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

Global trace delay and duration (#235)

Browse Source 
- The primary feature of this PR is the **addition of support for scoping the collection of tracing/profiling data into one or more time-based windows**
  - Closes #222 
  - Closes #207
  - Support for a real-clock time delay and/or a duration for tracing/profiling was added, *resembling the support for this feature during sampling and process-sampling*
  - However, above paradigm was enhanced for tracing 
    - Instead of one delay and/or one duration based on real time, ***tracing supports periodic and varying delays and durations and these delay+duration sets can be controlled with different clocks***  
    - At some point, this capability will be extended to sampling and process-sampling
- A secondary feature of this PR are the improvements to the handling of categories (by-product of the primary feature)
  - For example, previously setting `OMNITRACE_ENABLE_CATEGORIES` to a specific set of categories only eliminated the disabled categories from the perfetto trace, now these are applied to timemory profiles too
  - A new configuration variable `OMNITRACE_DISABLE_CATEGORIES` was added for when disabling only a handful of categories is easier
- There are quite a few miscellaneous modifications which pollute this PR a bit

## Multiple Tracing Windows

As noted above, tracing now supports specifying multiple delays and durations _and_ with different clocks. Consider the configuration below with two entries in the format `<DELAY>:<DURATION>:<REPEAT>:<CLOCK_TYPE>`:

```console
OMNITRACE_TRACE_PERIODS = 0.5:1.0:2:realtime    10.0:5.0:3:cputime
```

The above configuration defines:
1. `0.5:1.0:2:realtime`
  - A delay of 0.5 seconds (real-time)
  - Followed by a data collection duration of 1 second (real-time)
  - This delay + duration is repeated 2x
  - Summary: tracing data is collected for 2 out of the first 3 seconds of the application's execution
2. `10.0:5.0:3:cputime`
  - A delay of 10 seconds (process _CPU-time_)
  - Followed by a data collection duration of 5 seconds (process _CPU-time_)
  - This delay + duration is repeated 3x
  - Summary: tracing data is collected for a total of 15 seconds of process CPU-time in the ensuing 75 seconds of CPU-time during the application execution. 
    - Note: the elapsed CPU-time is the aggregate of the CPU-time consumed by all the threads in the process and should be scaled accordingly, e.g., 4 threads running constantly for 1 second of real-time is ~4 seconds of CPU time. 

## `omnitrace-sample` Changes

Formerly, `--wait` and `--duration` command-line options only applied to sampling delay and duration. The value of these options are now applied to the tracing delay and duration. To retain the ability to control sampling delay/duration without setting tracing delay/duration or vice versa, `--sampling-wait`, `--sampling-duration`, `--trace-wait`, and `--trace-duration` options were added. `omnitrace-sample` also has new options for most of the new configuration options detailed below.

## New configuration options

| Option | Description |
| ------- | ----------- |
| `OMNITRACE_DISABLE_CATEGORIES` | inverse behavior from `OMNITRACE_ENABLE_CATEGORIES` -- populates list of all available categories and then removes the specified ones. |
| `OMNITRACE_TRACE_DELAY` | Single floating-point number specifying time to wait before starting data collection. Analagous to `OMNITRACE_SAMPLING_DELAY` and `OMNITRACE_PROCESS_SAMPLING_DELAY` |
| `OMNITRACE_TRACE_DURATION` | Single floating-point number specifying data collection duration. Analagous to `OMNITRACE_SAMPLING_DURATION` and `OMNITRACE_PROCESS_SAMPLING_DURATION` |
| `OMNITRACE_TRACE_PERIOD_CLOCK_ID` | Sets the default clock-type for tracing delay/duration. Always applied to above two options, can be overridden in below option. Accepts `CLOCK_REALTIME`, `CLOCK_MONOTONIC`, `CLOCK_PROCESS_CPUTIME_ID`, `CLOCK_MONOTONIC_RAW`, `CLOCK_REALTIME_COARSE`, `CLOCK_MONOTONIC_COARSE`, `CLOCK_BOOTTIME`. See `man 2 clock_gettime` for details on differences. |
| `OMNITRACE_TRACE_PERIODS` | More powerful version for specifying delay + duration. Supports formats: `<DELAY>`, `<DELAY>:<DURATION>`, `<DELAY>:<DURATION>:<REPEAT>`, and `<DELAY>:<DURATION>:<REPEAT>:<CLOCK_ID>`.  |

