Refactoring cmake files to build data types separately. (#932)

* refactor cmake files for the tests

* refactor cmake files for examples

* fix cmake for gemm example

* fix the cmake file for all examples

* add splitting by data types in gemm_splitk instance header

* rename test to reflect only dl instances are used

* clean up CI workspace, update cmake for instances

* change the jenkinsfile syntax

* build all instances except DL on gfx11

* move workspace cleanup after stages

* clean up workspace after every stage

* isolate data types in grouped_conv_fwd header

* isolate dl instances for grouped_conv2d_fwd

* fix syntax

* fix cmake and batchnorm instances

* fix typo

* fix reduction instances

* fix grouped_conv headers

* fix syntax

* replace parsing logic for instances, replace bfp16 with bf16

* fix the client examples build

* clean up DTYPES from instances cmake files

* update the parsing logic in cmake files

* make an exception for reduction kernels

* update few remaining cmake files to handle DTYPES

* fix syntax

* fix cmake conflicts

* replace f8 with fp8 test name

* resolve conflicts for dpp instances

example/26_contraction/CMakeLists.txt

-if(DTYPES MATCHES "fp32" OR NOT DEFINED DTYPES)
-    add_example_executable(example_contraction_bilinear_xdl_fp32 contraction_bilinear_xdl_fp32.cpp)
-    add_example_executable(example_contraction_scale_xdl_fp32 contraction_scale_xdl_fp32.cpp)
-endif()
-if(DTYPES MATCHES "fp64" OR NOT DEFINED DTYPES)
-    add_example_executable(example_contraction_bilinear_xdl_fp64 contraction_bilinear_xdl_fp64.cpp)
-    add_example_executable(example_contraction_scale_xdl_fp64 contraction_scale_xdl_fp64.cpp)
-endif()
+add_example_executable(example_contraction_bilinear_xdl_fp32 contraction_bilinear_xdl_fp32.cpp)
+add_example_executable(example_contraction_scale_xdl_fp32 contraction_scale_xdl_fp32.cpp)
+add_example_executable(example_contraction_bilinear_xdl_fp64 contraction_bilinear_xdl_fp64.cpp)
+add_example_executable(example_contraction_scale_xdl_fp64 contraction_scale_xdl_fp64.cpp)

example/27_layernorm/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_layernorm_fp16 layernorm_fp16.cpp)
-    add_example_executable(example_layernorm_splitk_fp16 layernorm_splitk_fp16.cpp)
-endif()
+add_example_executable(example_layernorm_fp16 layernorm_fp16.cpp)
+add_example_executable(example_layernorm_splitk_fp16 layernorm_splitk_fp16.cpp)

example/28_grouped_gemm_bias_e_permute/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_grouped_gemm_bias_e_permute_xdl_fp16 grouped_gemm_bias_e_permute_xdl_fp16.cpp)
-endif()
+add_example_executable(example_grouped_gemm_bias_e_permute_xdl_fp16 grouped_gemm_bias_e_permute_xdl_fp16.cpp)

example/29_batched_gemm_bias_e_permute/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_batched_gemm_bias_e_permute_xdl_fp16 batched_gemm_bias_e_permute_xdl_fp16.cpp)
+add_example_executable(example_batched_gemm_bias_e_permute_xdl_fp16 batched_gemm_bias_e_permute_xdl_fp16.cpp)
 
-    if(GPU_TARGETS MATCHES "gfx1100" OR GPU_TARGETS MATCHES "gfx1101" OR GPU_TARGETS MATCHES "gfx1102")
-        add_example_executable(example_batched_gemm_bias_e_permute_wmma_fp16 batched_gemm_bias_e_permute_wmma_fp16.cpp)
-    endif()
+if(GPU_TARGETS MATCHES "gfx1100" OR GPU_TARGETS MATCHES "gfx1101" OR GPU_TARGETS MATCHES "gfx1102")
+    add_example_executable(example_batched_gemm_bias_e_permute_wmma_fp16 batched_gemm_bias_e_permute_wmma_fp16.cpp)
 endif()

