Merge commit '57fadf6f' into gemm_layernorm_welford

example/35_splitK_gemm/CMakeLists.txt

+add_custom_target(example_splitK_gemm_xdl)
+
+add_example_executable(example_splitK_gemm_xdl_fp32 splitK_gemm_xdl_fp32.cpp)
+add_example_executable(example_splitK_gemm_xdl_fp16 splitK_gemm_xdl_fp16.cpp)
+add_example_executable(example_splitK_gemm_xdl_bfp16 splitK_gemm_xdl_bfp16.cpp)
+add_example_executable(example_splitK_gemm_xdl_int8 splitK_gemm_xdl_int8.cpp)
+
+add_dependencies(example_splitK_gemm_xdl
+                 example_splitK_gemm_xdl_fp32
+                 example_splitK_gemm_xdl_fp16
+                 example_splitK_gemm_xdl_bfp16
+                 example_splitK_gemm_xdl_int8)
+
+if(USE_BITINT_EXTENSION_INT4)
+  add_example_executable(example_splitK_gemm_xdl_int4 splitK_gemm_xdl_int4.cpp)
+  add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_int4)
+endif()

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

@@ -3,9 +3,6 @@
 
 #pragma once
 
-#include <iostream>
-#include <vector>
-
 #include "ck/tensor_operation/gpu/device/device_base.hpp"
 
 namespace ck {