Remove conditional cleaning c_thread buff.

include/ck/tensor_operation/gpu/grid/gridwise_gemm_multiple_d_xdl_splitk_cshuffle_v2.hpp

@@ -818,8 +818,6 @@ class GridwiseGemmMultipleD_xdl_splitk_cshuffle_v2
         const bool has_k_block_main_loop =
             gridwise_gemm_pipeline.CalculateHasMainLoop(num_k_block_main_loop);
 
-        bool clear_c_thread_buf = true;
-
         auto blockwise_gemm = BlockwiseGemmXdlops_k0mk1_k0nk1_m0n0m1n1m2m3m4n2_Selector<
             BlockSize,
             ComputeType,
@@ -850,8 +848,7 @@ class GridwiseGemmMultipleD_xdl_splitk_cshuffle_v2
                                                       b_block_slice_copy_step,
                                                       blockwise_gemm,
                                                       c_thread_buf,
-                                                      num_k_block_main_loop,
-                                                      clear_c_thread_buf);
+                                                      num_k_block_main_loop);
         }
         else
         {
@@ -869,8 +866,7 @@ class GridwiseGemmMultipleD_xdl_splitk_cshuffle_v2
                                                        b_block_slice_copy_step,
                                                        blockwise_gemm,
                                                        c_thread_buf,
-                                                       num_k_block_main_loop,
-                                                       clear_c_thread_buf);
+                                                       num_k_block_main_loop);
         }
     }
 

include/ck/tensor_operation/gpu/grid/gridwise_gemm_pipeline_v1.hpp

@@ -54,8 +54,7 @@ struct GridwiseGemmPipeline_v1<1, true, true>
                                const BBlockTransferStep& b_block_copy_step,
                                const BlockwiseGemm& blockwise_gemm,
                                CThreadBuffer& c_thread_buf,
-                               index_t num_loop,
-                               bool clear_c_thread_buf = true)
+                               index_t num_loop)
     {
         // preload data into LDS
         a_blockwise_copy.RunRead(a_grid_desc, a_grid_buf);
@@ -65,7 +64,6 @@ struct GridwiseGemmPipeline_v1<1, true, true>
         b_blockwise_copy.MoveSrcSliceWindow(b_grid_desc, b_block_copy_step);
 
         // Initialize C
-        if(clear_c_thread_buf)
         c_thread_buf.Clear();
 
         a_blockwise_copy.RunWrite(a_block_desc, a_block_buf);
@@ -154,8 +152,7 @@ struct GridwiseGemmPipeline_v1<2, true, true>
                                const BBlockTransferStep& b_block_copy_step,
                                const BlockwiseGemm& blockwise_gemm,
                                CThreadBuffer& c_thread_buf,
-                               index_t num_loop,
-                               bool clear_c_thread_buf = true)
+                               index_t num_loop)
     {
         // preload data into LDS
         {
@@ -173,7 +170,6 @@ struct GridwiseGemmPipeline_v1<2, true, true>
         }
 
         // Initialize C
-        if(clear_c_thread_buf)
         c_thread_buf.Clear();
 
         // main body
@@ -699,8 +695,7 @@ struct GridwiseGemmPipelineInterwave_v1<1>
                                const BBlockTransferStep& b_block_copy_step,
                                const BlockwiseGemm& blockwise_gemm,
                                CThreadBuffer& c_thread_buf,
-                               index_t num_loop,
-                               bool clear_c_thread_buf = true)
+                               index_t num_loop)
     {
         // preload data into LDS
         a_blockwise_copy.RunRead(a_grid_desc, a_grid_buf);
@@ -710,7 +705,6 @@ struct GridwiseGemmPipelineInterwave_v1<1>
         b_blockwise_copy.MoveSrcSliceWindow(b_grid_desc, b_block_copy_step);
 
         // Initialize C
-        if(clear_c_thread_buf)
         c_thread_buf.Clear();
 
         a_blockwise_copy.RunWrite(a_block_desc, a_block_buf);

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

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 
@@ -49,8 +49,7 @@ struct GridwiseGemmPipeline_v2
                                const BBlockTransferStep& b_block_copy_step,
                                const BlockwiseGemm& blockwise_gemm,
                                CThreadBuffer& c_thread_buf,
-                               index_t num_loop,
-                               bool clear_c_thread_buf = true)
+                               index_t num_loop)
     {
         // global read 0
         a_blockwise_copy.RunRead(a_grid_desc, a_grid_buf);
@@ -61,7 +60,6 @@ struct GridwiseGemmPipeline_v2
         b_blockwise_copy.MoveSrcSliceWindow(b_grid_desc, b_block_copy_step);
 
         // Initialize C
-        if(clear_c_thread_buf)
         c_thread_buf.Clear();
 
         // LDS write 0