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

SWDEV-492625 memory free functions (#11)

Browse Source 
* SWDEV-492625: Track free memory HSA functions to help determine total amount of memory allocated on the system at any one time

* Minor fixes to address comments

* Update allocation size description

* Moved get function back to specialization, minor typo fixes

* Removed memory_operation_type field, removed memory_pool allocation enum, converted starting address to hex string for json format.

* Made conversion to hex_string a function, changed address to use union rocprofiler_address_t type, changed VMEM descriptors

* Removed as_hex from the global namespace

* Formatting

* Removed TRACK_EVENT for memory allocation, now TRACK_COUNTER for memory allocation is being performed

* Check if address was recorded before retrieving allocation size in generate Perfetto

* Formatting

* Update source/lib/output/generatePerfetto.cpp

* Explicitly disable app-abort tests

* Remove excluding app-abort test from workflow CI

- redundant bc these tests are explicitly marked as disabled now

---------

Co-authored-by: Madsen, Jonathan <Jonathan.Madsen@amd.com>
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>

[ROCm/rocprofiler-sdk commit: 79006bb896]
This commit is contained in:
Trowbridge, Ian
2024-12-06 00:05:30 -06:00
committed by GitHub
parent a79f8a0198
commit 792329fefd
23 changed files with 564 additions and 248 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rocprofiler-sdk/tests/tools/json-tool.cpp
+33 -30
View File
@@ -49,6 +49,7 @@
#include <rocprofiler-sdk/internal_threading.h>
#include <rocprofiler-sdk/registration.h>
#include <rocprofiler-sdk/rocprofiler.h>
#include <rocprofiler-sdk/cxx/utility.hpp>
#include <unistd.h>
#include <algorithm>
@@ -1843,6 +1844,7 @@ write_perfetto()
auto tids = std::set<rocprofiler_thread_id_t>{};
auto agent_ids = std::set<uint64_t>{};
auto agent_ids_alloc = std::set<uint64_t>{};
auto agent_queue_ids = std::map<uint64_t, std::set<uint64_t>>{};
auto _get_agent = [](uint64_t id_handle) -> const rocprofiler_agent_t* {
@@ -1875,7 +1877,7 @@ write_perfetto()
for(auto itr : memory_allocation_bf_records)
{
tids.emplace(itr.thread_id);
agent_ids.emplace(itr.agent_id.handle);
agent_ids_alloc.emplace(itr.agent_id.handle);
}
for(auto itr : kernel_dispatch_bf_records)
@@ -1934,6 +1936,36 @@ write_perfetto()
agent_tracks.emplace(itr, _track);
}
for(auto itr : agent_ids_alloc)
{
const auto* _agent = _get_agent(itr);
auto _namess = std::stringstream{};
if(_agent != nullptr)
{
if(_agent->type == ROCPROFILER_AGENT_TYPE_CPU)
_namess << "CPU MEMORY OPERATION [" << itr << "] ";
else if(_agent->type == ROCPROFILER_AGENT_TYPE_GPU)
_namess << "GPU MEMORY OPERATION [" << itr << "] ";
if(!std::string_view{_agent->model_name}.empty())
_namess << _agent->model_name;
else
_namess << _agent->product_name;
}
else
{
_namess << "UNKNOWN MEMORY OPERATION [" << itr << "] ";
}
auto _track = ::perfetto::Track{get_hash_id(_namess.str())};
auto _desc = _track.Serialize();
_desc.set_name(_namess.str());
perfetto::TrackEvent::SetTrackDescriptor(_track, _desc);
agent_tracks.emplace(itr, _track);
}
auto agent_queue_tracks =
std::unordered_map<uint64_t, std::unordered_map<uint64_t, ::perfetto::Track>>{};
@@ -2155,35 +2187,6 @@ write_perfetto()
itr.end_timestamp);
}
for(auto itr : memory_allocation_bf_records)
{
auto name = buffer_names.at(itr.kind, itr.operation);
auto& track = agent_tracks.at(itr.agent_id.handle);
TRACE_EVENT_BEGIN(sdk::perfetto_category<sdk::category::memory_allocation>::name,
::perfetto::StaticString(name.data()),
track,
itr.start_timestamp,
::perfetto::Flow::ProcessScoped(itr.correlation_id.internal),
"begin_ns",
itr.start_timestamp,
"kind",
itr.kind,
"operation",
itr.operation,
"agent",
agents_map.at(itr.agent_id).logical_node_id,
"Allocation_size",
itr.allocation_size,
"Starting_address",
itr.starting_address);
TRACE_EVENT_END(sdk::perfetto_category<sdk::category::memory_allocation>::name,
track,
itr.end_timestamp,
"end_ns",
itr.end_timestamp);
}
auto demangled = std::unordered_map<std::string_view, std::string>{};
for(auto itr : kernel_dispatch_bf_records)
{