Merge branch 'develop' into codegen_hiprtc

include/ck_tile/ops/topk.hpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 

include/ck_tile/ops/topk_softmax.hpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 

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

@@ -15,6 +15,9 @@
 #ifdef DL_KERNELS
 #include "gemm_dl.inc"
 #endif
+#ifdef DPP_KERNELS
+#include "gemm_dpp.inc"
+#endif
 #ifdef CK_USE_WMMA
 #include "gemm_wmma.inc"
 #endif
@@ -92,32 +95,24 @@ struct DeviceOperationInstanceFactory<
             {
                 add_device_gemm_dl_f16_f16_f16_mk_kn_mn_instances(op_ptrs);
                 add_device_gemm_dl_f16_f16_f16_mk_kn_mn_irregular_instances(op_ptrs);
-                add_device_gemm_dpp_f16_f16_f16_mk_kn_mn_instances(op_ptrs);
-                add_device_gemm_dpp_f16_f16_f16_mk_kn_mn_irregular_instances(op_ptrs);
             }
             else if constexpr(is_same_v<ALayout, Row> && is_same_v<BLayout, Col> &&
                               is_same_v<CLayout, Row>)
             {
                 add_device_gemm_dl_f16_f16_f16_mk_nk_mn_instances(op_ptrs);
                 add_device_gemm_dl_f16_f16_f16_mk_nk_mn_irregular_instances(op_ptrs);
-                add_device_gemm_dpp_f16_f16_f16_mk_nk_mn_instances(op_ptrs);
-                add_device_gemm_dpp_f16_f16_f16_mk_nk_mn_irregular_instances(op_ptrs);
             }
             else if constexpr(is_same_v<ALayout, Col> && is_same_v<BLayout, Row> &&
                               is_same_v<CLayout, Row>)
             {
                 add_device_gemm_dl_f16_f16_f16_km_kn_mn_instances(op_ptrs);
                 add_device_gemm_dl_f16_f16_f16_km_kn_mn_irregular_instances(op_ptrs);
-                add_device_gemm_dpp_f16_f16_f16_km_kn_mn_instances(op_ptrs);
-                add_device_gemm_dpp_f16_f16_f16_km_kn_mn_irregular_instances(op_ptrs);
             }
             else if constexpr(is_same_v<ALayout, Col> && is_same_v<BLayout, Col> &&
                               is_same_v<CLayout, Row>)
             {
                 add_device_gemm_dl_f16_f16_f16_km_nk_mn_instances(op_ptrs);
                 add_device_gemm_dl_f16_f16_f16_km_nk_mn_irregular_instances(op_ptrs);
-                add_device_gemm_dpp_f16_f16_f16_km_nk_mn_instances(op_ptrs);
-                add_device_gemm_dpp_f16_f16_f16_km_nk_mn_irregular_instances(op_ptrs);
             }
         }
 #endif
@@ -153,6 +148,39 @@ struct DeviceOperationInstanceFactory<
 #endif
 #endif // DL_KERNELS
 
+#ifdef DPP_KERNELS
+#ifdef CK_ENABLE_FP16
+        if constexpr(is_same_v<ADataType, half_t> && is_same_v<BDataType, half_t> &&
+                     is_same_v<CDataType, half_t>)
+        {
+            if constexpr(is_same_v<ALayout, Row> && is_same_v<BLayout, Row> &&
+                         is_same_v<CLayout, Row>)
+            {
+                add_device_gemm_dpp_f16_f16_f16_mk_kn_mn_instances(op_ptrs);
+                add_device_gemm_dpp_f16_f16_f16_mk_kn_mn_irregular_instances(op_ptrs);
+            }
+            else if constexpr(is_same_v<ALayout, Row> && is_same_v<BLayout, Col> &&
+                              is_same_v<CLayout, Row>)
+            {
+                add_device_gemm_dpp_f16_f16_f16_mk_nk_mn_instances(op_ptrs);
+                add_device_gemm_dpp_f16_f16_f16_mk_nk_mn_irregular_instances(op_ptrs);
+            }
+            else if constexpr(is_same_v<ALayout, Col> && is_same_v<BLayout, Row> &&
+                              is_same_v<CLayout, Row>)
+            {
+                add_device_gemm_dpp_f16_f16_f16_km_kn_mn_instances(op_ptrs);
+                add_device_gemm_dpp_f16_f16_f16_km_kn_mn_irregular_instances(op_ptrs);
+            }
+            else if constexpr(is_same_v<ALayout, Col> && is_same_v<BLayout, Col> &&
+                              is_same_v<CLayout, Row>)
+            {
+                add_device_gemm_dpp_f16_f16_f16_km_nk_mn_instances(op_ptrs);
+                add_device_gemm_dpp_f16_f16_f16_km_nk_mn_irregular_instances(op_ptrs);
+            }
+        }
+#endif
+#endif // DPP_KERNELS
+
 #ifdef CK_USE_WMMA
 #ifdef CK_ENABLE_FP16
         if constexpr(is_same_v<ADataType, half_t> && is_same_v<BDataType, half_t> &&

library/include/ck/library/tensor_operation_instance/gpu/gemm_dl.inc

@@ -28,16 +28,6 @@ void add_device_gemm_dl_f16_f16_f16_km_kn_mn_irregular_instances(
         DeviceGemm<Col, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
         instances);
 