example/30_grouped_conv_fwd_multiple_d/CMakeLists.txt

@@ -5,27 +5,31 @@ set(target 0)
 foreach(gpu IN LISTS GPU_TARGETS)
  if(gpu IN_LIST gpu_list1 AND target EQUAL 0)
    add_custom_target(example_grouped_conv_fwd_multiple_d)
-   if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_fwd_bias_relu_add_xdl_fp16 grouped_conv_fwd_bias_relu_add_xdl_fp16.cpp)
+   add_example_executable(example_grouped_conv_fwd_bias_relu_add_xdl_fp16 grouped_conv_fwd_bias_relu_add_xdl_fp16.cpp)
+   if(result EQUAL 0)
       add_dependencies(example_grouped_conv_fwd_multiple_d example_grouped_conv_fwd_bias_relu_add_xdl_fp16)
-      add_example_executable(example_grouped_conv_fwd_xdl_fp16 grouped_conv_fwd_xdl_fp16.cpp)
+   endif()
+   add_example_executable(example_grouped_conv_fwd_xdl_fp16 grouped_conv_fwd_xdl_fp16.cpp)
+   if(result EQUAL 0)
       add_dependencies(example_grouped_conv_fwd_multiple_d example_grouped_conv_fwd_xdl_fp16)
    endif()
-   if(DTYPES MATCHES "fp32" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_fwd_bias_relu_add_xdl_fp32 grouped_conv_fwd_bias_relu_add_xdl_fp32.cpp)
+   add_example_executable(example_grouped_conv_fwd_bias_relu_add_xdl_fp32 grouped_conv_fwd_bias_relu_add_xdl_fp32.cpp)
+   if(result EQUAL 0)
       add_dependencies(example_grouped_conv_fwd_multiple_d example_grouped_conv_fwd_bias_relu_add_xdl_fp32)
    endif()
-   if(DTYPES MATCHES "bf16" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_fwd_bias_relu_add_xdl_bf16 grouped_conv_fwd_bias_relu_add_xdl_bf16.cpp)
+   add_example_executable(example_grouped_conv_fwd_bias_relu_add_xdl_bf16 grouped_conv_fwd_bias_relu_add_xdl_bf16.cpp)
+   if(result EQUAL 0)
       add_dependencies(example_grouped_conv_fwd_multiple_d example_grouped_conv_fwd_bias_relu_add_xdl_bf16)
    endif()
-   if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_fwd_bias_relu_add_xdl_int8 grouped_conv_fwd_bias_relu_add_xdl_int8.cpp)
+   add_example_executable(example_grouped_conv_fwd_bias_relu_add_xdl_int8 grouped_conv_fwd_bias_relu_add_xdl_int8.cpp)
+   if(result EQUAL 0)
       add_dependencies(example_grouped_conv_fwd_multiple_d example_grouped_conv_fwd_bias_relu_add_xdl_int8)
    endif()
    if(USE_BITINT_EXTENSION_INT4)
       add_example_executable(example_grouped_conv_fwd_bias_relu_add_xdl_int4 grouped_conv_fwd_bias_relu_add_xdl_int4.cpp)
-      add_dependencies(example_grouped_conv_fwd_multiple_d example_grouped_conv_fwd_bias_relu_add_xdl_int4)
+      if(result EQUAL 0)
+         add_dependencies(example_grouped_conv_fwd_multiple_d example_grouped_conv_fwd_bias_relu_add_xdl_int4)
+      endif()
    endif() # USE_BITINT_EXTENSION_INT4
 
    set(target 1)
@@ -35,12 +39,8 @@ endforeach()
 set(target 0)
 foreach(gpu IN LISTS GPU_TARGETS)
  if(gpu IN_LIST gpu_list2 AND target EQUAL 0)
-   if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_fwd_bias_relu_add_wmma_fp16 grouped_conv_fwd_bias_relu_add_wmma_fp16.cpp)
-   endif()
-   if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_fwd_bias_relu_add_wmma_int8 grouped_conv_fwd_bias_relu_add_wmma_int8.cpp)
-   endif()
+   add_example_executable(example_grouped_conv_fwd_bias_relu_add_wmma_fp16 grouped_conv_fwd_bias_relu_add_wmma_fp16.cpp)
+   add_example_executable(example_grouped_conv_fwd_bias_relu_add_wmma_int8 grouped_conv_fwd_bias_relu_add_wmma_int8.cpp)
    set(target 1)
  endif()
 endforeach()

