Rework sampling and colorized logs (#140)
## Overview
This is a significant PR which has 3 very notable characteristics:
1. Omnitrace colorizes most of it's logging
2. Completely reworked the sampling
- Samples now record the current instruction pointers instead of strings
- This _dramatically_ decreases the overhead of taking a sample
- The collection of metrics during a sample are split out into another component, enabling that data collection to be disabled -- which decreases the sampling overhead even further
- When both `OMNITRACE_SAMPLING_CPUTIME` and `OMNITRACE_SAMPLING_REALTIME` are ON:
- `OMNITRACE_SAMPLING_CPUTIME_FREQ` and `OMNITRACE_SAMPLING_REALTIME_FREQ` can be used to individually control the sampling frequency
- `OMNITRACE_SAMPLING_CPUTIME_DELAY` and `OMNITRACE_SAMPLING_REALTIME_DELAY` can be used to individually control the delay time before starting
- Now, omnitrace does not start a real-time sampler on the main thread unless `OMNITRACE_SAMPLING_REALTIME` is ON
- In the future, an `OMNITRACE_SAMPLING_TIDS` (and real-time, cpu-time variants) configuration variable(s) will allow you to select which threads will be sampled
3. Files produced by `omnitrace` exe -- `available-instr.txt`, `instrumented-instr.txt`, etc. -- now no longer has `-instr` suffix and are placed in `instrumentation/` subfolder, i.e. `available-instr.txt` -> instrumentation/available.txt`
- This helped de-clutter the output folder
Most of the other edits were reorganization (e.g. internal namespace changes), cleanup, and splitting up functionality.
## Bug Fixes
There is a bug fix with respect to the HSA callbacks which disabled sampling on child threads when an HSA API call was made
## Details
- created thread_info struct for mapping different thread IDs
- reorganized file structure significantly
- added categories.hpp, concepts.hpp
- moved around name trait definitions
- moved all omnitrace components into `omnitrace::component` namespace
- there was a lot of inconsistency b/t using `tim::component` in some places and `omnitrace::component`
- added macros like OMNITRACE_DECLARE_COMPONENT in lieu of TIMEMORY_DECLARE_COMPONENT
- OMNITRACE_CRITICAL_TRACE_NUM_THREADS -> OMNITRACE_THREAD_POOL_SIZE
- roctracer and critical_trace use same thread pool
- critical_trace functions do not lock anymore bc of thread-local TaskGroup
- added `component::local_category_region` to support using `component::category_region` without explicitly passing in name
- removed `component::omnitrace` (unused)
- migrated KokkosP and OMPT to use `component::local_category_region`
- removed `component::user_region` as a result
- migrated omnitrace_{push,pop}_{trace,region}_hidden to use component::category_region
- removed `component::functors` as a result
- migrated some ppdefs
- `api::omnitrace` -> `project::omnitrace`
- `api::(...)` -> `category::(...)`
- improved recording the execution time of threads
- migrated this functionality out of pthread_create_gotcha and into thread_info
- moved mpi_gotcha, fork_gotcha, exit_gotcha, rcclp into omnitrace::component namespace
- split backtrace up into backtrace, backtrace_metrics, backtrace_timestamp components
- sampling.cpp handles setup and post-processing that was formerly in backtrace
- updated logging to use colors
- `OMNITRACE_COLORIZED_LOG` config variable
- updated docs on JSON output from timemory
- instrumentation info in instrumentation subfolder
- added testing for KokkosP entries
- added testing for ompt entries
- add_critical_trace function defined in critical_trace.hpp
- disable push_thread_state and pop_thread_state when thread state is Disabled or Completed
- add comp::page_rss to main bundle
- thread_data supports std::optional instead of std::unique_ptr
