Split the instances by architecture. (#1223)

* parse examples inside the add_example_executable function

* fix the example 64 cmake file

* add xdl flag to the gemm_bias_softmax_gemm_permute example

* add filtering of tests based on architecture type

* enable test_grouped_gemm for gfx9 only

* enable test_transpose only for gfx9

* only linnk test_transpose if it gets built

* split the gemm instances by architectures

* split gemm_bilinear,grouped_conv_bwd_weight instances by targets

* split instances by architecture

* split grouped_conv instances by architecture

* fix clang format

* fix the if-else logic in group_conv headers

* small fix for grouped convolution instances

* fix the grouped conv bwd weight dl instances

* fix client examples

* only enable client examples 3 and 4 on gfx9

* set the gfx9 macro

* make sure the architecture macros are set by cmake

* use separate set of xdl/wmma flags for host code

* sinmplify the main cmake file

* add conv_fwd_bf8 instance declaration

library/src/tensor_operation_instance/gpu/CMakeLists.txt

@@ -36,12 +36,27 @@ function(add_instance_library INSTANCE_NAME)
             endif()
         endforeach()
     endif()
+    # 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")
             message("removing dl instance ${source} ")
             list(REMOVE_ITEM ARGN "${source}")
         endif()
     endforeach()
+    # Do not build XDL instances if gfx9 targets are not on the target list
+    foreach(source IN LISTS ARGN)
+        if(NOT GPU_TARGETS MATCHES "gfx9" AND source MATCHES "_xdl")
+            message("removing xdl instance ${source} ")
+            list(REMOVE_ITEM ARGN "${source}")
+        endif()
+    endforeach()
+    # Do not build WMMA instances if gfx11 targets are not on the target list
+    foreach(source IN LISTS ARGN)
+        if(NOT GPU_TARGETS MATCHES "gfx11" AND source MATCHES "_wmma")
+            message("removing wmma instance ${source} ")
+            list(REMOVE_ITEM ARGN "${source}")
+        endif()
+    endforeach()
     #only continue if there are some source files left on the list
     if(ARGN)
         add_library(${INSTANCE_NAME} OBJECT ${ARGN})
@@ -124,6 +139,26 @@ FOREACH(subdir_path ${dir_list})
             message("Found only dl instances, but DL_KERNELS is not set. Skipping.")
             set(add_inst 0)
         endif()
+        if(("${cmake_instance}" MATCHES "ONLY XDL_KERNELS") AND (NOT GPU_TARGETS MATCHES "gfx9"))
+            message("Found only xdl instances, but gfx9 is not on the targets list. Skipping.")
+            set(add_inst 0)
+        endif()
+        if(("${cmake_instance}" MATCHES "ONLY WMMA_KERNELS") AND (NOT GPU_TARGETS MATCHES "gfx11"))
+            message("Found only wmma instances, but gfx11 is not on the targets list. Skipping.")
+            set(add_inst 0)
+        endif()
+        if(("${cmake_instance}" MATCHES "ONLY XDL_AND_DL_KERNELS") AND (NOT DEFINED DL_KERNELS) AND (NOT GPU_TARGETS MATCHES "gfx9"))
+            message("Found only xdl and dl instances, but gfx9 is not on the targets listand DL_KERNELS is not set. Skipping.")
+            set(add_inst 0)
+        endif()
+        if(("${cmake_instance}" MATCHES "ONLY XDL_AND_WMMA_KERNELS") AND (NOT GPU_TARGETS MATCHES "gfx11") AND (NOT GPU_TARGETS MATCHES "gfx9"))
+            message("Found only xdl and wmma instances, but gfx11 and gfx9 are not on the targets list. Skipping.")
+            set(add_inst 0)
+        endif()
+        if(("${cmake_instance}" MATCHES "XDL_DL_WMMA_KERNELS") AND (NOT GPU_TARGETS MATCHES "gfx11") AND (NOT GPU_TARGETS MATCHES "gfx9") AND (NOT DEFINED DL_KERNELS))
+            message("Found xdl, dl, and wmma instances, but none of those meet the target list. Skipping.")
+            set(add_inst 0)
+        endif()
         if((add_inst EQUAL 1))
             get_filename_component(target_dir ${subdir_path} NAME)
             add_subdirectory(${target_dir})

