Improve external interface for GEMM and GEMM+add+add+fastgelu (#311)

* interface for GEMM and GEMM+add+add+fastgelu

* rename namespace

* instance factory

* fix build

* fix build; add GEMM client example

* clean

library/src/tensor_operation_instance/gpu/gemm_reduce/device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_km_kn_mn_instance.cpp

@@ -9,12 +9,12 @@
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_reduce_xdl_cshuffle.hpp"
 #include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16              = ck::half_t;
 using F32              = float;
@@ -74,7 +74,7 @@ void add_device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_km_kn_mn_instances(
         instances, device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_km_kn_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_reduce/device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_km_nk_mn_instance.cpp

@@ -9,12 +9,12 @@
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_reduce_xdl_cshuffle.hpp"
 #include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16              = ck::half_t;
 using F32              = float;
@@ -74,7 +74,7 @@ void add_device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_km_nk_mn_instances(
         instances, device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_km_nk_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_reduce/device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_mk_kn_mn_instance.cpp

@@ -9,12 +9,12 @@
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_reduce_xdl_cshuffle.hpp"
 #include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16              = ck::half_t;
 using F32              = float;
@@ -74,7 +74,7 @@ void add_device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_mk_kn_mn_instances(
         instances, device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_mk_kn_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_reduce/device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_mk_nk_mn_instance.cpp

@@ -9,12 +9,12 @@
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_reduce_xdl_cshuffle.hpp"
 #include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16              = ck::half_t;
 using F32              = float;
@@ -71,7 +71,7 @@ void add_device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_mk_nk_mn_instances(
         instances, device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_mk_nk_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_splitk/device_gemm_xdl_splitk_f16_f16_f16_km_kn_mn_instance.cpp

@@ -7,12 +7,13 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_xdl_splitk_c_shuffle.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -46,13 +47,15 @@ using device_gemm_xdl_splitk_f16_f16_f16_km_kn_mn_instances = std::tuple<
     >;
 
 void add_device_gemm_xdl_splitk_f16_f16_f16_km_kn_mn_instances(
-    std::vector<DeviceGemmSplitKPtr<PassThrough, PassThrough, PassThrough>>& instances)
+    std::vector<std::unique_ptr<
+        DeviceGemmSplitK<Col, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances)
 {
     add_device_operation_instances(instances,
                                    device_gemm_xdl_splitk_f16_f16_f16_km_kn_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_splitk/device_gemm_xdl_splitk_f16_f16_f16_km_nk_mn_instance.cpp

@@ -7,12 +7,13 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_xdl_splitk_c_shuffle.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -46,13 +47,15 @@ using device_gemm_xdl_splitk_f16_f16_f16_km_nk_mn_instances = std::tuple<
     >;
 
 void add_device_gemm_xdl_splitk_f16_f16_f16_km_nk_mn_instances(
-    std::vector<DeviceGemmSplitKPtr<PassThrough, PassThrough, PassThrough>>& instances)
+    std::vector<std::unique_ptr<
+        DeviceGemmSplitK<Col, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances)
 {
     add_device_operation_instances(instances,
                                    device_gemm_xdl_splitk_f16_f16_f16_km_nk_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_splitk/device_gemm_xdl_splitk_f16_f16_f16_mk_kn_mn_instance.cpp

@@ -7,12 +7,13 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_xdl_splitk_c_shuffle.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -46,13 +47,15 @@ using device_gemm_xdl_splitk_f16_f16_f16_mk_kn_mn_instances = std::tuple<
     >;
 
 void add_device_gemm_xdl_splitk_f16_f16_f16_mk_kn_mn_instances(
-    std::vector<DeviceGemmSplitKPtr<PassThrough, PassThrough, PassThrough>>& instances)
+    std::vector<std::unique_ptr<
+        DeviceGemmSplitK<Row, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances)
 {
     add_device_operation_instances(instances,
                                    device_gemm_xdl_splitk_f16_f16_f16_mk_kn_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_splitk/device_gemm_xdl_splitk_f16_f16_f16_mk_nk_mn_instance.cpp

