Files
rocm-systems/tests/rocprofv3/tracing/validate.py
T
Ammar ELWazir 2905fb5e95 Update run-ci.py (#641)
* Temp: Fixing node id

* source formatting (clang-format v11) (#709)

Co-authored-by: ammarwa <3832908+ammarwa@users.noreply.github.com>

* Using logical node id

* Update agent.cpp

* Update agent.cpp

* Python formatting

* Update run-ci.py

* Update run-ci.py

* Update continuous_integration.yml

* Update continuous_integration.yml

running directly using the prepared runner container

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update run-ci.py

* Clean up

* Fixing install paths

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Update continuous_integration.yml

* Fixing GPU Agents Test Validation

* python formatting (black) (#712)

Co-authored-by: ammarwa <3832908+ammarwa@users.noreply.github.com>

* Fixing the issue with rocclr detected kernels __amd_rocclr_.*

* python formatting (black) (#713)

Co-authored-by: ammarwa <3832908+ammarwa@users.noreply.github.com>

* Fixing the issue with rocclr detected kernels __amd_rocclr_.*

* Fixing static number of async copies and using hsa_api instead for validation

* python formatting (black) (#714)

Co-authored-by: ammarwa <3832908+ammarwa@users.noreply.github.com>

* Increasing the time limit for waiting on active signals

* Update continuous_integration.yml

* Update async_copy.cpp

* Update CMakeLists.txt

* changing node id to logical node id in rocprofv3

* Update tool.cpp

* testing async mem copy signal decrement

* Update logging.cpp

* Update validate.py

---------

Co-authored-by: Ammar ELWazir <aelwazir@rocprofiler1.amd.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: ammarwa <3832908+ammarwa@users.noreply.github.com>
Co-authored-by: Ammar ELWazir <aelwazir@rocprofiler2.amd.com>
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
2024-04-02 01:39:24 -05:00

130 lines
4.3 KiB
Python

#!/usr/bin/env python3
import sys
import subprocess
import pytest
def test_hsa_api_trace(hsa_input_data):
functions = []
correlation_ids = []
for row in hsa_input_data:
assert row["Domain"] in (
"HSA_CORE_API",
"HSA_AMD_EXT_API",
"HSA_IMAGE_EXT_API",
"HSA_FINALIZE_EXT_API",
)
assert int(row["Process_Id"]) > 0
assert int(row["Thread_Id"]) >= int(row["Process_Id"])
assert int(row["End_Timestamp"]) >= int(row["Start_Timestamp"])
functions.append(row["Function"])
correlation_ids.append(int(row["Correlation_Id"]))
correlation_ids = sorted(list(set(correlation_ids)))
hsa_api_calls_offset = 2 # roctxRangePush is first
num_marker_api_calls = 7 # seven marker API calls, only six entries in
# marker csv data because roctxRangePush + roctxRangePop is one entry
# all correlation ids are unique
assert len(correlation_ids) == len(hsa_input_data)
# correlation ids are numbered from 1 to N
assert correlation_ids[0] == hsa_api_calls_offset
assert correlation_ids[-1] == len(correlation_ids) + num_marker_api_calls
functions = list(set(functions))
assert "hsa_amd_memory_async_copy_on_engine" in functions
def test_kernel_trace(kernel_input_data):
valid_kernel_names = (
"_Z15matrixTransposePfS_i.kd",
"matrixTranspose(float*, float*, int)",
)
assert len(kernel_input_data) == 1
for row in kernel_input_data:
assert row["Kind"] == "KERNEL_DISPATCH"
assert int(row["Agent_Id"]) > 0
assert int(row["Queue_Id"]) > 0
assert int(row["Kernel_Id"]) > 0
assert row["Kernel_Name"] in valid_kernel_names
assert int(row["Correlation_Id"]) > 0
assert int(row["Workgroup_Size_X"]) == 4
assert int(row["Workgroup_Size_Y"]) == 4
assert int(row["Workgroup_Size_Z"]) == 1
assert int(row["Grid_Size_X"]) == 1024
assert int(row["Grid_Size_Y"]) == 1024
assert int(row["Grid_Size_Z"]) == 1
assert int(row["End_Timestamp"]) >= int(row["Start_Timestamp"])
def test_memory_copy_trace(memory_copy_input_data):
for row in memory_copy_input_data:
assert row["Kind"] == "MEMORY_COPY"
assert len(memory_copy_input_data) == 2
row = memory_copy_input_data[0]
assert row["Direction"] == "HOST_TO_DEVICE"
output = subprocess.check_output(
'rocminfo | grep "Node: *'
+ row["Source_Agent_Id"]
+ '" -A 1 | grep "Device Type" | sed \'s/.*: *//\'',
shell=True,
)
assert int(str(output).find("CPU")) >= 0
output = subprocess.check_output(
'rocminfo | grep "Node: *'
+ row["Destination_Agent_Id"]
+ '" -A 1 | grep "Device Type" | sed \'s/.*: *//\'',
shell=True,
)
assert int(str(output).find("GPU")) >= 0
assert int(row["Correlation_Id"]) > 0
assert int(row["End_Timestamp"]) >= int(row["Start_Timestamp"])
row = memory_copy_input_data[1]
assert row["Direction"] == "DEVICE_TO_HOST"
output = subprocess.check_output(
'rocminfo | grep "Node: *'
+ row["Source_Agent_Id"]
+ '" -A 1 | grep "Device Type" | sed \'s/.*: *//\'',
shell=True,
)
assert int(str(output).find("GPU")) >= 0
output = subprocess.check_output(
'rocminfo | grep "Node: *'
+ row["Destination_Agent_Id"]
+ '" -A 1 | grep "Device Type" | sed \'s/.*: *//\'',
shell=True,
)
assert int(str(output).find("CPU")) >= 0
assert int(row["Correlation_Id"]) > 0
assert int(row["End_Timestamp"]) >= int(row["Start_Timestamp"])
def test_marker_api_trace(marker_input_data):
functions = []
for row in marker_input_data:
assert row["Domain"] in [
"MARKER_CORE_API",
"MARKER_CONTROL_API",
"MARKER_NAME_API",
]
assert int(row["Process_Id"]) > 0
assert int(row["Thread_Id"]) == 0 or int(row["Thread_Id"]) >= int(
row["Process_Id"]
)
assert int(row["End_Timestamp"]) >= int(row["Start_Timestamp"])
functions.append(row["Function"])
functions = list(set(functions))
assert "main" in functions
if __name__ == "__main__":
exit_code = pytest.main(["-x", __file__] + sys.argv[1:])
sys.exit(exit_code)