AI/rocm-systems
2
0
Garpu 0
Kode Masalah Tarik Permintaan Actions Packages Projects Rilis Wiki Kegiatan
Files
2ef9dfd002076093ba97dea74fc54870e4e74ae4
rocm-systems/projects/rocprofiler-systems/cmake/BuildSettings.cmake
T
Jonathan R. Madsen 473f452d39 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

[ROCm/rocprofiler-systems commit: 808ea7dfa7]
2022-08-31 01:24:31 -05:00

343 baris
13 KiB
CMake
Mentah Salahkan Riwayat

# include guard
include_guard(DIRECTORY)
# ########################################################################################
#
# Handles the build settings
#
# ########################################################################################
include(GNUInstallDirs)
include(Compilers)
include(FindPackageHandleStandardArgs)
include(MacroUtilities)
omnitrace_add_option(OMNITRACE_BUILD_DEVELOPER
"Extra build flags for development like -Werror" OFF)
omnitrace_add_option(OMNITRACE_BUILD_EXTRA_OPTIMIZATIONS "Extra optimization flags" OFF)
omnitrace_add_option(OMNITRACE_BUILD_LTO "Build with link-time optimization" OFF)
omnitrace_add_option(OMNITRACE_USE_COMPILE_TIMING
"Build with timing metrics for compilation" OFF)
omnitrace_add_option(OMNITRACE_USE_COVERAGE "Build with code-coverage flags" OFF)
omnitrace_add_option(OMNITRACE_USE_SANITIZER
"Build with -fsanitze=\${OMNITRACE_SANITIZER_TYPE}" OFF)
omnitrace_add_option(OMNITRACE_BUILD_STATIC_LIBGCC
"Build with -static-libgcc if possible" OFF)
omnitrace_add_option(OMNITRACE_BUILD_STATIC_LIBSTDCXX
"Build with -static-libstdc++ if possible" OFF)
omnitrace_add_option(OMNITRACE_BUILD_STACK_PROTECTOR "Build with -fstack-protector" ON)
omnitrace_add_interface_library(omnitrace-static-libgcc
"Link to static version of libgcc")
omnitrace_add_interface_library(omnitrace-static-libstdcxx
"Link to static version of libstdc++")
omnitrace_add_interface_library(omnitrace-static-libgcc-optional
"Link to static version of libgcc")
omnitrace_add_interface_library(omnitrace-static-libstdcxx-optional
"Link to static version of libstdc++")
target_compile_definitions(omnitrace-compile-options INTERFACE $<$<CONFIG:DEBUG>:DEBUG>)
set(OMNITRACE_SANITIZER_TYPE
"leak"
CACHE STRING "Sanitizer type")
if(OMNITRACE_USE_SANITIZER)
omnitrace_add_feature(OMNITRACE_SANITIZER_TYPE
"Sanitizer type, e.g. leak, thread, address, memory, etc.")
