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
Tá an tiomantas seo le fáil i:
Jonathan R. Madsen
2022-08-31 01:24:31 -05:00
tiomanta ag GitHub
tuismitheoir a1dcd1bc4b
tiomantas 808ea7dfa7
D'athraigh 113 comhad le 4284 breiseanna agus 2880 scriosta
+12
Féach ar an gComhad
@@ -2,6 +2,14 @@ cmake_minimum_required(VERSION 3.16 FATAL_ERROR)
project(omnitrace-examples LANGUAGES C CXX)
if("${CMAKE_BUILD_TYPE}" STREQUAL "")
set(CMAKE_BUILD_TYPE
"RelWithDebInfo"
CACHE STRING "Build type" FORCE)
endif()
string(TOUPPER "${CMAKE_BUILD_TYPE}" BUILD_TYPE)
set(CMAKE_VISIBILITY_INLINES_HIDDEN OFF)
set(CMAKE_CXX_VISIBILITY_PRESET "default")
set(CMAKE_CXX_STANDARD 17)
@@ -9,6 +17,10 @@ set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CXX_CLANG_TIDY)
set(CMAKE_INSTALL_DEFAULT_COMPONENT_NAME examples)
if(OMNITRACE_BUILD_DEBUG)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -g -gdwarf-3 -fno-omit-frame-pointer")
endif()
option(BUILD_SHARED_LIBS "Build dynamic libraries" ON)
if(CMAKE_PROJECT_NAME STREQUAL "omnitrace")
+1 -1
Féach ar an gComhad
@@ -10,7 +10,7 @@ option(LULESH_BUILD_KOKKOS "Build Kokkos from submodule" ON)
if(LULESH_BUILD_KOKKOS)
add_subdirectory(external)
if(LULESH_USE_CUDA)
kokkos_compilation(PROJECT)
kokkos_compilation(PROJECT COMPILER ${Kokkos_NVCC_WRAPPER})
elseif(LULESH_USE_HIP AND NOT "${CMAKE_CXX_COMPILER}" MATCHES "hipcc")
if(NOT HIPCC_EXECUTABLE)
find_package(hip QUIET HINTS ${ROCmVersion_DIR} PATHS ${ROCmVersion_DIR})
+10 -8
Féach ar an gComhad
@@ -158,8 +158,8 @@ run(MPI_Comm _comm, int nitr)
MPI_Comm_rank(_comm, &_rank);
MPI_Comm_size(_comm, &_size);
printf("[%s][%i] running %i iterations on %i ranks...\n", _name.c_str(), _rank, nitr,
_size);
printf("[%s][%i][%s] running %i iterations on %i ranks... \n", _name.c_str(), _rank,
__FUNCTION__, nitr, _size);
MPI_Barrier(_comm);
for(int i = 0; i < nitr; ++i)
@@ -175,6 +175,9 @@ run(MPI_Comm _comm, int nitr)
all2all<double, 6>(_rank, _comm);
}
MPI_Barrier(_comm);
printf("[%s][%i][%s] running %i iterations on %i ranks... Done\n", _name.c_str(),
_rank, __FUNCTION__, nitr, _size);
}
void
@@ -235,8 +238,6 @@ run_main(int argc, char** argv)
print_info(MPI_COMM_WORLD, true, "MPI_COMM_WORLD");
printf("[%s]\n", _name.c_str());
if(size > 1)
{
MPI_Comm dup;
@@ -313,6 +314,8 @@ run_main(int argc, char** argv)
print_info(dup, false);
}
printf("[%s][%i of %i] %s... Done", _name.c_str(), rank, size, __FUNCTION__);
}
int
@@ -325,13 +328,12 @@ main(int argc, char** argv)
auto _prom = std::promise<void>{};
auto _fut = _prom.get_future();
std::thread _thr{ [&]() {
std::thread{ [&]() {
_prom.set_value_at_thread_exit();
run_main(argc, argv);
_prom.set_value();
} };
} }.join();
_fut.wait();
_thr.join();
MPI_Finalize();
return EXIT_SUCCESS;
+9
Féach ar an gComhad
@@ -14,6 +14,9 @@ add_executable(openmp-lu ${CMAKE_CURRENT_SOURCE_DIR}/LU/lu.cpp
if(CMAKE_CXX_COMPILER_ID MATCHES "Clang")
find_package(OpenMP REQUIRED)
target_link_libraries(openmp-common PUBLIC OpenMP::OpenMP_CXX)
set(OMNITRACE_OPENMP_USING_LIBOMP_LIBRARY
ON
CACHE INTERNAL "Used by omnitrace testing" FORCE)
else()
find_program(CLANGXX_EXECUTABLE NAMES clang++)
find_library(
@@ -27,9 +30,15 @@ else()