example/31_batched_gemm_gemm/CMakeLists.txt

 list(APPEND gpu_list1 gfx908 gfx90a gfx940 gfx941 gfx942)
-list(APPEND gpu_list2 gfx908 gfx90a)
+
 set(target 0)
 foreach(gpu IN LISTS GPU_TARGETS)
  if(gpu IN_LIST gpu_list1 AND target EQUAL 0)
-   if(DTYPES MATCHES "fp32" OR NOT DEFINED DTYPES)
-      add_example_executable(example_batched_gemm_gemm_xdl_fp32 batched_gemm_gemm_xdl_fp32.cpp)
-   endif()
-   if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-      add_example_executable(example_batched_gemm_gemm_xdl_fp16 batched_gemm_gemm_xdl_fp16.cpp)
-   endif()
-   if(DTYPES MATCHES "bf16" OR NOT DEFINED DTYPES)
-      add_example_executable(example_batched_gemm_gemm_xdl_bf16 batched_gemm_gemm_xdl_bf16.cpp)
-   endif()
+   add_example_executable(example_batched_gemm_gemm_xdl_fp32 batched_gemm_gemm_xdl_fp32.cpp)
+   add_example_executable(example_batched_gemm_gemm_xdl_fp16 batched_gemm_gemm_xdl_fp16.cpp)
+   add_example_executable(example_batched_gemm_gemm_xdl_bf16 batched_gemm_gemm_xdl_bf16.cpp)
    if(USE_BITINT_EXTENSION_INT4)
       add_example_executable(example_batched_gemm_gemm_xdl_int4 batched_gemm_gemm_xdl_int4.cpp)
    endif(USE_BITINT_EXTENSION_INT4)
@@ -20,7 +14,5 @@ foreach(gpu IN LISTS GPU_TARGETS)
 endforeach()
 
 if(NOT GPU_TARGETS MATCHES "gfx94" AND NOT GPU_TARGETS MATCHES "gfx1")
-   if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
-      add_example_executable(example_batched_gemm_gemm_xdl_int8 batched_gemm_gemm_xdl_int8.cpp)
-   endif()
+   add_example_executable(example_batched_gemm_gemm_xdl_int8 batched_gemm_gemm_xdl_int8.cpp)
 endif()

example/32_batched_gemm_scale_softmax_gemm/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_batched_gemm_scale_softmax_gemm_xdl_fp16 batched_gemm_scale_softmax_gemm_xdl_fp16.cpp)
-    add_example_executable(example_batched_gemm_scale_softmax_gemm_permute_xdl_fp16 batched_gemm_scale_softmax_gemm_permute_xdl_fp16.cpp)
-    add_example_executable(example_grouped_gemm_scale_softmax_gemm_permute_xdl_fp16 grouped_gemm_scale_softmax_gemm_permute_xdl_fp16.cpp)
-    add_example_executable(example_batched_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16 batched_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16.cpp)
-    add_example_executable(example_grouped_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16 grouped_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16.cpp)
-endif()
-if(DTYPES MATCHES "bf16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_batched_gemm_scale_softmax_gemm_xdl_bf16 batched_gemm_scale_softmax_gemm_xdl_bf16.cpp)
-    add_example_executable(example_batched_gemm_scale_softmax_gemm_permute_xdl_bf16 batched_gemm_scale_softmax_gemm_permute_xdl_bf16.cpp)
-endif()
-
 add_custom_target(example_gemm_scale_softmax_gemm)
-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
+
+add_example_executable(example_batched_gemm_scale_softmax_gemm_xdl_fp16 batched_gemm_scale_softmax_gemm_xdl_fp16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_gemm_scale_softmax_gemm example_batched_gemm_scale_softmax_gemm_xdl_fp16)
+endif()
+add_example_executable(example_batched_gemm_scale_softmax_gemm_permute_xdl_fp16 batched_gemm_scale_softmax_gemm_permute_xdl_fp16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_gemm_scale_softmax_gemm example_batched_gemm_scale_softmax_gemm_permute_xdl_fp16)
+endif()
+add_example_executable(example_grouped_gemm_scale_softmax_gemm_permute_xdl_fp16 grouped_gemm_scale_softmax_gemm_permute_xdl_fp16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_gemm_scale_softmax_gemm example_grouped_gemm_scale_softmax_gemm_permute_xdl_fp16)
+endif()
+add_example_executable(example_batched_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16 batched_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_gemm_scale_softmax_gemm example_batched_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16)
+endif()
+add_example_executable(example_grouped_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16 grouped_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_gemm_scale_softmax_gemm example_grouped_gemm_lower_triangle_scale_softmax_gemm_permute_xdl_fp16)
 endif()
-if(DTYPES MATCHES "bf16" OR NOT DEFINED DTYPES)
+add_example_executable(example_batched_gemm_scale_softmax_gemm_xdl_bf16 batched_gemm_scale_softmax_gemm_xdl_bf16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_gemm_scale_softmax_gemm example_batched_gemm_scale_softmax_gemm_xdl_bf16)
+endif()
+add_example_executable(example_batched_gemm_scale_softmax_gemm_permute_xdl_bf16 batched_gemm_scale_softmax_gemm_permute_xdl_bf16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_gemm_scale_softmax_gemm example_batched_gemm_scale_softmax_gemm_permute_xdl_bf16)
 endif()
+