endif()
if(OMNITRACE_BUILD_CI)
omnitrace_target_compile_definitions(${LIBNAME}-compile-options
INTERFACE OMNITRACE_CI)
endif()
# ----------------------------------------------------------------------------------------#
# dynamic linking and runtime libraries
#
if(CMAKE_DL_LIBS AND NOT "${CMAKE_DL_LIBS}" STREQUAL "dl")
# if cmake provides dl library, use that
set(dl_LIBRARY
${CMAKE_DL_LIBS}
CACHE FILEPATH "dynamic linking system library")
endif()
foreach(_TYPE dl rt dw)
if(NOT ${_TYPE}_LIBRARY)
find_library(${_TYPE}_LIBRARY NAMES ${_TYPE})
endif()
endforeach()
find_package_handle_standard_args(dl-library REQUIRED_VARS dl_LIBRARY)
find_package_handle_standard_args(rt-library REQUIRED_VARS rt_LIBRARY)
# find_package_handle_standard_args(dw-library REQUIRED_VARS dw_LIBRARY)
if(dl_LIBRARY)
target_link_libraries(omnitrace-compile-options INTERFACE ${dl_LIBRARY})
endif()
# ----------------------------------------------------------------------------------------#
# set the compiler flags
#
add_flag_if_avail(
"-W" "-Wall" "-Wno-unknown-pragmas" "-Wno-unused-function" "-Wno-ignored-attributes"
"-Wno-attributes" "-Wno-missing-field-initializers")
if(OMNITRACE_BUILD_DEBUG)
add_flag_if_avail("-g" "-gdwarf-3" "-fno-omit-frame-pointer")
endif()
if(WIN32)
# suggested by MSVC for spectre mitigation in rapidjson implementation
add_cxx_flag_if_avail("/Qspectre")
endif()
if(CMAKE_CXX_COMPILER_IS_CLANG)
add_cxx_flag_if_avail("-Wno-mismatched-tags")
endif()
# ----------------------------------------------------------------------------------------#
# extra flags for debug information in debug or optimized binaries
#
omnitrace_add_interface_library(
omnitrace-compile-debuginfo
"Attempts to set best flags for more expressive profiling information in debug or optimized binaries"
)
add_target_flag_if_avail(omnitrace-compile-debuginfo "-g" "-fno-omit-frame-pointer"
"-fno-optimize-sibling-calls")
if(CMAKE_CUDA_COMPILER_IS_NVIDIA)
add_target_cuda_flag(omnitrace-compile-debuginfo "-lineinfo")
endif()
target_compile_options(
omnitrace-compile-debuginfo
INTERFACE $<$<COMPILE_LANGUAGE:C>:$<$<C_COMPILER_ID:GNU>:-rdynamic>>
$<$<COMPILE_LANGUAGE:CXX>:$<$<CXX_COMPILER_ID:GNU>:-rdynamic>>)
if(NOT APPLE)
target_link_options(omnitrace-compile-debuginfo INTERFACE
$<$<CXX_COMPILER_ID:GNU>:-rdynamic>)
endif()
if(CMAKE_CUDA_COMPILER_IS_NVIDIA)
target_compile_options(
omnitrace-compile-debuginfo
INTERFACE
$<$<COMPILE_LANGUAGE:CUDA>:$<$<CXX_COMPILER_ID:GNU>:-Xcompiler=-rdynamic>>)
endif()
if(dl_LIBRARY)
target_link_libraries(omnitrace-compile-debuginfo INTERFACE ${dl_LIBRARY})
endif()
if(rt_LIBRARY)
target_link_libraries(omnitrace-compile-debuginfo INTERFACE ${rt_LIBRARY})
endif()
# ----------------------------------------------------------------------------------------#
# non-debug optimizations
#
omnitrace_add_interface_library(omnitrace-compile-extra "Extra optimization flags")
if(NOT OMNITRACE_USE_COVERAGE)
add_target_flag_if_avail(
omnitrace-compile-extra "-finline-functions" "-funroll-loops" "-ftree-vectorize"
"-ftree-loop-optimize" "-ftree-loop-vectorize")
endif()
if(NOT "${CMAKE_BUILD_TYPE}" STREQUAL "Debug" AND OMNITRACE_BUILD_EXTRA_OPTIMIZATIONS)
target_link_libraries(omnitrace-compile-options
INTERFACE $<BUILD_INTERFACE:omnitrace-compile-extra>)
add_flag_if_avail(
"-fno-signaling-nans" "-fno-trapping-math" "-fno-signed-zeros"
