Add instance for layout=NT

library/include/ck/library/tensor_operation_instance/gpu/gemm.hpp

@@ -309,7 +309,7 @@ struct DeviceOperationInstanceFactory<
             else if constexpr(is_same_v<ALayout, Col> && is_same_v<BLayout, Row> &&
                               is_same_v<CLayout, Row>)
             {
-                // add_device_gemm_xdl_f16_f16_f16_km_kn_mn_instances(op_ptrs);
+                add_device_gemm_xdl_f16_f16_f16_km_kn_mn_instances(op_ptrs);
                 // add_device_gemm_dl_f16_f16_f16_km_kn_mn_instances(op_ptrs);
                 // add_device_gemm_xdl_c_shuffle_f16_f16_f16_km_kn_mn_instances(op_ptrs);
             }

library/src/tensor_operation_instance/gpu/gemm/CMakeLists.txt

@@ -9,7 +9,7 @@ add_instance_library(device_gemm_instance
    # device_gemm_xdl_f32_f32_f32_km_nk_mn_instance.cpp
    device_gemm_xdl_f16_f16_f16_mk_kn_mn_instance.cpp
    device_gemm_xdl_f16_f16_f16_mk_nk_mn_instance.cpp
-   # device_gemm_xdl_f16_f16_f16_km_kn_mn_instance.cpp
+   device_gemm_xdl_f16_f16_f16_km_kn_mn_instance.cpp
    device_gemm_xdl_f16_f16_f16_km_nk_mn_instance.cpp
    # device_gemm_xdl_c_shuffle_i8_i8_i8_mk_kn_mn_instance.cpp
    # device_gemm_xdl_c_shuffle_i8_i8_i8_mk_nk_mn_instance.cpp
@@ -44,6 +44,10 @@ add_instance_library(device_gemm_instance
 
 set(ENABLE_IGLP_OPT ON)
 if(ENABLE_IGLP_OPT)
+# layout=NT
+set_source_files_properties(device_gemm_xdl_f16_f16_f16_km_kn_mn_instance.cpp PROPERTIES
+  COMPILE_OPTIONS ";--save-temps;-Wno-gnu-line-marker;"
+  COMPILE_DEFINITIONS ";USE_IGLP_OPT;USE_WAVES_PER_EU;")
 # layout=NN
 set_source_files_properties(device_gemm_xdl_f16_f16_f16_km_nk_mn_instance.cpp PROPERTIES
   COMPILE_OPTIONS ";--save-temps;-Wno-gnu-line-marker;"