example/35_splitK_gemm/CMakeLists.txt

@@ -3,25 +3,28 @@ set(target 0)
 foreach(gpu IN LISTS GPU_TARGETS)
  if(gpu IN_LIST gpu_list AND target EQUAL 0)
    add_custom_target(example_splitK_gemm_xdl)
-   if(DTYPES MATCHES "fp32" OR NOT DEFINED DTYPES)
-    add_example_executable(example_splitK_gemm_xdl_fp32 splitK_gemm_xdl_fp32.cpp)
+
+  add_example_executable(example_splitK_gemm_xdl_fp32 splitK_gemm_xdl_fp32.cpp)
+  if(result EQUAL 0)
     add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_fp32)
    endif()
-   if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_splitK_gemm_xdl_fp16 splitK_gemm_xdl_fp16.cpp)
+  add_example_executable(example_splitK_gemm_xdl_fp16 splitK_gemm_xdl_fp16.cpp)
+  if(result EQUAL 0)
     add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_fp16)
    endif()
-   if(DTYPES MATCHES "bf16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_splitK_gemm_xdl_bfp16 splitK_gemm_xdl_bfp16.cpp)
-    add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_bfp16)
+  add_example_executable(example_splitK_gemm_xdl_bf16 splitK_gemm_xdl_bf16.cpp)
+  if(result EQUAL 0)
+    add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_bf16)
    endif()
-   if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
-    add_example_executable(example_splitK_gemm_xdl_int8 splitK_gemm_xdl_int8.cpp)
+  add_example_executable(example_splitK_gemm_xdl_int8 splitK_gemm_xdl_int8.cpp)
+  if(result EQUAL 0)
     add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_int8)
    endif()
    if(USE_BITINT_EXTENSION_INT4)
       add_example_executable(example_splitK_gemm_xdl_int4 splitK_gemm_xdl_int4.cpp)
-      add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_int4)
+      if(result EQUAL 0)
+        add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_int4)
+      endif()
    endif()
    set(target 1)
  endif()

example/37_batched_gemm_add_add_relu_gemm_add/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_batched_gemm_add_add_relu_gemm_add_xdl_fp16 batched_gemm_add_add_relu_gemm_add_xdl_fp16.cpp)
-endif()
+add_example_executable(example_batched_gemm_add_add_relu_gemm_add_xdl_fp16 batched_gemm_add_add_relu_gemm_add_xdl_fp16.cpp)

