AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Put cached perfetto traces as default one (#2138)

Browse Source 
* Put cached perfetto traces as default one

* Improve cached data and perfetto traces in order to be more aligned with E2E tests

* Addressing PR comments and findings

* Force early instrumentation bundle instantiation

* Sync-up insturumented containers with thread growth data

* Revert ompvv number of host threads to default 8

* Fixed counter track namings for amd-smi

* AIPROFSYST-34 [rocprof-sys] Update documentation describing newly introduced changes to default tracing mechanism
This commit is contained in:
marantic-amd
2025-12-22 12:47:35 +01:00
committed by GitHub
parent 7da3275b42
commit ba1380a75d
36 changed files with 557 additions and 183 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rocprofiler-systems/tests/rocprof-sys-python-tests.cmake
+65 -13
View File
@@ -164,15 +164,28 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
TEST
""
"NAME;TIMEMORY_METRIC;TIMEMORY_FILE;PERFETTO_FILE"
"ARGS;PERFETTO_METRIC;ROCPD_FILE;ROCPD_RULES"
"ARGS;TIMEMORY_ARGS;PERFETTO_ARGS;PERFETTO_METRIC;ROCPD_FILE;ROCPD_RULES"
${ARGN}
)
# Use specific args if provided, otherwise fall back to common ARGS
if(TEST_TIMEMORY_ARGS)
set(_TIMEMORY_VALIDATION_ARGS ${TEST_TIMEMORY_ARGS})
else()
set(_TIMEMORY_VALIDATION_ARGS ${TEST_ARGS})
endif()
if(TEST_PERFETTO_ARGS)
set(_PERFETTO_VALIDATION_ARGS ${TEST_PERFETTO_ARGS})
else()
set(_PERFETTO_VALIDATION_ARGS ${TEST_ARGS})
endif()
rocprofiler_systems_add_python_test(
NAME ${TEST_NAME}-validate-timemory
COMMAND
${_PYTHON_EXECUTABLE} ${CMAKE_CURRENT_LIST_DIR}/validate-timemory-json.py
-m ${TEST_TIMEMORY_METRIC} ${TEST_ARGS} -i
-m ${TEST_TIMEMORY_METRIC} ${_TIMEMORY_VALIDATION_ARGS} -i
PYTHON_VERSION ${_VERSION}
FILE rocprof-sys-tests-output/${TEST_NAME}/${_VERSION}/${TEST_TIMEMORY_FILE}
DEPENDS ${TEST_NAME}-${_VERSION}
@@ -185,7 +198,7 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
NAME ${TEST_NAME}-validate-perfetto
COMMAND
${_PYTHON_EXECUTABLE} ${CMAKE_CURRENT_LIST_DIR}/validate-perfetto-proto.py
-m ${TEST_PERFETTO_METRIC} ${TEST_ARGS} -p -t
-m ${TEST_PERFETTO_METRIC} ${_PERFETTO_VALIDATION_ARGS} -p -t
/opt/trace_processor/bin/trace_processor_shell -i
PYTHON_VERSION ${_VERSION}
FILE rocprof-sys-tests-output/${TEST_NAME}/${_VERSION}/${TEST_PERFETTO_FILE}
@@ -217,7 +230,8 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
endif()
endfunction()
set(python_source_labels
# Timemory validation uses hierarchical output with multiple entries at different depths
set(python_source_timemory_labels
main_loop
run
fib
@@ -228,7 +242,7 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
inefficient
_sum
)
set(python_source_count
set(python_source_timemory_count
5
3
3
@@ -239,7 +253,7 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
3
3
)
set(python_source_depth
set(python_source_timemory_depth
0
1
2
@@ -251,6 +265,29 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
3
)
# Perfetto (cached mode) aggregates entries by name
set(python_source_perfetto_labels
main_loop
run
fib
inefficient
_sum
)
set(python_source_perfetto_count
5
3
24
3
3
)
set(python_source_perfetto_depth
0
1
2
2
3
)
set(python_source_categories python user)
rocprofiler_systems_add_python_validation_test(
@@ -259,14 +296,17 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
TIMEMORY_FILE "trip_count.json"
PERFETTO_FILE "perfetto-trace.proto"
PERFETTO_METRIC ${python_source_categories}
ARGS -l ${python_source_labels} -c ${python_source_count} -d
${python_source_depth}
TIMEMORY_ARGS -l ${python_source_timemory_labels} -c ${python_source_timemory_count} -d
${python_source_timemory_depth}
PERFETTO_ARGS -l ${python_source_perfetto_labels} -c ${python_source_perfetto_count} -d
${python_source_perfetto_depth}
ROCPD_FILE "rocpd.db"
ROCPD_RULES
"${CMAKE_CURRENT_LIST_DIR}/rocpd-validation-rules/python/python-source-rules.json"
)
set(python_builtin_labels
# Timemory validation uses hierarchical output with multiple entries at different depths
set(python_builtin_timemory_labels
[run][builtin.py:31]
[fib][builtin.py:13]
[fib][builtin.py:13]
@@ -280,7 +320,7 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
[fib][builtin.py:13]
[inefficient][builtin.py:17]
)
set(python_builtin_count
set(python_builtin_timemory_count
5
5
10
@@ -294,7 +334,7 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
10
5
)
set(python_builtin_depth
set(python_builtin_timemory_depth
0
1
2
@@ -309,14 +349,26 @@ foreach(_VERSION ${ROCPROFSYS_PYTHON_VERSIONS})
1
)
# Perfetto validation with trace caching aggregates all calls to the same function,
# so we only expect one entry per unique label rather than hierarchical entries.
set(python_builtin_perfetto_labels
[run][builtin.py:31]
[fib][builtin.py:13]
[inefficient][builtin.py:17]
)
set(python_builtin_perfetto_count 5 445 5)
set(python_builtin_perfetto_depth 0 1 1)
rocprofiler_systems_add_python_validation_test(
NAME python-builtin
TIMEMORY_METRIC "trip_count"
TIMEMORY_FILE "trip_count.json"
PERFETTO_METRIC "python"
PERFETTO_FILE "perfetto-trace.proto"
ARGS -l ${python_builtin_labels} -c ${python_builtin_count} -d
${python_builtin_depth}
TIMEMORY_ARGS -l ${python_builtin_timemory_labels} -c ${python_builtin_timemory_count} -d
${python_builtin_timemory_depth}
PERFETTO_ARGS -l ${python_builtin_perfetto_labels} -c ${python_builtin_perfetto_count} -d
${python_builtin_perfetto_depth}
ROCPD_FILE "rocpd.db"
ROCPD_RULES
"${CMAKE_CURRENT_LIST_DIR}/rocpd-validation-rules/python/python-builtin-rules.json"