 ## Miscellaneous Changes

- Expanded `critical_trace_categories_t` to include tracing data from MPI, pthread, HIP, HSA, RCCL, NUMA, and Python.
- Added categories `thread_wall_time` and `thread_cpu_time` (derived from sampling)
- Read DWARF info for breakpoints
- Relocated some source code
  - Reason: necessary to make `libomnitrace` a bit more modular. Eventually, a large chunk will be separated into `libomnitrace-core`, `libomnitrace-binary`, etc. in order to facilitate re-usability
  - Relocated some functionality from `runtime.cpp` to `config.cpp`
  - Relocated code using rocm-smi library to query number of devices to `gpu.cpp` (where the code for using HIP to query number of devices is)
  - Relocated code for perfetto config and perfetto session out of tracing namespace to reside with other perfetto code
- `OMNITRACE_COLORIZED_LOG` configuration option renamed to `OMNITRACE_MONOCHROME`
  - Backwards compatibility via a deprecated option was not retained here since the logic changed (i.e. true in former means false in latter)
- Replaced `TIMEMORY_DEFAULT_OBJECT` macro with `OMNITRACE_DEFAULT_OBJECT` macro 
- Updated some code in roctracer to use `component::category_region` instead of explicitly using `tracing::` functions
- Updated `backtrace_metrics` to better support controlling their presence in the traces/profiles via categories
- Added support for `--print` in `validate-timemory-json.py`
- Generic `OMNITRACE_ADD_VALIDATION_TEST` CMake function

## Git Log

* OMNITRACE_DEFAULT_OBJECT

- replace TIMEMORY_DEFAULT_OBJECT with TIMEMORY_DEFAULT_OBJECT

* trace-time-window example + tests

- adds cmake OMNITRACE_ADD_VALIDATION_TEST function for testing
- validate-timemory-json.py now supports printing (-p)
- update to OMNITRACE_STRIP_TARGET

* Update timemory submodule

- detailed backtrace print /proc/<PID>/maps
- operation::push_node verbosity change
- storage::insert_hierarchy use emplace + at instead of operator[]
- concepts::is_type_listing
- argparse updates for start/end group
- argparse color fixes

* perfetto updates

- Remove OMNITRACE_CUSTOM_DATA_SOURCE CMake option
- move tracing::get_perfetto_config and tracing::get_perfetto_session to perfetto.cpp

* config and runtime updates

- OMNITRACE_DISABLE_CATEGORIES option
  - get_enabled_categories() + get_disabled_categories()
  - config impl handles populating them
- OMNITRACE_TRACE_DELAY option
- OMNITRACE_TRACE_DURATION option
- OMNITRACE_TRACE_PERIODS option
- {get,set}_signal_handler
  - removes config.cpp link dependency for omnitrace_finalize
- get_realtime_signal() + get_cputime_signal() + get_sampling_signals()
  - moved from runtime.cpp to config.cpp

* utility::convert

- helper function for converting string to a type

* pthread_create_gotcha + thread_info updates

- thread_index_data::as_string()
- tweak printing info about new thread / exited thread

* binary updates

- get_binary_info has arg to disable dwarf parsing
- binary_info contains vector of breakpoint addresses
- binary_info:filename() function
- binary::get_linked_path
- binary::get_link_map has args for dlopen mode
- symbol::read_dwarf -> symbol::read_dwarf_entries
- symbol::read_dwarf_breakpoints

* library updates + categories impl

- implement config::set_signal_handler
- categories.cpp for handling trace delays
  - implement trace delay/duration/periods

* concepts + debug + defines

- tuple_element in concepts
- removed runtime header from debug header
- OMNITRACE_DEFAULT_COPY_MOVE

* gpu + rocm_smi

- moved rsmi_num_monitor_devices call to gpu.cpp
  - gpu::rsmi_device_count()

* roctracer updates

- roctracer_bundle_t -> roctracer_hip_bundle_t
- use category_region instead of explicit tracing push/pop calls

* sampling + backtrace_metrics

- rework backtrace_metrics to support categories

* tracing updates

- category stack counters (i.e. push vs. pop counter) for profiling and tracing
- push_timemory and pop_timemory accept string_view instead of const char*
- tweaked the pop_timemory hash search
- {push,pop}_perfetto theoretically supports same invocations as for {push,pop}_perfetto_ts and {push,pop}_perfetto_track
- mark_perfetto, mark_perfetto_ts, mark_perfetto_track

* category_region update

- expanded the critical trace categories
- use category_push_disabled
- use category_pop_disabled
- use category_mark_disabled