example/38_grouped_conv_bwd_data_multiple_d/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
 list(APPEND gpu_list gfx908 gfx90a gfx940 gfx941 gfx942)
 set(target 0)
 foreach(gpu IN LISTS GPU_TARGETS)
  if(gpu IN_LIST gpu_list AND target EQUAL 0)
    add_custom_target(example_grouped_conv_bwd_data)
    add_example_executable(example_grouped_conv_bwd_data_fp16 grouped_conv_bwd_data_fp16.cpp)
+   if(result EQUAL 0)
+      add_dependencies(example_grouped_conv_bwd_data example_grouped_conv_bwd_data_fp16)
+   endif()
    add_example_executable(example_grouped_conv_bwd_data_bias_relu_fp16 grouped_conv_bwd_data_bias_relu_fp16.cpp)
-
-   add_dependencies(example_grouped_conv_bwd_data example_grouped_conv_bwd_data_fp16)
-   add_dependencies(example_grouped_conv_bwd_data example_grouped_conv_bwd_data_bias_relu_fp16)
+   if(result EQUAL 0)
+     add_dependencies(example_grouped_conv_bwd_data example_grouped_conv_bwd_data_bias_relu_fp16)
+   endif()
    set(target 1)
  endif()
 endforeach()
-endif()

example/39_permute/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_custom_target(example_permute)
-
-    add_example_executable(example_permute_1xHxW_fp16 permute_1xHxW_fp16.cpp)
-    add_example_executable(example_permute_NxHxW_fp16 permute_NxHxW_fp16.cpp)
-    add_example_executable(example_permute_HxWx4_fp16 permute_HxWx4_fp16.cpp)
+add_custom_target(example_permute)
 
+add_example_executable(example_permute_1xHxW_fp16 permute_1xHxW_fp16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_permute example_permute_1xHxW_fp16)
+endif()
+add_example_executable(example_permute_NxHxW_fp16 permute_NxHxW_fp16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_permute example_permute_NxHxW_fp16)
+endif()
+add_example_executable(example_permute_HxWx4_fp16 permute_HxWx4_fp16.cpp)
+if(result EQUAL 0)
     add_dependencies(example_permute example_permute_HxWx4_fp16)
 endif()

example/40_conv2d_fwd_quantization/CMakeLists.txt

-if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
 list(APPEND gpu_list gfx908 gfx90a gfx940 gfx941 gfx942)
 set(target 0)
 foreach(gpu IN LISTS GPU_TARGETS)
@@ -11,7 +10,6 @@ foreach(gpu IN LISTS GPU_TARGETS)
  endif()
 endforeach()
 
-  if(DL_KERNELS)
   # Conv perlayer quantization
   add_example_executable(example_conv2d_fwd_dl_perlayer_quantization_int8 conv2d_fwd_dl_perlayer_quantization_int8.cpp)
   # Conv perchannel quantization
@@ -24,5 +22,3 @@ endforeach()
   add_example_executable(example_conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8 conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8.cpp)
   # Conv + bias + tanh perchannel quantization
   add_example_executable(example_conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8 conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8.cpp)
-  endif()
-endif()
\ No newline at end of file

example/41_grouped_conv_conv_fwd/CMakeLists.txt

@@ -3,15 +3,9 @@ list(APPEND gpu_list2 gfx908 gfx90a)
 set(target 0)
 foreach(gpu IN LISTS GPU_TARGETS)
  if(gpu IN_LIST gpu_list1 AND target EQUAL 0)
-   if(DTYPES MATCHES "fp32" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_conv_fwd_xdl_fp32 grouped_conv_conv_fwd_xdl_fp32.cpp)
-   endif()
-   if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_conv_fwd_xdl_fp16 grouped_conv_conv_fwd_xdl_fp16.cpp)
-   endif()
-   if(DTYPES MATCHES "bf16" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_conv_fwd_xdl_bf16 grouped_conv_conv_fwd_xdl_bf16.cpp)
-   endif()
+   add_example_executable(example_grouped_conv_conv_fwd_xdl_fp32 grouped_conv_conv_fwd_xdl_fp32.cpp)
+   add_example_executable(example_grouped_conv_conv_fwd_xdl_fp16 grouped_conv_conv_fwd_xdl_fp16.cpp)
+   add_example_executable(example_grouped_conv_conv_fwd_xdl_bf16 grouped_conv_conv_fwd_xdl_bf16.cpp)
    if(USE_BITINT_EXTENSION_INT4)
       add_example_executable(example_grouped_conv_conv_fwd_xdl_int4 grouped_conv_conv_fwd_xdl_int4.cpp)
    endif(USE_BITINT_EXTENSION_INT4)
@@ -20,7 +14,5 @@ foreach(gpu IN LISTS GPU_TARGETS)
 endforeach()
 