@@ -7,12 +7,13 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_xdl_splitk_c_shuffle.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -50,50 +51,16 @@ using device_gemm_xdl_splitk_f16_f16_f16_mk_nk_mn_instances = std::tuple<
     // clang-format on
     >;
 
-// using device_gemm_xdl_splitk_f16_f16_f16_mk_nk_mn_irregular_tile_instances = std::tuple<
-//     // clang-format off
-//         //#########################|AData| BData| CData| AccData| ALayout| BLayout| CLayout| A|
-//         B|           C|          GEMM| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|
-//         ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer|
-//         ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer|
-//         BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds|    CShuffle|    CShuffle|
-//         CBlockTransferClusterLengths|  CBlockTransfer|
-//         //#########################| Type|  Type|  Type|    Type|        |        |        |
-//         Elementwise| Elementwise| Elementwise|Spacialization|  Size| Block| Block| Block|   |
-//         XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|
-//         SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|
-//         SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave|
-//         _MBlock_MXdlPerWave_MWaveMPerXdl| ScalarPerVector|
-//         //#########################|     |      |      |        |        |        |        |
-//         Operation|   Operation|   Operation|              |      |      |      |      |   |     |
-//         | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               | PerVector|
-//         PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |
-//         PerVector|   PerVector_K1|          |  PerShuffle|  PerShuffle|
-//         _NBlock_NXdlPerWave_NWaveNPerXdl|   _NWaveNPerXdl|
-//         //#########################|     |      |      |        |        |        |        | | |
-//         |              |      |      |      |      |   |     |     |     |     |                |
-//         |               |               |               |               |          | | | | | | |
-//         |            |            |                                 |                |
-//         DeviceGemmXdlSplitKCShuffle<   F16,   F16,   F16,     F32,     Row,     Col,     Row,
-//         PassThrough, PassThrough, PassThrough,   GemmDefault,   256,   128,   144,     4,  8, 16,
-//         16,    2,    9,  S<1, 4, 64, 1>,  S<0, 2, 1, 3>,  S<0, 2, 1, 3>,              3, 8, 8,
-//         true,  S<1, 4, 16, 4>,  S<0, 2, 1, 3>,  S<0, 2, 1, 3>,             3,              2, 2,
-//         true,           1,           9,                   S<1, 2, 1, 72>,               2>
-//     // clang-format on
-//     >;
-
 void add_device_gemm_xdl_splitk_f16_f16_f16_mk_nk_mn_instances(
-    std::vector<DeviceGemmSplitKPtr<PassThrough, PassThrough, PassThrough>>& instances)
+    std::vector<std::unique_ptr<
+        DeviceGemmSplitK<Row, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances)
 {
     add_device_operation_instances(instances,
                                    device_gemm_xdl_splitk_f16_f16_f16_mk_nk_mn_instances{});
-
-    // FIXME - IsSupportedArgument() is false, need to check validity
-    // add_device_operation_instances(
-    //     instances, device_gemm_xdl_splitk_f16_f16_f16_mk_nk_mn_irregular_tile_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_splitk/device_gemm_xdl_splitk_f32_f32_f32_km_kn_mn_instance.cpp

@@ -7,12 +7,13 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_xdl_splitk.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -46,13 +47,15 @@ using device_gemm_xdl_splitk_f32_f32_f32_km_kn_mn_instances = std::tuple<
     >;
 
 void add_device_gemm_xdl_splitk_f32_f32_f32_km_kn_mn_instances(
-    std::vector<DeviceGemmSplitKPtr<PassThrough, PassThrough, PassThrough>>& instances)
+    std::vector<std::unique_ptr<
+        DeviceGemmSplitK<Col, Row, Row, F32, F32, F32, PassThrough, PassThrough, PassThrough>>>&
+        instances)
 {
     add_device_operation_instances(instances,
                                    device_gemm_xdl_splitk_f32_f32_f32_km_kn_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_splitk/device_gemm_xdl_splitk_f32_f32_f32_km_nk_mn_instance.cpp