"-ffinite-math-only" "-fno-math-errno" "-fpredictive-commoning"
"-fvariable-expansion-in-unroller")
# add_flag_if_avail("-freciprocal-math" "-fno-signed-zeros" "-mfast-fp")
endif()
# ----------------------------------------------------------------------------------------#
# debug-safe optimizations
#
add_cxx_flag_if_avail("-faligned-new")
omnitrace_save_variables(FLTO VARIABLES CMAKE_CXX_FLAGS)
set(_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
set(CMAKE_CXX_FLAGS "-flto=thin ${_CXX_FLAGS}")
omnitrace_add_interface_library(omnitrace-lto "Adds link-time-optimization flags")
add_target_flag_if_avail(omnitrace-lto "-flto=thin")
if(NOT cxx_omnitrace_lto_flto_thin)
set(CMAKE_CXX_FLAGS "-flto ${_CXX_FLAGS}")
add_target_flag_if_avail(omnitrace-lto "-flto")
if(NOT cxx_omnitrace_lto_flto)
set(OMNITRACE_BUILD_LTO OFF)
else()
target_link_options(omnitrace-lto INTERFACE -flto)
endif()
add_target_flag_if_avail(omnitrace-lto "-fno-fat-lto-objects")
if(cxx_omnitrace_lto_fno_fat_lto_objects)
target_link_options(omnitrace-lto INTERFACE -fno-fat-lto-objects)
endif()
else()
target_link_options(omnitrace-lto INTERFACE -flto=thin)
endif()
omnitrace_restore_variables(FLTO VARIABLES CMAKE_CXX_FLAGS)
# ----------------------------------------------------------------------------------------#
# print compilation timing reports (Clang compiler)
#
omnitrace_add_interface_library(
omnitrace-compile-timing
"Adds compiler flags which report compilation timing metrics")
if(CMAKE_CXX_COMPILER_IS_CLANG)
add_target_flag_if_avail(omnitrace-compile-timing "-ftime-trace")
if(NOT cxx_omnitrace_compile_timing_ftime_trace)
add_target_flag_if_avail(omnitrace-compile-timing "-ftime-report")
endif()
else()
add_target_flag_if_avail(omnitrace-compile-timing "-ftime-report")
endif()
if(OMNITRACE_USE_COMPILE_TIMING)
target_link_libraries(omnitrace-compile-options INTERFACE omnitrace-compile-timing)
endif()
# ----------------------------------------------------------------------------------------#
# fstack-protector
#
omnitrace_add_interface_library(omnitrace-stack-protector
"Adds stack-protector compiler flags")
add_target_flag_if_avail(omnitrace-stack-protector "-fstack-protector-strong"
"-Wstack-protector")
if(OMNITRACE_BUILD_STACK_PROTECTOR)
target_link_libraries(omnitrace-compile-options INTERFACE omnitrace-stack-protector)
endif()
# ----------------------------------------------------------------------------------------#
# developer build flags
#
omnitrace_add_interface_library(omnitrace-develop-options "Adds developer compiler flags")
if(OMNITRACE_BUILD_DEVELOPER)
add_target_flag_if_avail(omnitrace-develop-options "-Werror" "-Wdouble-promotion"
"-Wshadow" "-Wextra" "-Wpedantic" "-Werror" "/showIncludes")
endif()
# ----------------------------------------------------------------------------------------#
# visibility build flags
#
omnitrace_add_interface_library(omnitrace-default-visibility
"Adds -fvisibility=default compiler flag")
omnitrace_add_interface_library(omnitrace-hidden-visibility
"Adds -fvisibility=hidden compiler flag")
add_target_flag_if_avail(omnitrace-default-visibility "-fvisibility=default")
add_target_flag_if_avail(omnitrace-hidden-visibility "-fvisibility=hidden"
"-fvisibility-inlines-hidden")
# ----------------------------------------------------------------------------------------#
# developer build flags
#
if(dl_LIBRARY)
