Addressing (Post Merge) code review comments for PR 1845 (#1883)

* Addressing code review comments.

* Addressing code review comments.

* Reorganized code for better readability.

* add ck_tile gemms for new types in CI

* fix jenkins syntax

* fix script syntax

* Add the test cases back

* Address the review comments

* Address review comments

* clang format

* Solve the merging issues

* Addressed the comments

* clang format

---------

Co-authored-by: illsilin <Illia.Silin@amd.com>
Co-authored-by: ThomasNing <thomas.ning@amd.com>
Co-authored-by: Adam Osewski <19374865+aosewski@users.noreply.github.com>

[ROCm/composable_kernel commit: 66c5f5b0b6]

script/process_perf_data.py

@@ -344,6 +344,30 @@ def main():
             for i in range(1, len(results)+1):
                 testlist.append("Test%i"%i)
             table_name="ck_tile_gemm_mem_pipeline_fp16_tflops"
+        if 'gemm_basic_bf16' in filename:
+            for i in range(1, len(results)+1):
+                testlist.append("Test%i"%i)
+            table_name="ck_tile_gemm_basic_bf16_tflops"
+        if 'gemm_mem_pipeline_bf16' in filename:
+            for i in range(1, len(results)+1):
+                testlist.append("Test%i"%i)
+            table_name="ck_tile_gemm_mem_pipeline_bf16_tflops"
+        if 'gemm_basic_fp8' in filename:
+            for i in range(1, len(results)+1):
+                testlist.append("Test%i"%i)
+            table_name="ck_tile_gemm_basic_fp8_tflops"
+        if 'gemm_mem_pipeline_fp8' in filename:
+            for i in range(1, len(results)+1):
+                testlist.append("Test%i"%i)
+            table_name="ck_tile_gemm_mem_pipeline_fp8_tflops"
+        if 'gemm_basic_bf8' in filename:
+            for i in range(1, len(results)+1):
+                testlist.append("Test%i"%i)
+            table_name="ck_tile_gemm_basic_bf8_tflops"
+        if 'gemm_mem_pipeline_bf8' in filename:
+            for i in range(1, len(results)+1):
+                testlist.append("Test%i"%i)
+            table_name="ck_tile_gemm_mem_pipeline_bf8_tflops"
 
         tflops_base = get_baseline(table_name,conn)
         store_new_test_result(table_name, results, testlist, branch_name, node_id, gpu_arch, compute_units, rocm_vers, hip_vers, environment, sqlEngine)

script/process_perf_data.sh

@@ -43,19 +43,12 @@ file=./perf_fmha_bwd_gfx90a.log
 if [ -e "$file" ]; then
     python3 process_perf_data.py perf_fmha_bwd_gfx90a.log
 fi
-file=./perf_tile_gemm_basic_fp16_gfx942.log
-if [ -e "$file" ]; then
-    python3 process_perf_data.py perf_tile_gemm_basic_fp16_gfx942.log
-fi
-file=./perf_tile_gemm_basic_fp16_gfx90a.log
-if [ -e "$file" ]; then
-    python3 process_perf_data.py perf_tile_gemm_basic_fp16_gfx90a.log
-fi
-file=./perf_tile_gemm_mem_pipeline_fp16_gfx942.log
-if [ -e "$file" ]; then
-    python3 process_perf_data.py perf_tile_gemm_mem_pipeline_fp16_gfx942.log
-fi
-file=./perf_tile_gemm_mem_pipeline_fp16_gfx90a.log
-if [ -e "$file" ]; then
-    python3 process_perf_data.py perf_tile_gemm_mem_pipeline_fp16_gfx90a.log
-fi
+
+for gpu in "gfx90a" "gfx942"; do
+    for dtype in "fp16" "bf16" "fp8" "bf8"; do
+        file=./perf_tile_gemm_mem_pipeline_${dtype}_${gpu}.log
+        if [ -e "$file" ]; then
+            python3 process_perf_data.py perf_tile_gemm_mem_pipeline_${dtype}_${gpu}.log
+        fi
+    done
+done

script/process_qa_data.sh

@@ -52,19 +52,12 @@ file=./perf_fmha_bwd_gfx90a.log
 if [ -e "$file" ]; then
     python3 process_perf_data.py perf_fmha_bwd_gfx90a.log
 fi
-file=./perf_gemm_basic_gfx942.log
-if [ -e "$file" ]; then
-    python3 process_perf_data.py perf_gemm_basic_gfx942.log
-fi
-file=./perf_gemm_basic_gfx90a.log
-if [ -e "$file" ]; then
-    python3 process_perf_data.py perf_gemm_basic_gfx90a.log
-fi
-file=./perf_gemm_mem_pipeline_gfx942.log
-if [ -e "$file" ]; then
-    python3 process_perf_data.py perf_gemm_mem_pipeline_gfx942.log
-fi
-file=./perf_gemm_mem_pipeline_gfx90a.log
-if [ -e "$file" ]; then
-    python3 process_perf_data.py perf_gemm_mem_pipeline_gfx90a.log
-fi
+
+for gpu in "gfx90a" "gfx942"; do
+    for dtype in "fp16" "bf16" "fp8" "bf8"; do
+        file=./perf_tile_gemm_mem_pipeline_${dtype}_${gpu}.log
+        if [ -e "$file" ]; then
+            python3 process_perf_data.py perf_tile_gemm_mem_pipeline_${dtype}_${gpu}.log
+        fi
+    done
+done