omnitrace_custom_compilation(COMPILER ${CLANGXX_EXECUTABLE} TARGET openmp-common)
omnitrace_custom_compilation(COMPILER ${CLANGXX_EXECUTABLE} TARGET openmp-cg)
omnitrace_custom_compilation(COMPILER ${CLANGXX_EXECUTABLE} TARGET openmp-lu)
set(OMNITRACE_OPENMP_USING_LIBOMP_LIBRARY
ON
CACHE INTERNAL "Used by omnitrace testing" FORCE)
else()
find_package(OpenMP REQUIRED)
target_link_libraries(openmp-common PUBLIC OpenMP::OpenMP_CXX)
set(OMNITRACE_OPENMP_USING_LIBOMP_LIBRARY
OFF
CACHE INTERNAL "Used by omnitrace testing" FORCE)
endif()
endif()
+6 -2
Féach ar an gComhad
@@ -14,11 +14,13 @@
#if USE_LOCKS > 0
# include <mutex>
using auto_lock_t = std::unique_lock<std::mutex>;
long total = 0;
using auto_lock_t = std::unique_lock<std::mutex>;
long total = 0;
long lock_count = 0;
std::mutex mtx{};
#else
std::atomic<long> total{ 0 };
long lock_count = 0;
#endif
long
@@ -52,6 +54,7 @@ run(size_t nitr, long n)
auto _v = fib(_get_n());
auto_lock_t _lk{ mtx };
total += _v;
++lock_count;
}
#else
long local = 0;
@@ -110,6 +113,7 @@ main(int argc, char** argv)
printf("[%s] fibonacci(%li) x %lu = %li\n", _name.c_str(), nfib, nthread,
static_cast<long>(total));
printf("[%s] number of mutex locks = %li\n", _name.c_str(), lock_count);
return 0;
}
+13 -10
Féach ar an gComhad
@@ -2,19 +2,22 @@ cmake_minimum_required(VERSION 3.16 FATAL_ERROR)
project(omnitrace-python)
set(PYTHON_FILES builtin.py external.py source.py noprofile.py)
set(PYTHON_FILES builtin.py external.py source.py noprofile.py fill.py)
if(OMNITRACE_INSTALL_EXAMPLES)
find_package(Python3 COMPONENTS Interpreter)
if(Python3_FOUND)
set(PYTHON_EXECUTABLE "${Python3_EXECUTABLE}")
foreach(_FILE ${PYTHON_FILES})
configure_file(${PROJECT_SOURCE_DIR}/${_FILE} ${PROJECT_BINARY_DIR}/${_FILE}
@ONLY)
find_package(Python3 COMPONENTS Interpreter)
if(Python3_FOUND)
set(PYTHON_EXECUTABLE "${Python3_EXECUTABLE}")
foreach(_FILE ${PYTHON_FILES})
configure_file(${PROJECT_SOURCE_DIR}/${_FILE} ${PROJECT_BINARY_DIR}/${_FILE}
@ONLY)
if(OMNITRACE_INSTALL_EXAMPLES)
install(
PROGRAMS ${PROJECT_BINARY_DIR}/${_FILE}
DESTINATION bin
COMPONENT omnitrace-examples)
endforeach()
endif()
endif()
endforeach()
endif()
+39
Féach ar an gComhad
@@ -0,0 +1,39 @@
#!@PYTHON_EXECUTABLE@
import os
import sys
import time
import omnitrace
from omnitrace.user import region as omni_user_region
from omnitrace.profiler import config as omni_config
_prefix = ""
def loop(n):
pass
@omnitrace.profile()
def run(i, n, v):
for l in range(n * n):
loop(v + l)
return v + (n * n)
if __name__ == "__main__":
import argparse
parser = argparse.ArgumentParser()
parser.add_argument("-n", "--num-iterations", help="Number", type=int, default=100)
parser.add_argument("-v", "--value", help="Starting value", type=int, default=10)
args = parser.parse_args()
omni_config.include_args = True
_prefix = os.path.basename(__file__)
print(f"[{_prefix}] Executing {args.num_iterations} iterations...\n")
ans = 0
for i in range(args.num_iterations):
beg = ans
ans = run(i, args.value, beg)
print(f"[{_prefix}] [{i}] result of run({args.value}, {beg}) = {ans}")
-5
Féach ar an gComhad
@@ -1,5 +0,0 @@
[tool.black]
line-length = 80
target-version = ['py38']
include = '\.py'