# This instructs the linker to add all symbols, not only used ones, to the dynamic
# symbol table. This option is needed for some uses of dlopen or to allow obtaining
# backtraces from within a program.
add_flag_if_avail("-rdynamic")
endif()
# ----------------------------------------------------------------------------------------#
# sanitizer
#
set(OMNITRACE_SANITIZER_TYPES
address
memory
thread
leak
undefined
unreachable
null
bounds
alignment)
set_property(CACHE OMNITRACE_SANITIZER_TYPE PROPERTY STRINGS
"${OMNITRACE_SANITIZER_TYPES}")
omnitrace_add_interface_library(omnitrace-sanitizer-compile-options
"Adds compiler flags for sanitizers")
omnitrace_add_interface_library(
omnitrace-sanitizer
"Adds compiler flags to enable ${OMNITRACE_SANITIZER_TYPE} sanitizer (-fsanitizer=${OMNITRACE_SANITIZER_TYPE})"
)
set(COMMON_SANITIZER_FLAGS "-fno-optimize-sibling-calls" "-fno-omit-frame-pointer"
"-fno-inline-functions")
add_target_flag(omnitrace-sanitizer-compile-options ${COMMON_SANITIZER_FLAGS})
foreach(_TYPE ${OMNITRACE_SANITIZER_TYPES})
set(_FLAG "-fsanitize=${_TYPE}")
omnitrace_add_interface_library(
omnitrace-${_TYPE}-sanitizer
"Adds compiler flags to enable ${_TYPE} sanitizer (${_FLAG})")
add_target_flag(omnitrace-${_TYPE}-sanitizer ${_FLAG})
target_link_libraries(omnitrace-${_TYPE}-sanitizer
INTERFACE omnitrace-sanitizer-compile-options)
set_property(TARGET omnitrace-${_TYPE}-sanitizer
PROPERTY INTERFACE_LINK_OPTIONS ${_FLAG} ${COMMON_SANITIZER_FLAGS})
endforeach()
unset(_FLAG)
unset(COMMON_SANITIZER_FLAGS)
if(OMNITRACE_USE_SANITIZER)
foreach(_TYPE ${OMNITRACE_SANITIZER_TYPE})
if(TARGET omnitrace-${_TYPE}-sanitizer)
target_link_libraries(omnitrace-sanitizer
INTERFACE omnitrace-${_TYPE}-sanitizer)
else()
message(
FATAL_ERROR "Error! Target 'omnitrace-${_TYPE}-sanitizer' does not exist!"
)
endif()
endforeach()
else()
set(OMNITRACE_USE_SANITIZER OFF)
endif()
# ----------------------------------------------------------------------------------------#
# static lib flags
#
target_compile_options(
omnitrace-static-libgcc
INTERFACE $<$<COMPILE_LANGUAGE:C>:$<$<C_COMPILER_ID:GNU>:-static-libgcc>>
$<$<COMPILE_LANGUAGE:CXX>:$<$<CXX_COMPILER_ID:GNU>:-static-libgcc>>)
target_link_options(
omnitrace-static-libgcc INTERFACE
$<$<COMPILE_LANGUAGE:C>:$<$<C_COMPILER_ID:GNU,Clang>:-static-libgcc>>
$<$<COMPILE_LANGUAGE:CXX>:$<$<CXX_COMPILER_ID:GNU,Clang>:-static-libgcc>>)
target_compile_options(
omnitrace-static-libstdcxx
INTERFACE $<$<COMPILE_LANGUAGE:CXX>:$<$<CXX_COMPILER_ID:GNU,Clang>:-static-libstdc++>>
)
target_link_options(
omnitrace-static-libstdcxx INTERFACE
$<$<COMPILE_LANGUAGE:CXX>:$<$<CXX_COMPILER_ID:GNU,Clang>:-static-libstdc++>>)
if(OMNITRACE_BUILD_STATIC_LIBGCC)
target_link_libraries(omnitrace-static-libgcc-optional
INTERFACE omnitrace-static-libgcc)
endif()
if(OMNITRACE_BUILD_STATIC_LIBSTDCXX)
target_link_libraries(omnitrace-static-libstdcxx-optional
INTERFACE omnitrace-static-libstdcxx)
endif()
# ----------------------------------------------------------------------------------------#
# user customization
#
get_property(LANGUAGES GLOBAL PROPERTY ENABLED_LANGUAGES)
if(NOT APPLE OR "$ENV{CONDA_PYTHON_EXE}" STREQUAL "")
add_user_flags(omnitrace-compile-options "CXX")
endif()
Reference in New Issue View Git Blame Copy Permalink