* constraint implementation

- This provides generic functionality for constraining data collection within a windows of time.
 - E.g., delay, delay + duration, (delay + duration) * nrepeat

* COLORIZED_LOG -> MONOCHROME

* constraint + omnitrace-causal + omnitrace-sample updates

- support for using different clock IDs for constraints
- OMNITRACE_TRACE_PERIOD_CLOCK_ID option
- tweak to trace-time-window example
- tweak to trace-time-window tests

* Fix formatting

* Update time-window tests

- Fix detection of validation support for perfetto
- Using the --caller-include feature + runtime instrumentation on Ubuntu 18.04 and OpenSUSE 15.2 results in a segfault in the internals of Dyninst.
  - For now, mark that these tests will fail
  - Later, determine if updating Dyninst submodule fixes this problem

* Fix OMNITRACE_OUTPUT_PATH for all tests

- Provide absolute path instead of relative

* Tweak lambda for checking whether HW counters are enabled

- causing strange build errors on older GCC compilers

* Update dyninst submodule

- fix issues with using --caller-include for Ubuntu 18.04, OpenSUSE 15.x

* cmake formatting

* fix sampling compiler issue for GCC 8

* Tweak thread create message

* Increase causal validation iterations
This commit is contained in:
Jonathan R. Madsen
2023-02-03 14:10:42 -06:00
committed by GitHub
parent 2fb67c394b
commit 8feb6bf8b6
82 changed files with 2538 additions and 632 deletions
Download Patch File Download Diff File Expand all files Collapse all files
source/lib/omnitrace/library/config.cpp
+186 -35
View File
@@ -22,12 +22,12 @@
#include "library/config.hpp"
#include "common/defines.h"
#include "library/constraint.hpp"
#include "library/debug.hpp"
#include "library/defines.hpp"
#include "library/gpu.hpp"
#include "library/mproc.hpp"
#include "library/perfetto.hpp"
#include "library/runtime.hpp"
#include <timemory/backends/dmp.hpp>
#include <timemory/backends/mpi.hpp>
@@ -43,12 +43,14 @@
#include <timemory/settings/types.hpp>
#include <timemory/utility/argparse.hpp>
#include <timemory/utility/declaration.hpp>
#include <timemory/utility/delimit.hpp>
#include <timemory/utility/filepath.hpp>
#include <timemory/utility/join.hpp>
#include <timemory/utility/signals.hpp>
#include <algorithm>
#include <array>
#include <atomic>
#include <csignal>
#include <cstdint>
#include <cstdlib>
@@ -98,7 +100,7 @@ get_setting_name(std::string _v)
template <typename Tp>
Tp
get_available_perfetto_categories()
get_available_categories()
{
auto _v = Tp{};
for(auto itr : { OMNITRACE_PERFETTO_CATEGORIES })
@@ -287,8 +289,8 @@ configure_settings(bool _init)
"for continuous integration)",
false, "debugging", "advanced");
OMNITRACE_CONFIG_SETTING(bool, "OMNITRACE_COLORIZED_LOG", "Enable colorized logging",
true, "debugging", "advanced");
OMNITRACE_CONFIG_SETTING(bool, "OMNITRACE_MONOCHROME", "Disable colorized logging",
false, "debugging", "advanced");
OMNITRACE_CONFIG_EXT_SETTING(int, "OMNITRACE_DL_VERBOSE",
"Verbosity within the omnitrace-dl library", 0,
@@ -392,10 +394,45 @@ configure_settings(bool _init)
"Enable support for code coverage", false, "coverage",
"backend", "advanced");
OMNITRACE_CONFIG_SETTING(size_t, "OMNITRACE_INSTRUMENTATION_INTERVAL",
"Instrumentation only takes measurements once every N "
"function calls (not statistical)",
size_t{ 1 }, "instrumentation", "data_sampling", "advanced");
OMNITRACE_CONFIG_SETTING(
double, "OMNITRACE_TRACE_DELAY",
"Time in seconds to wait before enabling trace/profile data collection. If "
"multiple delays + durations are needed, see OMNITRACE_TRACE_PERIODS.",
0.0, "trace", "profile", "perfetto", "timemory");