- thread_data supports tim::identity<T> to avoid unique_ptr or optional
- tracing::record_thread_start_time()
- tracing::push_timemory and tracing::pop_timemory are templated on CategoryT
- removed anonymous namespace from omnitrace::utility
- sampling backtrace stores instruction pointers instead of strings
- component::category_region updates
- handle disabled thread state
- handle finalized state
- fewer debug messages
- invoke thread_init()
- invoke thread_init_sampling()
- handle push/pop count based on category
- push/pop count only modified when used
- component::cpu_freq
- components/ensure_storage.hpp
- reworked the pthread_create replacement function
- updated parallel-overhead example to report # of times locked
- OMNITRACE_MAX_UNWIND_DEPTH build option
- update timemory submodule
This commit is contained in:
Jonathan R. Madsen
zatwierdzone przez
GitHub
GitHub
rodzic
a1dcd1bc4b
commit
808ea7dfa7
@@ -22,30 +22,31 @@
|
||||
|
||||
#pragma once
|
||||
|
||||
#include "library/api.hpp"
|
||||
#include "api.hpp"
|
||||
#include "library/common.hpp"
|
||||
#include "library/concepts.hpp"
|
||||
#include "library/config.hpp"
|
||||
#include "library/defines.hpp"
|
||||
#include "library/state.hpp"
|
||||
#include "library/timemory.hpp"
|
||||
|
||||
#include <timemory/utility/types.hpp>
|
||||
|
||||
#include <array>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <memory>
|
||||
#include <optional>
|
||||
#include <type_traits>
|
||||
|
||||
#if !defined(OMNITRACE_MAX_THREADS)
|
||||
# define OMNITRACE_MAX_THREADS 1024
|
||||
#endif
|
||||
|
||||
namespace omnitrace
|
||||
{
|
||||
ThreadState set_thread_state(ThreadState);
|
||||
|
||||
// bundle of components used in instrumentation
|
||||
using instrumentation_bundle_t =
|
||||
tim::component_bundle<api::omnitrace, comp::wall_clock*, comp::user_global_bundle*>;
|
||||
tim::component_bundle<project::omnitrace, comp::wall_clock*,
|
||||
comp::user_global_bundle*>;
|
||||
|
||||
// allocator for instrumentation_bundle_t
|
||||
using bundle_allocator_t = tim::data::ring_buffer_allocator<instrumentation_bundle_t>;
|
||||
@@ -77,18 +78,56 @@ struct thread_deleter
|
||||
}
|
||||
};
|
||||
|
||||
template <typename Tp>
|
||||
struct generate
|
||||
{
|
||||
using type = Tp;
|
||||
|
||||
template <typename... Args>
|
||||
auto operator()(Args&&... _args) const
|
||||
{
|
||||
if constexpr(concepts::is_unique_pointer<Tp>::value)
|
||||
{
|
||||
using value_type = typename type::element_type;
|
||||
return type{ new value_type{ invoke(std::forward<Args>(_args), 0)... } };
|
||||
}
|
||||
else
|
||||
{
|
||||
return type{ invoke(std::forward<Args>(_args), 0)... };
|
||||
}
|
||||
}
|
||||
|
||||
private:
|
||||
template <typename Up>
|
||||
static auto invoke(Up&& _v, int,
|
||||
std::enable_if_t<std::is_invocable<Up>::value, int> = 0)
|
||||
-> decltype(std::forward<Up>(_v)())
|
||||
{
|
||||
return std::forward<Up>(_v)();
|
||||
}
|
||||
|
||||
template <typename Up>
|
||||
static auto&& invoke(Up&& _v, long)
|
||||
{
|
||||
return std::forward<Up>(_v);
|
||||
}
|
||||
};
|
||||
|
||||
using construct_on_init = std::true_type;
|
||||
|
||||
template <typename Tp, typename Tag = void, size_t MaxThreads = max_supported_threads>
|
||||
struct thread_data
|
||||
{
|
||||
using instance_array_t = std::array<unique_ptr_t<Tp>, MaxThreads>;
|
||||
using value_type = unique_ptr_t<Tp>;
|
||||
using instance_array_t = std::array<value_type, MaxThreads>;
|
||||
using construct_on_init = std::true_type;
|
||||
|
||||
template <typename... Args>
|
||||
static void construct(Args&&...);
|
||||
static unique_ptr_t<Tp>& instance();
|
||||
static value_type& instance();
|
||||
static instance_array_t& instances();
|
||||
template <typename... Args>
|
||||
static unique_ptr_t<Tp>& instance(construct_on_init, Args&&...);