@@ -7,12 +7,13 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_xdl_splitk.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -46,13 +47,15 @@ using device_gemm_xdl_splitk_f32_f32_f32_km_nk_mn_instances = std::tuple<
     >;
 
 void add_device_gemm_xdl_splitk_f32_f32_f32_km_nk_mn_instances(
-    std::vector<DeviceGemmSplitKPtr<PassThrough, PassThrough, PassThrough>>& instances)
+    std::vector<std::unique_ptr<
+        DeviceGemmSplitK<Col, Col, Row, F32, F32, F32, PassThrough, PassThrough, PassThrough>>>&
+        instances)
 {
     add_device_operation_instances(instances,
                                    device_gemm_xdl_splitk_f32_f32_f32_km_nk_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_splitk/device_gemm_xdl_splitk_f32_f32_f32_mk_kn_mn_instance.cpp

@@ -7,12 +7,13 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_xdl_splitk.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -51,13 +52,15 @@ using device_gemm_xdl_splitk_f32_f32_f32_mk_kn_mn_instances = std::tuple<
     >;
 
 void add_device_gemm_xdl_splitk_f32_f32_f32_mk_kn_mn_instances(
-    std::vector<DeviceGemmSplitKPtr<PassThrough, PassThrough, PassThrough>>& instances)
+    std::vector<std::unique_ptr<
+        DeviceGemmSplitK<Row, Row, Row, F32, F32, F32, PassThrough, PassThrough, PassThrough>>>&
+        instances)
 {
     add_device_operation_instances(instances,
                                    device_gemm_xdl_splitk_f32_f32_f32_mk_kn_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/gemm_splitk/device_gemm_xdl_splitk_f32_f32_f32_mk_nk_mn_instance.cpp

@@ -7,12 +7,13 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_gemm_xdl_splitk.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -51,13 +52,15 @@ using device_gemm_xdl_splitk_f32_f32_f32_mk_nk_mn_instances = std::tuple<
     >;
 
 void add_device_gemm_xdl_splitk_f32_f32_f32_mk_nk_mn_instances(
-    std::vector<DeviceGemmSplitKPtr<PassThrough, PassThrough, PassThrough>>& instances)
+    std::vector<std::unique_ptr<
+        DeviceGemmSplitK<Row, Col, Row, F32, F32, F32, PassThrough, PassThrough, PassThrough>>>&
+        instances)
 {
     add_device_operation_instances(instances,
                                    device_gemm_xdl_splitk_f32_f32_f32_mk_nk_mn_instances{});
 }
 
-} // namespace device_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/grouped_gemm/device_grouped_gemm_xdl_f16_f16_f16_km_kn_mn_instance.cpp

@@ -7,12 +7,12 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_grouped_gemm_xdl.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_grouped_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -52,7 +52,7 @@ void add_device_grouped_gemm_xdl_f16_f16_f16_km_kn_mn_instances(
                                    device_grouped_gemm_xdl_f16_f16_f16_km_kn_mn_instances{});
 }
 
-} // namespace device_grouped_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/grouped_gemm/device_grouped_gemm_xdl_f16_f16_f16_km_nk_mn_instance.cpp

@@ -7,12 +7,12 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_grouped_gemm_xdl.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_grouped_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -52,7 +52,7 @@ void add_device_grouped_gemm_xdl_f16_f16_f16_km_nk_mn_instances(
                                    device_grouped_gemm_xdl_f16_f16_f16_km_nk_mn_instances{});
 }
 
-} // namespace device_grouped_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/grouped_gemm/device_grouped_gemm_xdl_f16_f16_f16_mk_kn_mn_instance.cpp