OMNITRACE_CONFIG_SETTING(
double, "OMNITRACE_TRACE_DURATION",
"If > 0.0, time (in seconds) to collect trace/profile data. If multiple delays + "
"durations are needed, see OMNITRACE_TRACE_PERIODS.",
0.0, "trace", "profile", "perfetto", "timemory");
auto _clock_s =
config::get_setting_value<std::string>("OMNITRACE_TRACE_PERIOD_CLOCK_ID").second;
auto _clock_choices = std::vector<std::string>{};
for(const auto& itr : constraint::get_valid_clock_ids())
{
_clock_choices.emplace_back(
join("", "(", join('|', itr.name, itr.value, itr.raw_name), ")"));
}
OMNITRACE_CONFIG_SETTING(std::string, "OMNITRACE_TRACE_PERIODS",
"Similar to specify trace delay and/or duration except in "
"the form <DELAY>:<DURATION>, <DELAY>:<DURATION>:<REPEAT>, "
"and/or <DELAY>:<DURATION>:<REPEAT>:<CLOCK_ID>",
std::string{}, "trace", "profile", "perfetto", "timemory");
OMNITRACE_CONFIG_SETTING(
std::string, "OMNITRACE_TRACE_PERIOD_CLOCK_ID",
"Set the default clock ID for OMNITRACE_TRACE_DELAY, OMNITRACE_TRACE_DURATION, "
"and/or OMNITRACE_TRACE_PERIODS. E.g. \"realtime\" == the delay/duration is "
"governed by the elapsed realtime, \"cputime\" == the delay/duration is governed "
"by the elapsed CPU-time within the process, etc. Note: when using CPU-based "
"timing, it is recommened to scale the value by the number of threads and be "
"aware that omnitrace may contribute to advancing the process CPU-time",
"CLOCK_REALTIME", "trace", "profile", "perfetto", "timemory")
->set_choices(_clock_choices);
OMNITRACE_CONFIG_SETTING(
double, "OMNITRACE_SAMPLING_FREQ",
@@ -639,10 +676,18 @@ configure_settings(bool _init)
"discard", "perfetto", "data")
->set_choices({ "fill", "discard" });
OMNITRACE_CONFIG_SETTING(std::string, "OMNITRACE_PERFETTO_CATEGORIES",
"Categories to collect within perfetto", "", "perfetto",
"data", "advanced")
->set_choices(get_available_perfetto_categories<std::vector<std::string>>());
OMNITRACE_CONFIG_SETTING(std::string, "OMNITRACE_ENABLE_CATEGORIES",
"Enable collecting profiling and trace data for these "
"categories and disable all other categories",
"", "trace", "profile", "perfetto", "timemory", "data",
"advanced")
->set_choices(get_available_categories<std::vector<std::string>>());
OMNITRACE_CONFIG_SETTING(
std::string, "OMNITRACE_DISABLE_CATEGORIES",
"Disable collecting profiling and trace data for these categories", "", "trace",
"profile", "perfetto", "timemory", "data", "advanced")
->set_choices(get_available_categories<std::vector<std::string>>());
OMNITRACE_CONFIG_SETTING(bool, "OMNITRACE_PERFETTO_ANNOTATIONS",
"Include debug annotations in perfetto trace. When enabled, "
@@ -977,8 +1022,8 @@ configure_settings(bool _init)
settings::suppress_config() = true;
if(!get_env("OMNITRACE_COLORIZED_LOG", _config->get<bool>("OMNITRACE_COLORIZED_LOG")))
tim::log::colorized() = false;
if(get_env("OMNITRACE_MONOCHROME", _config->get<bool>("OMNITRACE_MONOCHROME")))
tim::log::monochrome() = true;
if(_init)
{
@@ -1105,8 +1150,6 @@ configure_mode_settings()
_set("OMNITRACE_USE_ROCM_SMI", false);
}
get_instrumentation_interval() = std::max<size_t>(get_instrumentation_interval(), 1);
if(get_use_kokkosp())
{
auto _current_kokkosp_lib = tim::get_env<std::string>("KOKKOS_PROFILE_LIBRARY");
@@ -1156,6 +1199,13 @@ namespace
using signal_settings = tim::signals::signal_settings;
using sys_signal = tim::signals::sys_signal;
std::atomic<signal_handler_t>&
get_signal_handler()
{
static auto _v = std::atomic<signal_handler_t>{ nullptr };
return _v;
}
void
omnitrace_exit_action(int nsig)
{
@@ -1163,7 +1213,8 @@ omnitrace_exit_action(int nsig)
tim::signals::sigmask_scope::process);
