,13 +28,12 @@ __global__ void mul_add_kernel(void* a, int an, void* x, int xn, void* b, int bn
     __half2* hb = reinterpret_cast<__half2*>(b);
     __half2* hx = reinterpret_cast<__half2*>(x);
     __half2* hr = reinterpret_cast<__half2*>(r);
-    if (id < n)
+    if(id < n)
     {
         hr[id] = __hadd2(__hmul2(ha[id % an], hx[id % xn]), hb[id % bn]);
     }
 }
 
-
 void mul_add(hipStream_t stream,
              const argument& result,
              const argument& arg1,
@@ -42,7 +41,7 @@ void mul_add(hipStream_t stream,
              const argument& arg3)
 {
     auto type = result.get_shape().type();
-    if (type == shape::half_type)
+    if(type == shape::half_type)
     {
         std::cout << "case1" << std::endl;
         int s1e      = arg1.get_shape().element_space() / 2;
@@ -52,10 +51,12 @@ void mul_add(hipStream_t stream,
         s1e          = (s1e == 0 ? 1 : s1e);
         s2e          = (s2e == 0 ? 1 : s2e);
         s3e          = (s3e == 0 ? 1 : s3e);
-        std::cout << "re =" << elem_num << ", s1e = " << s1e << ", s2e = " << s2e << ", s3e = " << s3e << std::endl;
+        std::cout << "re =" << elem_num << ", s1e = " << s1e << ", s2e = " << s2e
+                  << ", s3e = " << s3e << std::endl;
         int block_size = 1024;
         int block_num  = (elem_num + block_size - 1) / block_size;
-        mul_add_kernel<<<block_num, block_size>>>(arg1.data(), s1e, arg2.data(), s2e, arg3.data(), s3e, result.data(), elem_num);
+        mul_add_kernel<<<block_num, block_size>>>(
+            arg1.data(), s1e, arg2.data(), s2e, arg3.data(), s3e, result.data(), elem_num);
     }
     else
     {