Merge branch 'feature/split-gemm-instance-library' of...

Merge branch 'feature/split-gemm-instance-library' of github.com:ROCmSoftwarePlatform/composable_kernel into feature/split-gemm-instance-library

include/ck/tensor_operation/gpu/grid/gridwise_gemm_pipeline_v2.hpp

@@ -79,9 +79,9 @@ struct GridwiseGemmPipeline_v2
 
             do
             {
-                #if CK_EXPERIMENTAL_PIPELINE_V2_IGLP_OPT
+#if CK_EXPERIMENTAL_PIPELINE_V2_IGLP_OPT
                 __builtin_amdgcn_iglp_opt(CK_EXPERIMENTAL_PIPELINE_V2_IGLP_OPT);
-                #endif
+#endif
 
                 block_sync_lds();
 

include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_v2r3.hpp

@@ -29,7 +29,7 @@ __global__ void
     __launch_bounds__(CK_MAX_THREAD_PER_BLOCK, CK_MIN_BLOCK_PER_CU)
 #endif
 #if CK_USE_WAVES_PER_EU
-    __attribute__((amdgpu_waves_per_eu(CK_MIN_WAVES_PER_EU, CK_MAX_WAVES_PER_EU)))
+        __attribute__((amdgpu_waves_per_eu(CK_MIN_WAVES_PER_EU, CK_MAX_WAVES_PER_EU)))
 #endif
         kernel_gemm_xdlops_v2r3(const FloatAB* __restrict__ p_a_grid,
                                 const FloatAB* __restrict__ p_b_grid,
@@ -65,7 +65,7 @@ __global__ void
     __launch_bounds__(CK_MAX_THREAD_PER_BLOCK, CK_MIN_BLOCK_PER_CU)
 #endif
 #if CK_USE_WAVES_PER_EU
-    __attribute__((amdgpu_waves_per_eu(CK_MIN_WAVES_PER_EU, CK_MAX_WAVES_PER_EU)))
+        __attribute__((amdgpu_waves_per_eu(CK_MIN_WAVES_PER_EU, CK_MAX_WAVES_PER_EU)))
 #endif
         kernel_gemm_xdlops_v2r3(const typename GridwiseGemm::Argument karg)
 {