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(BATCHED_GEMM_INSTANCES)
 list(APPEND BATCHED_GEMM_INSTANCES device_batched_gemm_xdl_f16_f16_f16_gmk_gkn_gmn_instance.cpp
                                       device_batched_gemm_xdl_f16_f16_f16_gmk_gnk_gmn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_batched_gemm_add_relu_gemm_add_instance
     device_batched_gemm_add_relu_gemm_add_xdl_cshuffle_f16_f16_f16_f16_gmk_gnk_gno_gmo_instance.cpp
     device_batched_gemm_add_relu_gemm_add_xdl_cshuffle_f16_f16_f16_f16_gmk_gnk_gon_gmo_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_batched_gemm_bias_permute_instance
     device_batched_gemm_bias_permute_m2_n3_k1_xdl_c_shuffle_f16_f16_f16_f16_instance.cpp
 )

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_batched_gemm_gemm_instance
     device_batched_gemm_gemm_xdl_cshuffle_f16_f16_f16_f16_gmk_gnk_gno_gmo_instance.cpp
     device_batched_gemm_gemm_xdl_cshuffle_f16_f16_f16_f16_gmk_gnk_gon_gmo_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_batched_gemm_reduce_instance
     device_batched_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_gmk_gkn_gmn_instance.cpp
     device_batched_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_gmk_gnk_gmn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_batched_gemm_softmax_gemm_instance
     device_batched_gemm_softmax_gemm_xdl_cshuffle_f16_f16_f16_f16_gmk_gnk_gno_gmo_instance.cpp
 )

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(DEVICE_BATCHED_GEMM_SOFTMAX_GEMM_PERMUTE_INSTANCES)
 list(APPEND DEVICE_BATCHED_GEMM_SOFTMAX_GEMM_PERMUTE_INSTANCES
     device_batched_gemm_softmax_gemm_permute_xdl_cshuffle_f16_f16_f16_f16_gmk_gnk_gno_gmo_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(DEVICE_CONTRACTION_BILINEAR_INSTANCES)
 
 # FP32

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(DEVICE_CONTRACTION_SCALE_INSTANCES)
 
 # FP32

library/src/tensor_operation_instance/gpu/conv1d_bwd_data/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_conv1d_bwd_data_instance
    device_conv1d_bwd_data_xdl_nwc_kxc_nwk_f16_instance.cpp
    device_conv1d_bwd_data_xdl_nwc_kxc_nwk_f32_instance.cpp

library/src/tensor_operation_instance/gpu/conv2d_bwd_data/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_AND_DL_KERNELS
 set(CONV2D_BWD_DATA_INSTANCES)
 list(APPEND CONV2D_BWD_DATA_INSTANCES device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk_f32_instance.cpp
                                       device_conv2d_bwd_data_dl_nhwc_kyxc_nhwk_f32_instance.cpp

library/src/tensor_operation_instance/gpu/conv2d_fwd/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(DEVICE_CONV2D_FWD_INSTANCES)
 list(APPEND DEVICE_CONV2D_FWD_INSTANCES device_conv2d_fwd_xdl_nhwc_kyxc_nhwk_f32_instance.cpp
                                         device_conv2d_fwd_xdl_nhwc_kyxc_nhwk_f16_instance.cpp

library/src/tensor_operation_instance/gpu/conv2d_fwd_bias_relu/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_conv2d_fwd_bias_relu_instance
    device_conv2d_fwd_xdl_c_shuffle_bias_relu_nhwc_kyxc_nhwk_f16_instance.cpp
 )

library/src/tensor_operation_instance/gpu/conv2d_fwd_bias_relu_add/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_conv2d_fwd_bias_relu_add_instance
    device_conv2d_fwd_xdl_c_shuffle_bias_relu_add_nhwc_kyxc_nhwk_f16_instance.cpp
 )

