clang-format

include/ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp

@@ -116,7 +116,7 @@ template <typename GridwiseGemm,
           bool HasDoubleTailKBlockLoop>
 __global__ void
 #if CK_USE_LAUNCH_BOUNDS
-__launch_bounds__(CK_MAX_THREAD_PER_BLOCK, CK_MIN_BLOCK_PER_CU)
+    __launch_bounds__(CK_MAX_THREAD_PER_BLOCK, CK_MIN_BLOCK_PER_CU)
 #endif
         kernel_grouped_conv_fwd_dl_multiple_d(
             const ABDataType* __restrict__ p_a_grid,
@@ -380,8 +380,8 @@ struct DeviceGroupedConvFwdDlMultipleD_NHWC_KYXC_NHWK
     }
 
     // desc for problem definition
-    using AGridDesc_AK0_M_AK1 = remove_cvref_t<decltype(MakeAGridDescriptor_AK0_M_AK1<ALayout>(
-        {}, {}, {}, {}, {}, {}, {}, {}, {}, {}))>;
+    using AGridDesc_AK0_M_AK1 = remove_cvref_t<decltype(
+        MakeAGridDescriptor_AK0_M_AK1<ALayout>({}, {}, {}, {}, {}, {}, {}, {}, {}, {}))>;
     using BGridDesc_BK0_N_BK1 =
         remove_cvref_t<decltype(MakeBGridDescriptor_BK0_N_BK1<BLayout>({}, {}))>;
     using DsGridDesc_M_N = remove_cvref_t<decltype(MakeDsGridDescriptor_M_N({}, {}))>;

include/ck/tensor_operation/gpu/device/impl/device_gemm_multiple_d_dl.hpp

@@ -34,7 +34,7 @@ template <typename GridwiseGemm,
           bool HasDoubleTailKBlockLoop>
 __global__ void
 #if CK_USE_LAUNCH_BOUNDS
-__launch_bounds__(CK_MAX_THREAD_PER_BLOCK, CK_MIN_BLOCK_PER_CU)
+    __launch_bounds__(CK_MAX_THREAD_PER_BLOCK, CK_MIN_BLOCK_PER_CU)
 #endif
         kernel_gemm_dl_multiple_d(
             const ABDataType* __restrict__ p_a_grid,