Fix hardware counter summary files not being generated after profiling (#124)

- Register a cleanup function in tim::manager instance to write out data in
counter storages

- The counter_storage::write() calls in tool_fini happen after the storage is destroyed
which is too late for the write to happen.

- Adjust traits for counter_data_tracker

- Add MIN, MAX, VAR, STDDEV columns
- Remove DEPTH, UNITS, %SELF columns

- Update "add_validation_test" to test for the existence of output file(s).
- Added step to test perfetto output for `transpose-rocprofiler-sampling`
and `transpose-rocprofiler-binary-rewrite`

---------

Co-authored-by: David Galiffi <David.Galiffi@amd.com>

tests/rocprof-sys-rocm-tests.cmake

@@ -62,10 +62,6 @@ rocprofiler_systems_add_test(
     REWRITE_FAIL_REGEX "0 instrumented loops in procedure transpose")
 
 if(ROCPROFSYS_USE_ROCM)
-    set(_ROCP_PASS_REGEX
-        "rocprof-device-0-GRBM_COUNT.txt(.*)rocprof-device-0-SQ_INSTS_VALU.txt(.*)rocprof-device-0-SQ_WAVES.txt(.*)rocprof-device-0-TA_TA_BUSY.txt(.*)"
-        )
-
     rocprofiler_systems_add_test(
         SKIP_BASELINE SKIP_RUNTIME
         NAME transpose-rocprofiler
@@ -80,4 +76,19 @@ if(ROCPROFSYS_USE_ROCM)
         REWRITE_RUN_PASS_REGEX "${_ROCP_PASS_REGEX}"
         SAMPLING_PASS_REGEX "${_ROCP_PASS_REGEX}")
 
+    rocprofiler_systems_add_validation_test(
+        NAME transpose-rocprofiler-sampling
+        PERFETTO_FILE "perfetto-trace.proto"
+        ARGS --counter-names "TA_TA_BUSY" "SQ_WAVES" "GRBM_COUNT" "SQ_INSTS_VALU" -p
+        EXIST_FILES rocprof-device-0-GRBM_COUNT.txt rocprof-device-0-TA_TA_BUSY.txt
+                    rocprof-device-0-SQ_INSTS_VALU.txt rocprof-device-0-SQ_WAVES.txt
+        LABELS "rocprofiler")
+
+    rocprofiler_systems_add_validation_test(
+        NAME transpose-rocprofiler-binary-rewrite
+        PERFETTO_FILE "perfetto-trace.proto"
+        ARGS --counter-names "TA_TA_BUSY" "SQ_WAVES" "GRBM_COUNT" "SQ_INSTS_VALU" -p
+        EXIST_FILES rocprof-device-0-GRBM_COUNT.txt rocprof-device-0-TA_TA_BUSY.txt
+                    rocprof-device-0-SQ_INSTS_VALU.txt rocprof-device-0-SQ_WAVES.txt
+        LABELS "rocprofiler")
 endif()

tests/rocprof-sys-testing.cmake

@@ -929,7 +929,7 @@ function(ROCPROFILER_SYSTEMS_ADD_VALIDATION_TEST)
         TEST
         ""
         "NAME;TIMEOUT;TIMEMORY_METRIC;TIMEMORY_FILE;PERFETTO_METRIC;PERFETTO_FILE"
-        "ENVIRONMENT;LABELS;PROPERTIES;PASS_REGEX;FAIL_REGEX;SKIP_REGEX;DEPENDS;ARGS"
+        "ENVIRONMENT;LABELS;PROPERTIES;PASS_REGEX;FAIL_REGEX;SKIP_REGEX;DEPENDS;EXIST_FILES;ARGS"
         ${ARGN})
 
     if(NOT TEST "${TEST_NAME}")
@@ -963,6 +963,14 @@ function(ROCPROFILER_SYSTEMS_ADD_VALIDATION_TEST)
             )
     endif()
 
+    foreach(_FILE ${TEST_EXIST_FILES})
+        add_test(
+            NAME validate-${TEST_NAME}-${_FILE}-exists
+            COMMAND test -e
+                    ${PROJECT_BINARY_DIR}/rocprof-sys-tests-output/${TEST_NAME}/${_FILE}
+            WORKING_DIRECTORY ${PROJECT_BINARY_DIR})
+    endforeach()
+
     if(TEST_TIMEMORY_FILE)
         add_test(
             NAME validate-${TEST_NAME}-timemory