Merge remote-tracking branch 'origin/develop' into aosewski/ggemm_multi_d2

script/run_full_performance_tests.sh

@@ -121,26 +121,16 @@ print_log_header $reduction_log $env_type $branch $host_name
 ./profile_reduce_no_index.sh $verify 2 10 --half 2>&1 | tee -a $reduction_log
 
 #run splitK_gemm tests, first correctness verification, then performance
-export splitK_gemm_ver_log="perf_splitK_gemm_verify.log"
-print_log_header $splitK_gemm_ver_log $env_type $branch $host_name
-./profile_splitK_gemm.sh gemm_splitk 0 0 $verify 1 0 0 4 2>&1 | tee -a $splitK_gemm_ver_log
-./profile_splitK_gemm.sh gemm_splitk 0 1 $verify 1 0 0 4 2>&1 | tee -a $splitK_gemm_ver_log
-./profile_splitK_gemm.sh gemm_splitk 0 2 $verify 1 0 0 4 2>&1 | tee -a $splitK_gemm_ver_log
-./profile_splitK_gemm.sh gemm_splitk 0 3 $verify 1 0 0 4 2>&1 | tee -a $splitK_gemm_ver_log
-./profile_splitK_gemm.sh gemm_splitk 1 0 $verify 1 0 0 4 2>&1 | tee -a $splitK_gemm_ver_log
-./profile_splitK_gemm.sh gemm_splitk 1 1 $verify 1 0 0 4 2>&1 | tee -a $splitK_gemm_ver_log
-./profile_splitK_gemm.sh gemm_splitk 1 2 $verify 1 0 0 4 2>&1 | tee -a $splitK_gemm_ver_log
-./profile_splitK_gemm.sh gemm_splitk 1 3 $verify 1 0 0 4 2>&1 | tee -a $splitK_gemm_ver_log
 export splitK_gemm_log="perf_splitK_gemm.log"
 print_log_header $splitK_gemm_log $env_type $branch $host_name
-./profile_splitK_gemm.sh gemm_splitk 0 0 0 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
-./profile_splitK_gemm.sh gemm_splitk 0 1 0 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
-./profile_splitK_gemm.sh gemm_splitk 0 2 0 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
-./profile_splitK_gemm.sh gemm_splitk 0 3 0 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
-./profile_splitK_gemm.sh gemm_splitk 1 0 0 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
-./profile_splitK_gemm.sh gemm_splitk 1 1 0 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
-./profile_splitK_gemm.sh gemm_splitk 1 2 0 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
-./profile_splitK_gemm.sh gemm_splitk 1 3 0 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
+./profile_splitK_gemm.sh gemm_splitk 0 0 $verify 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
+./profile_splitK_gemm.sh gemm_splitk 0 1 $verify 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
+./profile_splitK_gemm.sh gemm_splitk 0 2 $verify 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
+./profile_splitK_gemm.sh gemm_splitk 0 3 $verify 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
+./profile_splitK_gemm.sh gemm_splitk 1 0 $verify 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
+./profile_splitK_gemm.sh gemm_splitk 1 1 $verify 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
+./profile_splitK_gemm.sh gemm_splitk 1 2 $verify 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
+./profile_splitK_gemm.sh gemm_splitk 1 3 $verify 1 0 1 4 2>&1 | tee -a $splitK_gemm_log
 
 #run ONNX gemm tests
 export onnx_log="perf_onnx_gemm.log"

test/grouped_convnd_bwd_data/CMakeLists.txt

 list(APPEND gpu_list_xdl gfx908 gfx90a gfx940)
-list(APPEND gpu_list_wmma gfx1100 gfx1101 gfx1102)
+list(APPEND gpu_list_wmma gfx1100 gfx1101 gfx1102 gfx1103)
 set(target 0)
 foreach(gpu IN LISTS GPU_TARGETS)
     if(gpu IN_LIST gpu_list_xdl AND target EQUAL 0)

test/grouped_convnd_bwd_weight/CMakeLists.txt

 list(APPEND gpu_list_xdl gfx908 gfx90a gfx940 gfx941 gfx942)
-list(APPEND gpu_list_wmma gfx1100 gfx1101 gfx1102)
+list(APPEND gpu_list_wmma gfx1100 gfx1101 gfx1102 gfx1103)
 
 set(target 0)
 foreach(gpu IN LISTS GPU_TARGETS)