-void add_device_gemm_dpp_f16_f16_f16_km_kn_mn_instances(
-    std::vector<std::unique_ptr<
-        DeviceGemm<Col, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
-        instances);
-
-void add_device_gemm_dpp_f16_f16_f16_km_kn_mn_irregular_instances(
-    std::vector<std::unique_ptr<
-        DeviceGemm<Col, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
-        instances);
-
 void add_device_gemm_dl_f16_f16_f16_km_nk_mn_instances(
     std::vector<std::unique_ptr<
         DeviceGemm<Col, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
@@ -48,16 +38,6 @@ void add_device_gemm_dl_f16_f16_f16_km_nk_mn_irregular_instances(
         DeviceGemm<Col, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
         instances);
 
-void add_device_gemm_dpp_f16_f16_f16_km_nk_mn_instances(
-    std::vector<std::unique_ptr<
-        DeviceGemm<Col, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
-        instances);
-
-void add_device_gemm_dpp_f16_f16_f16_km_nk_mn_irregular_instances(
-    std::vector<std::unique_ptr<
-        DeviceGemm<Col, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
-        instances);
-
 void add_device_gemm_dl_f16_f16_f16_mk_kn_mn_instances(
     std::vector<std::unique_ptr<
         DeviceGemm<Row, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
@@ -68,16 +48,6 @@ void add_device_gemm_dl_f16_f16_f16_mk_kn_mn_irregular_instances(
         DeviceGemm<Row, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
         instances);
 
-void add_device_gemm_dpp_f16_f16_f16_mk_kn_mn_instances(
-    std::vector<std::unique_ptr<
-        DeviceGemm<Row, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
-        instances);
-
-void add_device_gemm_dpp_f16_f16_f16_mk_kn_mn_irregular_instances(
-    std::vector<std::unique_ptr<
-        DeviceGemm<Row, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
-        instances);
-
 void add_device_gemm_dl_f16_f16_f16_mk_nk_mn_instances(
     std::vector<std::unique_ptr<
         DeviceGemm<Row, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&

library/include/ck/library/tensor_operation_instance/gpu/gemm_dpp.inc

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#pragma once
+
+#include <memory>
+#include <vector>
+#include "ck/ck.hpp"
+#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
+#include "ck/tensor_operation/gpu/device/device_gemm.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+
+#include "ck/library/tensor_operation_instance/device_operation_instance_factory.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+#if defined(CK_ENABLE_FP16)
+void add_device_gemm_dpp_f16_f16_f16_km_kn_mn_instances(
+    std::vector<std::unique_ptr<
+        DeviceGemm<Col, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances);
+
+void add_device_gemm_dpp_f16_f16_f16_km_kn_mn_irregular_instances(
+    std::vector<std::unique_ptr<
+        DeviceGemm<Col, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances);
+
+void add_device_gemm_dpp_f16_f16_f16_km_nk_mn_instances(
+    std::vector<std::unique_ptr<
+        DeviceGemm<Col, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances);
+
+void add_device_gemm_dpp_f16_f16_f16_km_nk_mn_irregular_instances(
+    std::vector<std::unique_ptr<
+        DeviceGemm<Col, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances);
+
+void add_device_gemm_dpp_f16_f16_f16_mk_kn_mn_instances(
+    std::vector<std::unique_ptr<
+        DeviceGemm<Row, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances);
+
+void add_device_gemm_dpp_f16_f16_f16_mk_kn_mn_irregular_instances(
+    std::vector<std::unique_ptr<
+        DeviceGemm<Row, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances);
+
+void add_device_gemm_dpp_f16_f16_f16_mk_nk_mn_instances(
+    std::vector<std::unique_ptr<
+        DeviceGemm<Row, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances);
+
+void add_device_gemm_dpp_f16_f16_f16_mk_nk_mn_irregular_instances(
+    std::vector<std::unique_ptr<
+        DeviceGemm<Row, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
+        instances);
+#endif
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/CMakeLists.txt

@@ -39,6 +39,13 @@ function(add_instance_library INSTANCE_NAME)
 
     set(INST_TARGETS ${SUPPORTED_GPU_TARGETS})
 
+    # Do not build DPP instances if DPP_KERNELS macro is not set
+    foreach(source IN LISTS ARGN)
+        if(NOT DEFINED DPP_KERNELS AND source MATCHES "_dpp")
+            message("removing dpp instance ${source} ")
+            list(REMOVE_ITEM ARGN "${source}")
+        endif()
+    endforeach()
     # Do not build DL instances if DL_KERNELS macro is not set
     foreach(source IN LISTS ARGN)
         if(NOT DEFINED DL_KERNELS AND source MATCHES "_dl")

test/CMakeLists.txt

@@ -43,6 +43,12 @@ function(add_test_executable TEST_NAME)
 
     set(TEST_TARGETS ${SUPPORTED_GPU_TARGETS})
 
+    foreach(source IN LISTS ARGN)
+        if(NOT DEFINED DPP_KERNELS AND source MATCHES "_dpp")
+            message("removing dpp test ${source} ")
+            list(REMOVE_ITEM ARGN "${source}")
+        endif()
+    endforeach()
     foreach(source IN LISTS ARGN)
         if(NOT DEFINED DL_KERNELS AND source MATCHES "_dl")
             message("removing dl test ${source} ")