OMNITRACE_BASIC_PRINT("Finalizing afer signal %i :: %s\n", nsig,
signal_settings::str(static_cast<sys_signal>(nsig)).c_str());
if(get_state() == State::Active) omnitrace_finalize();
auto _handler = get_signal_handler().load();
if(_handler) (*_handler)();
kill(process::get_id(), nsig);
}
@@ -1183,6 +1234,28 @@ omnitrace_trampoline_handler(int _v)
}
} // namespace
signal_handler_t
set_signal_handler(signal_handler_t _func)
{
if(_func)
{
auto _handler = get_signal_handler().load(std::memory_order_relaxed);
if(get_signal_handler().compare_exchange_strong(_handler, _func,
std::memory_order_relaxed))
{
return _handler;
}
else
{
_handler = get_signal_handler().load(std::memory_order_seq_cst);
get_signal_handler().store(_func);
return _handler;
}
}
return get_signal_handler().load();
}
void
configure_signal_handler()
{
@@ -1218,6 +1291,35 @@ configure_signal_handler()
}
}
int
get_realtime_signal()
{
return SIGRTMIN + get_sampling_rtoffset();
}
int
get_cputime_signal()
{
return SIGPROF;
}
std::set<int> get_sampling_signals(int64_t)
{
auto _v = std::set<int>{};
if(get_use_causal())
{
_v.emplace(get_cputime_signal());
_v.emplace(get_realtime_signal());
}
else
{
if(get_use_sampling_cputime()) _v.emplace(get_cputime_signal());
if(get_use_sampling_realtime()) _v.emplace(get_realtime_signal());
}
return _v;
}
void
configure_disabled_settings()
{
@@ -1964,18 +2066,74 @@ get_perfetto_fill_policy()
return static_cast<tim::tsettings<std::string>&>(*_v->second).get();
}
std::set<std::string>
get_perfetto_categories()
namespace
{
static auto _v = get_config()->find("OMNITRACE_PERFETTO_CATEGORIES");
static auto _avail = get_available_perfetto_categories<std::set<std::string>>();
auto _ret = std::set<std::string>{};
for(auto itr : tim::delimit(
static_cast<tim::tsettings<std::string>&>(*_v->second).get(), " ,;:"))
{
if(_avail.count(itr) > 0) _ret.emplace(itr);
}
return _ret;
auto
get_category_config()
{
using strset_t = std::set<std::string>;
static auto _v = []() {
auto _avail = get_available_categories<strset_t>();
auto _parse = [&_avail](const auto& _setting) {
auto _ret = strset_t{};
for(auto itr : tim::delimit(
static_cast<tim::tsettings<std::string>&>(*_setting->second).get(),
" ,;:\n\t"))
{
if(_avail.count(itr) > 0) _ret.emplace(itr);
}
return _ret;
};
auto _enabled = _parse(get_config()->find("OMNITRACE_ENABLE_CATEGORIES"));
auto _disabled = _parse(get_config()->find("OMNITRACE_DISABLE_CATEGORIES"));
if(_enabled.empty() && _disabled.empty())
{
_enabled = _avail;
}
else if(_enabled.empty() && !_disabled.empty())
{
for(auto itr : _avail)
{
if(_disabled.count(itr) == 0) _enabled.emplace(itr);
}
}
else if(!_enabled.empty() && _disabled.empty())
{
for(auto itr : _avail)
{
if(_enabled.count(itr) == 0) _disabled.emplace(itr);
}
}
else
{
OMNITRACE_ABORT("Error! Conflicting options OMNITRACE_ENABLE_CATEGORIES and "
"OMNITRACE_DISABLE_CATEGORIES were both provided.");
}
OMNITRACE_CI_THROW(_enabled.size() + _disabled.size() != _avail.size(),
"Error! Internal error for categories: %zu (enabled) + %zu "
"(disabled) != %zu (total)\n",
_enabled.size(), _disabled.size(), _avail.size());
return std::make_pair(_enabled, _disabled);
}();
return _v;
}
} // namespace
std::set<std::string>
get_enabled_categories()
{
return get_category_config().first;
}
std::set<std::string>
get_disabled_categories()
{
return get_category_config().second;
}
bool
@@ -2043,13 +2201,6 @@ get_perfetto_output_filename()
return _val;
}
size_t&
get_instrumentation_interval()
{
static auto _v = get_config()->find("OMNITRACE_INSTRUMENTATION_INTERVAL");
return static_cast<tim::tsettings<size_t>&>(*_v->second).get();
}
double
get_sampling_freq()
{