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

Fix agent node id + randomize offset id (#625)

Browse Source 
* Fix agent node id + randomize offset id

- fixes the node_id value
- randomizes a constant offset for the id.handle values
- switch to using node ids in rocprofiler-sdk-tool library
- update tests related to agents

* Logical node id

- sequential node id values from 0 to (N-1) where N is the number of agents
This commit is contained in:
Jonathan R. Madsen
2024-03-21 20:04:21 -05:00
committed by GitHub
parent 2f9b1767e9
commit 1addfed9f6
8 changed files with 170 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/common/serialization.hpp
+1
View File
@@ -487,6 +487,7 @@ save(ArchiveT& ar, const rocprofiler_agent_t& data)
SAVE_DATA_CSTR(model_name);
SAVE_DATA_FIELD(num_pc_sampling_configs);
SAVE_DATA_FIELD(node_id);
SAVE_DATA_FIELD(logical_node_id);
auto generate = [&](auto name, const auto* value, uint64_t size) {
using value_type = std::remove_const_t<std::remove_pointer_t<decltype(value)>>;
tests/rocprofv3/tracing/validate.py
+2 -2
View File
@@ -68,13 +68,13 @@ def test_memory_copy_trace(memory_copy_input_data):
row = memory_copy_input_data[0]
assert row["Direction"] == "HOST_TO_DEVICE"
assert int(row["Source_Agent_Id"]) == 0
assert int(row["Destination_Agent_Id"]) == 1
assert int(row["Destination_Agent_Id"]) >= 1
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"
assert int(row["Source_Agent_Id"]) == 1
assert int(row["Source_Agent_Id"]) >= 1
assert int(row["Destination_Agent_Id"]) == 0
assert int(row["Correlation_Id"]) > 0
assert int(row["End_Timestamp"]) >= int(row["Start_Timestamp"])