library/src/tensor_operation_instance/gpu/conv3d_bwd_data/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_conv3d_bwd_data_instance
    device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f16_instance.cpp
    device_conv3d_bwd_data_xdl_ndhwc_kzyxc_ndhwk_f32_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_add_instance
    device_gemm_add_xdl_c_shuffle_f16_i8_f16_f16_mk_kn_mn_mn_instance.cpp
    device_gemm_add_xdl_c_shuffle_bf16_i8_bf16_bf16_mk_kn_mn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_add_add_fastgelu_instance
    device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_f16_f16_km_kn_mn_mn_mn_instance.cpp
    device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_f16_f16_km_nk_mn_mn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_add_fastgelu_instance
    device_gemm_add_fastgelu_xdl_c_shuffle_bf16_i8_bf16_bf16_mk_kn_mn_mn_instance.cpp
    device_gemm_add_fastgelu_xdl_c_shuffle_f16_f16_f16_f16_km_kn_mn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_add_multiply_instance
    device_gemm_add_multiply_xdl_c_shuffle_f16_f16_f16_f16_f16_km_kn_mn_mn_mn_instance.cpp
    device_gemm_add_multiply_xdl_c_shuffle_f16_f16_f16_f16_f16_km_nk_mn_mn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_add_relu_instance
    device_gemm_add_relu_xdl_c_shuffle_f16_i8_f16_f16_mk_kn_mn_mn_instance.cpp
    device_gemm_add_relu_xdl_c_shuffle_bf16_i8_bf16_bf16_mk_kn_mn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_add_relu_add_layernorm_instance
    device_gemm_add_relu_add_xdl_c_shuffle_layernorm_f16_km_kn_mn_mn_mn_instance.cpp
    device_gemm_add_relu_add_xdl_c_shuffle_layernorm_f16_km_nk_mn_mn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_add_silu_instance
    device_gemm_add_silu_xdl_c_shuffle_f16_i8_f16_f16_mk_kn_mn_mn_instance.cpp
    device_gemm_add_silu_xdl_c_shuffle_bf16_i8_bf16_bf16_mk_kn_mn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_bias_add_reduce_instance
     device_gemm_bias_add_mean_squaremean_xdl_cshuffle_f16_f16_f16_f32_f32_mk_kn_mn_instance.cpp
     device_gemm_bias_add_mean_squaremean_xdl_cshuffle_f16_f16_f16_f32_f32_mk_nk_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_AND_WMMA_KERNELS
 add_instance_library(device_gemm_bilinear_instance
    device_gemm_bilinear_xdl_c_shuffle_f16_f16_f16_f16_km_kn_mn_mn_instance.cpp
    device_gemm_bilinear_xdl_c_shuffle_f16_f16_f16_f16_km_nk_mn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_fastgelu_instance
    device_gemm_fastgelu_xdl_c_shuffle_f16_f16_f16_km_kn_mn_instance.cpp
    device_gemm_fastgelu_xdl_c_shuffle_f16_f16_f16_km_nk_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(GEMM_MULTIPLY_ADD_INSTANCES)
 list(APPEND GEMM_MULTIPLY_ADD_INSTANCES device_gemm_multiply_add_xdl_c_shuffle_f16_f16_f16_f16_f16_mk_kn_mn_mn_mn_instance.cpp
                                         device_gemm_multiply_add_xdl_c_shuffle_f16_f16_f16_f16_f16_mk_nk_mn_mn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_reduce_instance
     device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_mk_kn_mn_instance.cpp
     device_gemm_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_mk_nk_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(GEMM_SPLITK_INSTANCES)
 
 list(APPEND GEMM_SPLITK_INSTANCES 

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_gemm_streamk_instance
    # device_gemm_xdl_streamk_f32_f32_f32_mk_kn_mn_instance.cpp
    # device_gemm_xdl_streamk_f32_f32_f32_mk_nk_mn_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv1d_bwd_weight/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_AND_DL_KERNELS
 set(GROUPED_CONV1D_BWD_WEIGHT
     xdl/device_grouped_conv1d_bwd_weight_xdl_gnwc_gkxc_gnwk_f16_instance.cpp
     xdl/device_grouped_conv1d_bwd_weight_xdl_gnwc_gkxc_gnwk_f32_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv1d_fwd/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_grouped_conv1d_fwd_instance
    xdl/device_grouped_conv1d_fwd_xdl_gnwc_gkxc_gnwk_bf16_instance.cpp
    xdl/device_grouped_conv1d_fwd_xdl_gnwc_gkxc_gnwk_f16_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_data/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_AND_WMMA_KERNELS
 add_instance_library(
 	device_grouped_conv2d_bwd_data_instance
 	xdl/device_grouped_conv2d_bwd_data_xdl_gnhwc_gkyxc_gnhwk_f16_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_weight/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_AND_DL_KERNELS
 set(GROUPED_CONV2D_BWD_WEIGHT
     xdl/device_grouped_conv2d_bwd_weight_xdl_gnhwc_gkyxc_gnhwk_f16_instance.cpp
     xdl/device_grouped_conv2d_bwd_weight_xdl_gnhwc_gkyxc_gnhwk_f32_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/CMakeLists.txt

@@ -1,3 +1,4 @@
+# XDL_DL_WMMA_KERNELS
 add_instance_library(device_grouped_conv2d_fwd_instance
    #xdl
    # GNHWC, GKYXC, GNHWK