 if(NOT GPU_TARGETS MATCHES "gfx94" AND NOT GPU_TARGETS MATCHES "gfx1")
-   if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
-      add_example_executable(example_grouped_conv_conv_fwd_xdl_int8 grouped_conv_conv_fwd_xdl_int8.cpp)
-   endif()
+   add_example_executable(example_grouped_conv_conv_fwd_xdl_int8 grouped_conv_conv_fwd_xdl_int8.cpp)
 endif()

example/42_groupnorm/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_groupnorm_sigmoid_mul_fp16 groupnorm_sigmoid_mul_fp16.cpp)
-    add_example_executable(example_groupnorm_splitk_fp16 groupnorm_splitk_fp16.cpp)
-    add_example_executable(example_groupnorm_swish_fp16 groupnorm_swish_fp16.cpp)
-endif()
+add_example_executable(example_groupnorm_sigmoid_mul_fp16 groupnorm_sigmoid_mul_fp16.cpp)
+add_example_executable(example_groupnorm_splitk_fp16 groupnorm_splitk_fp16.cpp)
+add_example_executable(example_groupnorm_swish_fp16 groupnorm_swish_fp16.cpp)

example/43_splitk_gemm_bias_e_permute/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_splitk_gemm_bias_e_permute_xdl_fp16 splitk_gemm_bias_e_permute_xdl_fp16.cpp)
-endif()
-if(DTYPES MATCHES "fp32" OR NOT DEFINED DTYPES)
-    add_example_executable(example_splitk_gemm_bias_e_permute_xdl_fp32 splitk_gemm_bias_e_permute_xdl_fp32.cpp)
-endif()
+add_example_executable(example_splitk_gemm_bias_e_permute_xdl_fp16 splitk_gemm_bias_e_permute_xdl_fp16.cpp)
+add_example_executable(example_splitk_gemm_bias_e_permute_xdl_fp32 splitk_gemm_bias_e_permute_xdl_fp32.cpp)

example/44_elementwise_permute/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_elementwise_permute_4D_fp16 elementwise_permute_4D_fp16.cpp)
-    add_example_executable(example_elementwise_permute_4D_fp16_2d elementwise_permute_4D_fp16_2d.cpp)
-endif()
+add_example_executable(example_elementwise_permute_4D_fp16 elementwise_permute_4D_fp16.cpp)
+add_example_executable(example_elementwise_permute_4D_fp16_2d elementwise_permute_4D_fp16_2d.cpp)

example/46_gemm_add_multiply/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    if(DL_KERNELS)
-        add_example_executable(example_gemm_add_multiply_dl_fp16 gemm_add_multiply_dl_fp16.cpp)
-    endif()
-    add_example_executable(example_gemm_add_multiply_xdl_fp16 gemm_add_multiply_xdl_fp16.cpp)
-endif()
+add_example_executable(example_gemm_add_multiply_dl_fp16 gemm_add_multiply_dl_fp16.cpp)
+add_example_executable(example_gemm_add_multiply_xdl_fp16 gemm_add_multiply_xdl_fp16.cpp)

example/48_pool3d_fwd/CMakeLists.txt

-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_pool3d_fwd_fp16 pool3d_fwd_fp16.cpp)
-endif()
+add_example_executable(example_pool3d_fwd_fp16 pool3d_fwd_fp16.cpp)

example/49_maxpool2d_bwd/CMakeLists.txt

-if(DTYPES MATCHES "bf16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_maxpool2d_bwd_bf16 maxpool2d_bwd_bf16.cpp)
-endif()
-if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
-    add_example_executable(example_maxpool2d_bwd_fp16 maxpool2d_bwd_fp16.cpp)
-endif()
-if(DTYPES MATCHES "fp32" OR NOT DEFINED DTYPES)
-    add_example_executable(example_maxpool2d_bwd_fp32 maxpool2d_bwd_fp32.cpp)
-endif()
+add_example_executable(example_maxpool2d_bwd_bf16 maxpool2d_bwd_bf16.cpp)
+add_example_executable(example_maxpool2d_bwd_fp16 maxpool2d_bwd_fp16.cpp)
+add_example_executable(example_maxpool2d_bwd_fp32 maxpool2d_bwd_fp32.cpp)