Formatting

src/targets/gpu/jit/ck_gemm.cpp

@@ -223,7 +223,10 @@ struct ck_gemm_compiler : compiler<ck_gemm_compiler>
         s       = shape{s.type(), {m1, m2}};
     }
 
-    std::vector<std::string> names() const { return {"ck_gemm", "gpu::ck_gemm", "ck_gemm_int8", "gpu::ck_gemm_int8"}; }
+    std::vector<std::string> names() const
+    {
+        return {"ck_gemm", "gpu::ck_gemm", "ck_gemm_int8", "gpu::ck_gemm_int8"};
+    }
 
     operation compile_op(context& /* ctx */, const std::vector<shape>& inputs, const value& v) const
     {
@@ -266,8 +269,6 @@ struct ck_gemm_compiler : compiler<ck_gemm_compiler>
         {
             cde_op = v.at("post").to<std::string>();
         }
-        
-            
 
         auto problem = ck::tensor_operation::device::device_gemm_multiple_d::Problem{
             static_cast<ck::index_t>(m),
@@ -343,7 +344,8 @@ struct ck_gemm_compiler : compiler<ck_gemm_compiler>
         return action_decorate(replace(compile_op(ctx, shapes, v)), [=] {
             if(enabled(MIGRAPHX_LOG_CK_GEMM{}))
             {
-                std::vector<shape> gemm_shapes{shapes[0], shapes[1], shapes.back().with_type(shapes[0].type())};
+                std::vector<shape> gemm_shapes{
+                    shapes[0], shapes[1], shapes.back().with_type(shapes[0].type())};
                 std::cout << "ck_gemm: " << to_json_string(to_value(gemm_shapes)) << std::endl;
             }
         });