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_data/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_AND_WMMA_KERNELS
 set(GROUPED_CONV3D_BWD_DATA
    xdl/device_grouped_conv3d_bwd_data_xdl_gndhwc_gkzyxc_gndhwk_f16_instance.cpp
    xdl/device_grouped_conv3d_bwd_data_xdl_gndhwc_gkzyxc_gndhwk_bf16_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_data_bilinear/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(GROUPED_CONV3D_BWD_DATA_BILINEAR
    xdl/device_grouped_conv3d_bwd_data_xdl_bilinear_ndhwgc_gkzyxc_ndhwgk_f16_instance.cpp
    xdl/device_grouped_conv3d_bwd_data_xdl_bilinear_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_data_scale/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(GROUPED_CONV3D_BWD_DATA_BILINEAR
    xdl/device_grouped_conv3d_bwd_data_xdl_scale_ndhwgc_gkzyxc_ndhwgk_f16_instance.cpp
    xdl/device_grouped_conv3d_bwd_data_xdl_scale_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv3d_bwd_weight/CMakeLists.txt

@@ -1,3 +1,4 @@
+# XDL_DL_WMMA_KERNELS
 set(GROUPED_CONV3D_BWD_WEIGHT
     xdl/device_grouped_conv3d_bwd_weight_xdl_gndhwc_gkzyxc_gndhwk_f16_instance.cpp
     xdl/device_grouped_conv3d_bwd_weight_xdl_gndhwc_gkzyxc_gndhwk_f32_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv3d_fwd/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_AND_WMMA_KERNELS
 set(GROUPED_CONV3D_FWD
    xdl/device_grouped_conv3d_fwd_xdl_gndhwc_gkzyxc_gndhwk_bf16_instance.cpp
    xdl/device_grouped_conv3d_fwd_xdl_gndhwc_gkzyxc_gndhwk_f16_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv3d_fwd_bilinear/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(GROUPED_CONV3D_FWD_BILINEAR
    xdl/device_grouped_conv3d_fwd_xdl_bilinear_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp
    xdl/device_grouped_conv3d_fwd_xdl_bilinear_ndhwgc_gkzyxc_ndhwgk_f16_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv3d_fwd_scale/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(GROUPED_CONV3D_FWD_BILINEAR
    xdl/device_grouped_conv3d_fwd_xdl_scale_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp
    xdl/device_grouped_conv3d_fwd_xdl_scale_ndhwgc_gkzyxc_ndhwgk_f16_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv3d_fwd_scaleadd_ab/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(GROUPED_CONV3D_FWD_SCALEADD_AB
    xdl/device_grouped_conv3d_fwd_xdl_scaleadd_ab_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp
    xdl/device_grouped_conv3d_fwd_xdl_scaleadd_ab_ndhwgc_gkzyxc_ndhwgk_f16_instance.cpp

library/src/tensor_operation_instance/gpu/grouped_conv3d_fwd_scaleadd_scaleadd_relu/CMakeLists.txt

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(GROUPED_CONV3D_FWD_scaleadd_scaleadd_RELU
    xdl/device_grouped_conv3d_fwd_xdl_scaleadd_scaleadd_relu_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp
    xdl/device_grouped_conv3d_fwd_xdl_scaleadd_scaleadd_relu_ndhwgc_gkzyxc_ndhwgk_f16_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_grouped_gemm_instance
    device_grouped_gemm_xdl_f16_f16_f16_mk_kn_mn_instance.cpp
    device_grouped_gemm_xdl_f16_f16_f16_mk_nk_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_grouped_gemm_bias_instance
    device_grouped_gemm_xdl_fixed_nk_bias_f16_f16_f16_mk_kn_mn_instance.cpp
    device_grouped_gemm_xdl_fixed_nk_bias_f16_f16_f16_mk_nk_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 add_instance_library(device_grouped_gemm_fastgelu_instance
    device_grouped_gemm_fastgelu_xdl_f16_f16_f16_mk_kn_mn_instance.cpp
    device_grouped_gemm_fastgelu_xdl_f16_f16_f16_mk_nk_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_KERNELS
 set(GROUPED_GEMM_FIXED_NK_INSTANCES)
 
 list(APPEND GROUPED_GEMM_FIXED_NK_INSTANCES device_grouped_gemm_xdl_fixed_nk_f16_f16_f16_mk_kn_mn_instance.cpp

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

@@ -1,3 +1,4 @@
+# ONLY XDL_AND_DL_KERNELS
 set(CONV2D_PERLAYER_QUANT_SRC conv2d_fwd/device_conv2d_xdl_perlayer_quantization_int8_instance.cpp)
 set(CONV2D_PERCHANNEL_QUANT_SRC conv2d_fwd/device_conv2d_xdl_perchannel_quantization_int8_instance.cpp)
 set(CONV2D_BIAS_PERLAYER_QUANT_SRC conv2d_fwd/device_conv2d_xdl_bias_perlayer_quantization_int8_instance.cpp)