@@ -7,12 +7,12 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_grouped_gemm_xdl.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_grouped_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -61,7 +61,7 @@ void add_device_grouped_gemm_xdl_f16_f16_f16_mk_kn_mn_instances(
                                    device_grouped_gemm_xdl_f16_f16_f16_mk_kn_mn_instances{});
 }
 
-} // namespace device_grouped_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/grouped_gemm/device_grouped_gemm_xdl_f16_f16_f16_mk_nk_mn_instance.cpp

@@ -7,12 +7,12 @@
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/device_grouped_gemm_xdl.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_grouped_gemm_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -72,7 +72,7 @@ void add_device_grouped_gemm_xdl_f16_f16_f16_mk_nk_mn_instances(
         instances, device_grouped_gemm_xdl_f16_f16_f16_mk_nk_mn_irregular_tile_instances{});
 }
 
-} // namespace device_grouped_gemm_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/normalization/device_softmax_f16_f16_instance.cpp

@@ -2,14 +2,15 @@
 // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "ck/ck.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 #include "ck/tensor_operation/gpu/device/device_softmax.hpp"
 #include "ck/utility/data_type.hpp"
 
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_normalization_instance {
+namespace instance {
 
 using F16 = ck::half_t;
 using F32 = float;
@@ -43,7 +44,7 @@ void add_device_softmax_f16_f16_rank4_instances(std::vector<DeviceNormalizationP
     add_device_operation_instances(instances, device_softmax_f16_f16_instances<4, 3>{});
 }
 
-} // namespace device_normalization_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/normalization/device_softmax_f32_f32_instance.cpp

@@ -2,14 +2,14 @@
 // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "ck/ck.hpp"
-#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
 #include "ck/tensor_operation/gpu/device/device_softmax.hpp"
 #include "ck/utility/data_type.hpp"
 
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_normalization_instance {
+namespace instance {
 
 using F32 = float;
 
@@ -42,7 +42,7 @@ void add_device_softmax_f32_f32_rank4_instances(std::vector<DeviceNormalizationP
     add_device_operation_instances(instances, device_softmax_f32_f32_instances<4, 3>{});
 }
 
-} // namespace device_normalization_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck

library/src/tensor_operation_instance/gpu/reduce/device_reduce_instance_blockwise_b16_f32_b16.cpp

@@ -6,7 +6,7 @@
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_reduce_instance {
+namespace instance {
 
 // clang-format off
 // InDataType | AccDataType | OutDataType | ReduceOpId | NanPropaOpt | IndicesOpt | Rank | NumReduceDim
@@ -49,7 +49,7 @@ ADD_BLOCKWISE_INST_BY_ID(bhalf_t, float, bhalf_t, 4, 0, 1, 4, 1);
 ADD_BLOCKWISE_INST_BY_ID(bhalf_t, float, bhalf_t, 4, 0, 1, 2, 1);
 // clang-format on
 
-} // namespace device_reduce_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation
 

library/src/tensor_operation_instance/gpu/reduce/device_reduce_instance_blockwise_f16_f16_f16.cpp

@@ -6,7 +6,7 @@
 namespace ck {
 namespace tensor_operation {
 namespace device {
-namespace device_reduce_instance {
+namespace instance {
 
 // clang-format off
 // InDataType | AccDataType | OutDataType | ReduceOpId | NanPropaOpt | IndicesOpt | Rank | NumReduceDim
@@ -36,7 +36,7 @@ ADD_BLOCKWISE_INST_BY_ID(half_t, half_t, half_t, 4, 0, 1, 4, 1);
 ADD_BLOCKWISE_INST_BY_ID(half_t, half_t, half_t, 4, 0, 1, 2, 1);
 // clang-format on
 
-} // namespace device_reduce_instance
+} // namespace instance
 } // namespace device
 } // namespace tensor_operation