clang-format

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

@@ -116,23 +116,23 @@ 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,
-        const ABDataType* __restrict__ p_b_grid,
-        DsPointer p_ds_grid,
-        EDataType* __restrict__ p_e_grid,
-        const AElementwiseOperation a_element_op,
-        const BElementwiseOperation b_element_op,
-        const CDEElementwiseOperation cde_element_op,
-        const index_t batch_count,
-        const AGridDesc_K0_M0_M1_K1 a_grid_desc_k0_m0_m1_k1,
-        const BGridDesc_K0_N0_N1_K1 b_grid_desc_k0_n0_n1_k1,
-        const DsGridDesc_M0_M10_M11_N0_N10_N11 ds_grid_desc_m0_m10_m11_n0_n10_n11,
-        const CGridDesc_M0_M10_M11_N0_N10_N11 e_grid_desc_m0_m10_m11_n0_n10_n11,
-        const Block2CTileMap block_2_ctile_map,
-        const ComputePtrOffsetOfBatch compute_ptr_offset_of_batch)
+        kernel_grouped_conv_fwd_dl_multiple_d(
+            const ABDataType* __restrict__ p_a_grid,
+            const ABDataType* __restrict__ p_b_grid,
+            DsPointer p_ds_grid,
+            EDataType* __restrict__ p_e_grid,
+            const AElementwiseOperation a_element_op,
+            const BElementwiseOperation b_element_op,
+            const CDEElementwiseOperation cde_element_op,
+            const index_t batch_count,
+            const AGridDesc_K0_M0_M1_K1 a_grid_desc_k0_m0_m1_k1,
+            const BGridDesc_K0_N0_N1_K1 b_grid_desc_k0_n0_n1_k1,
+            const DsGridDesc_M0_M10_M11_N0_N10_N11 ds_grid_desc_m0_m10_m11_n0_n10_n11,
+            const CGridDesc_M0_M10_M11_N0_N10_N11 e_grid_desc_m0_m10_m11_n0_n10_n11,
+            const Block2CTileMap block_2_ctile_map,
+            const ComputePtrOffsetOfBatch compute_ptr_offset_of_batch)
 {
 #if(!defined(__HIP_DEVICE_COMPILE__) || defined(__gfx906__) || defined(__gfx1030__) || \
     defined(__gfx90a__) || defined(__gfx908__))
@@ -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

@@ -34,21 +34,21 @@ 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,
-        const ABDataType* __restrict__ p_b_grid,
-        DsPointer p_ds_grid,
-        EDataType* __restrict__ p_e_grid,
-        const AElementwiseOperation a_element_op,
-        const BElementwiseOperation b_element_op,
-        const CDEElementwiseOperation cde_element_op,
-        const AGridDesc_K0_M0_M1_K1 a_grid_desc_k0_m0_m1_k1,
-        const BGridDesc_K0_N0_N1_K1 b_grid_desc_k0_n0_n1_k1,
-        const DsGridDesc_M0_M10_M11_N0_N10_N11 ds_grid_desc_m0_m10_m11_n0_n10_n11,
-        const CGridDesc_M0_M10_M11_N0_N10_N11 e_grid_desc_m0_m10_m11_n0_n10_n11,
-        const Block2CTileMap block_2_ctile_map)
+        kernel_gemm_dl_multiple_d(
+            const ABDataType* __restrict__ p_a_grid,
+            const ABDataType* __restrict__ p_b_grid,
+            DsPointer p_ds_grid,
+            EDataType* __restrict__ p_e_grid,
+            const AElementwiseOperation a_element_op,
+            const BElementwiseOperation b_element_op,
+            const CDEElementwiseOperation cde_element_op,
+            const AGridDesc_K0_M0_M1_K1 a_grid_desc_k0_m0_m1_k1,
+            const BGridDesc_K0_N0_N1_K1 b_grid_desc_k0_n0_n1_k1,
+            const DsGridDesc_M0_M10_M11_N0_N10_N11 ds_grid_desc_m0_m10_m11_n0_n10_n11,
+            const CGridDesc_M0_M10_M11_N0_N10_N11 e_grid_desc_m0_m10_m11_n0_n10_n11,
+            const Block2CTileMap block_2_ctile_map)
 {
 #if(!defined(__HIP_DEVICE_COMPILE__) || defined(__gfx906__) || defined(__gfx908__) || \
     defined(__gfx90a__) || defined(__gfx1030__))