Skip to content

benchmark: measure Regex time and heap allocations #13670

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
bryancall wants to merge 10 commits into
base: master
Choose a base branch
from
+759 −0
Open

benchmark: measure Regex time and heap allocations #13670

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
572b5a3
benchmark: measure Regex time and allocations
bryancall Sep 12, 2026
f70992e
benchmark: do not let a wrapper see a half-resolved allocator table
bryancall Sep 12, 2026
da6b2fb
benchmark: declare the interposed allocators noexcept and size the bo…
bryancall Sep 12, 2026
6f1309b
benchmark: name the JIT stack case honestly and stop timing a static'…
bryancall Sep 12, 2026
b4de4a8
benchmark: bound the bootstrap allocator, and ask whether the JIT is …
bryancall Sep 12, 2026
09886e9
benchmark: count the DFA cases, probe the JIT stack bound, fail reall…
bryancall Sep 12, 2026
9351514
benchmark: treat realloc(nullptr, size) as malloc(size)
bryancall Sep 12, 2026
58dcc88
benchmark: give the miss case a subject that actually misses
bryancall Sep 12, 2026
b3c273d
benchmark: keep the allocator interposer out of the timed binary
bryancall Sep 12, 2026
2669b2b
benchmark: build the timed cases only into the timing target
bryancall Sep 12, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 19 additions & 0 deletions tools/benchmark/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,3 +57,22 @@ target_include_directories(benchmark_HuffmanDecode PRIVATE ${CMAKE_SOURCE_DIR}/l

add_executable(benchmark_ascii_tolower benchmark_ascii_tolower.cc)
target_link_libraries(benchmark_ascii_tolower PRIVATE Catch2::Catch2WithMain ts::tscore)

# Timing. No allocator interposition is compiled in, so a timed case measures the regex
# path alone; a wrapper call costs about 1.5 ns, which would be a fifth of an operation as
# short as copying a compiled pattern.
add_executable(benchmark_Regex benchmark_Regex.cc)
target_link_libraries(benchmark_Regex PRIVATE Catch2::Catch2WithMain ts::tsutil)

# Allocation counting, from the same source. Nothing here is timed, so the interposer's
# own cost does not matter.
add_executable(benchmark_Regex_alloc benchmark_Regex.cc)
target_link_libraries(benchmark_Regex_alloc PRIVATE Catch2::Catch2WithMain ts::tsutil)
target_compile_definitions(benchmark_Regex_alloc PRIVATE BENCHMARK_REGEX_ALLOC)
if(CMAKE_SYSTEM_NAME STREQUAL "Linux")
# The counters interpose the system allocator, which needs the real symbols from the next
# object in the search order. RTLD_NEXT is a GNU extension; g++ and clang++ already define
# _GNU_SOURCE for C++ on Linux, but say so rather than relying on it.
target_link_libraries(benchmark_Regex_alloc PRIVATE ${CMAKE_DL_LIBS})
target_compile_definitions(benchmark_Regex_alloc PRIVATE _GNU_SOURCE)
endif()
Loading