|
||||
static value_type& instance(construct_on_init, Args&&...);
|
||||
template <typename... Args>
|
||||
static instance_array_t& instances(construct_on_init, Args&&...);
|
||||
|
||||
@@ -108,9 +147,9 @@ thread_data<Tp, Tag, MaxThreads>::construct(Args&&... _args)
|
||||
{
|
||||
// construct outside of lambda to prevent data-race
|
||||
static auto& _instances = instances();
|
||||
static thread_local bool _v = [&_args...]() {
|
||||
static thread_local bool _v = [&]() {
|
||||
_instances.at(threading::get_id()) =
|
||||
unique_ptr_t<Tp>{ new Tp(std::forward<Args>(_args)...) };
|
||||
generate<value_type>{}(std::forward<Args>(_args)...);
|
||||
return true;
|
||||
}();
|
||||
(void) _v;
|
||||
@@ -148,7 +187,180 @@ thread_data<Tp, Tag, MaxThreads>::instances(construct_on_init, Args&&... _args)
|
||||
static auto& _v = [&]() -> instance_array_t& {
|
||||
auto& _internal = instances();
|
||||
for(size_t i = 0; i < MaxThreads; ++i)
|
||||
_internal.at(i) = unique_ptr_t<Tp>{ new Tp(std::forward<Args>(_args)...) };
|
||||
_internal.at(i) = generate<value_type>{}(std::forward<Args>(_args)...);
|
||||
return _internal;
|
||||
}();
|
||||
return _v;
|
||||
}
|
||||
|
||||
//--------------------------------------------------------------------------------------//
|
||||
//
|
||||
// thread_data with std::optional
|
||||
//
|
||||
//--------------------------------------------------------------------------------------//
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
struct thread_data<std::optional<Tp>, Tag, MaxThreads>
|
||||
{
|
||||
using value_type = std::optional<Tp>;
|
||||
using instance_array_t = std::array<value_type, MaxThreads>;
|
||||
using construct_on_init = std::true_type;
|
||||
|
||||
template <typename... Args>
|
||||
static void construct(Args&&...);
|
||||
static value_type& instance();
|
||||
static instance_array_t& instances();
|
||||
template <typename... Args>
|
||||
static value_type& instance(construct_on_init, Args&&...);
|
||||
template <typename... Args>
|
||||
static instance_array_t& instances(construct_on_init, Args&&...);
|
||||
|
||||
static constexpr size_t size() { return MaxThreads; }
|
||||
|
||||
decltype(auto) begin() { return instances().begin(); }
|
||||
decltype(auto) end() { return instances().end(); }
|
||||
|
||||
decltype(auto) begin() const { return instances().begin(); }
|
||||
decltype(auto) end() const { return instances().end(); }
|
||||
};
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
template <typename... Args>
|
||||
void
|
||||
thread_data<std::optional<Tp>, Tag, MaxThreads>::construct(Args&&... _args)
|
||||
{
|
||||
// construct outside of lambda to prevent data-race
|
||||
static auto& _instances = instances();
|
||||
static thread_local bool _v = [&]() {
|
||||
_instances.at(threading::get_id()) =
|
||||
generate<value_type>{}(std::forward<Args>(_args)...);
|
||||
return true;
|
||||
}();
|
||||
(void) _v;
|
||||
}
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
std::optional<Tp>&
|
||||
thread_data<std::optional<Tp>, Tag, MaxThreads>::instance()
|
||||
{
|
||||
return instances().at(threading::get_id());
|
||||
}
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
typename thread_data<std::optional<Tp>, Tag, MaxThreads>::instance_array_t&
|
||||
thread_data<std::optional<Tp>, Tag, MaxThreads>::instances()
|
||||
{
|
||||
static auto _v = instance_array_t{};
|
||||
return _v;
|
||||
}
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
template <typename... Args>
|
||||
std::optional<Tp>&
|
||||
thread_data<std::optional<Tp>, Tag, MaxThreads>::instance(construct_on_init,
|
||||
Args&&... _args)
|
||||
{
|
||||
construct(std::forward<Args>(_args)...);
|
||||
return instances().at(threading::get_id());
