Add GGemmSplitK mk-kn-mn instances

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

@@ -81,6 +81,19 @@ void add_device_grouped_gemm_xdl_splitk_f16_f16_f16_mk_nk_mn_instances(
                                                   PassThrough,
                                                   PassThrough>>>& instances);
 
+void add_device_grouped_gemm_xdl_splitk_f16_f16_f16_mk_kn_mn_instances(
+    std::vector<std::unique_ptr<DeviceGroupedGemm<Row,
+                                                  Row,
+                                                  Empty_Tuple,
+                                                  Row,
+                                                  F16,
+                                                  F16,
+                                                  Empty_Tuple,
+                                                  F16,
+                                                  PassThrough,
+                                                  PassThrough,
+                                                  PassThrough>>>& instances);
+
 template <typename ALayout,
           typename BLayout,
           typename ELayout,
@@ -122,6 +135,7 @@ struct DeviceOperationInstanceFactory<ck::tensor_operation::device::DeviceGroupe
                          is_same_v<ELayout, Row>)
             {
                 add_device_grouped_gemm_xdl_f16_f16_f16_mk_kn_mn_instances(op_ptrs);
+                add_device_grouped_gemm_xdl_splitk_f16_f16_f16_mk_kn_mn_instances(op_ptrs);
             }
             else if constexpr(is_same_v<ALayout, Row> && is_same_v<BLayout, Col> &&
                               is_same_v<ELayout, Row>)

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

@@ -3,5 +3,6 @@ add_instance_library(device_grouped_gemm_instance
    device_grouped_gemm_xdl_f16_f16_f16_mk_nk_mn_instance.cpp
    device_grouped_gemm_xdl_f16_f16_f16_km_kn_mn_instance.cpp
    device_grouped_gemm_xdl_f16_f16_f16_km_nk_mn_instance.cpp
+   device_grouped_gemm_xdl_splitk_f16_f16_f16_mk_kn_mn_instance.cpp
    device_grouped_gemm_xdl_splitk_f16_f16_f16_mk_nk_mn_instance.cpp
 )