|
||||
}
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
template <typename... Args>
|
||||
typename thread_data<std::optional<Tp>, Tag, MaxThreads>::instance_array_t&
|
||||
thread_data<std::optional<Tp>, Tag, MaxThreads>::instances(construct_on_init,
|
||||
Args&&... _args)
|
||||
{
|
||||
static auto& _v = [&]() -> instance_array_t& {
|
||||
auto& _internal = instances();
|
||||
for(size_t i = 0; i < MaxThreads; ++i)
|
||||
_internal.at(i) = generate<value_type>{}(std::forward<Args>(_args)...);
|
||||
return _internal;
|
||||
}();
|
||||
return _v;
|
||||
}
|
||||
|
||||
//--------------------------------------------------------------------------------------//
|
||||
//
|
||||
// thread_data with raw data (no pointer)
|
||||
//
|
||||
//--------------------------------------------------------------------------------------//
|
||||
|
||||
using tim::identity;
|
||||
using tim::identity_t;
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
struct thread_data<identity<Tp>, Tag, MaxThreads>
|
||||
{
|
||||
using value_type = Tp;
|
||||
using instance_array_t = std::array<value_type, MaxThreads>;
|
||||
using construct_on_init = std::true_type;
|
||||
|
||||
template <typename... Args>
|
||||
static void construct(Args&&...);
|
||||
static value_type& instance();
|
||||
static instance_array_t& instances();
|
||||
template <typename... Args>
|
||||
static value_type& instance(construct_on_init, Args&&...);
|
||||
template <typename... Args>
|
||||
static instance_array_t& instances(construct_on_init, Args&&...);
|
||||
|
||||
static constexpr size_t size() { return MaxThreads; }
|
||||
|
||||
decltype(auto) begin() { return instances().begin(); }
|
||||
decltype(auto) end() { return instances().end(); }
|
||||
|
||||
decltype(auto) begin() const { return instances().begin(); }
|
||||
decltype(auto) end() const { return instances().end(); }
|
||||
};
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
template <typename... Args>
|
||||
void
|
||||
thread_data<identity<Tp>, Tag, MaxThreads>::construct(Args&&... _args)
|
||||
{
|
||||
// construct outside of lambda to prevent data-race
|
||||
static auto& _instances = instances();
|
||||
static thread_local bool _v = [&]() {
|
||||
_instances.at(threading::get_id()) =
|
||||
generate<value_type>{}(std::forward<Args>(_args)...);
|
||||
return true;
|
||||
}();
|
||||
(void) _v;
|
||||
}
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
Tp&
|
||||
thread_data<identity<Tp>, Tag, MaxThreads>::instance()
|
||||
{
|
||||
return instances().at(threading::get_id());
|
||||
}
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
typename thread_data<identity<Tp>, Tag, MaxThreads>::instance_array_t&
|
||||
thread_data<identity<Tp>, Tag, MaxThreads>::instances()
|
||||
{
|
||||
static auto _v = instance_array_t{};
|
||||
return _v;
|
||||
}
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
template <typename... Args>
|
||||
Tp&
|
||||
thread_data<identity<Tp>, Tag, MaxThreads>::instance(construct_on_init, Args&&... _args)
|
||||
{
|
||||
construct(std::forward<Args>(_args)...);
|
||||
return instances().at(threading::get_id());
|
||||
}
|
||||
|
||||
template <typename Tp, typename Tag, size_t MaxThreads>
|
||||
template <typename... Args>
|
||||
typename thread_data<identity<Tp>, Tag, MaxThreads>::instance_array_t&
|
||||
thread_data<identity<Tp>, Tag, MaxThreads>::instances(construct_on_init, Args&&... _args)
|
||||
{
|
||||
static auto& _v = [&]() -> instance_array_t& {
|
||||
auto& _internal = instances();
|
||||
for(size_t i = 0; i < MaxThreads; ++i)
|
||||
_internal.at(i) = generate<value_type>{}(std::forward<Args>(_args)...);
|
||||
return _internal;
|
||||
}();
|
||||
return _v;
|
||||
|
||||
Reference in New Issue
Block a user