* bump boost.json and see if it works in simdjson CI
* enable boost json
* clean up
* add boost json to deps
* use boost if std::string_view is available
* add build with c++20
* use docker image which has the proper libc++ installed
* Bump minimum CMake version
* Remove unnecessary git checks
* Move benchmark options where they are used
* Declare helper functions for dependencies
The custom solution here is tailored for fast configure times, but only
works for dependencies on Github.
* Import dependencies using the declared commands
* Remove git submodules
* Call target_link_libraries properly
target_link_libraries must not be called without a requirement
specifier.
* Fix includes for competition
Co-authored-by: friendlyanon <friendlyanon@users.noreply.github.com>
* Updating main branch for legacy libc++ support
* Adopting
* Removing unnecessary math header.
* Updating the single-header files so we can pass the new tests.
* Portable infinite-value detection is hard.
* Working toward disabling boost json selectively.
* Selectively disabling Boost JSON
* More work toward selectively disabling boost json.
Introduce cmake option SIMDJSON_DISABLE_DEPRECATED_API (default Off)
which turns off deprecated simdjson api functions by setting the macro
SIMDJSON_DISABLE_DEPRECATED_API.
For non-cmake users, users will have to set SIMDJSON_DISABLE_DEPRECATED_API
by some other means to disable the api.
Closes#1264
* Initial PPC64 support
* Add travis CI
* Fix outdated cmake version for travis
* Fix indendtation
* Try another workaround for outdated cmake in travis
* Try beta cmake
* Add dash before beta
* Use builtin snaps
* Use cmake as rocksdb
* Test cmake on bionic
* Remove unnecessary things from travis
* Remove unnecessary things from travis
* Another try of compiler install
* Add all major compilers
* Add all major compilers
* Add all major compilers
* Tweak travis a bit
* Typo
* More robust travis
* Typos typos typos
* Add fewer compilers, add non specific build for clang and gcc, should be the final config
* CMAKE_FLAGS is in incorrect place
* Remove default implementation
* Limit build thread number
* Fall back prefix_xor to a usual implementation, no performance boost is noticed
* Test for power9 as it is the main architecture for OpenPOWER right now
* Add to documentation to build with power9 as the implementation is compatible but compiler optimizations is not
* Replace ARM with PPC in the comment
* Adding a distinct user id benchmark
* reenabling everything
* Removing an unnecessary "value()".
* Better tests of the examples and some fixes.
* Guarding exception code.
* This would allow users to find out what builtin is.
* Trying another approach.
* Added instructions.
* Cleaning up the printout.
* Let us be less invasive.
* Adding a comment.
* initial try at adding boost json to the benchmark
* clean up
* qualify memcpy etc. with std::
* clang format
* extra space
* update benchmark with help from Vinnie Falco from Boost.json
* add missing separators
- Allow user to specify SIMDJSON_BUILTIN_IMPLEMENTATION
- Make cmake -DSIMDJSON_IMPLEMENTATION=haswell *only* specify haswell
- Move negative implementation selection to
-DSIMDJSON_EXCLUDE_IMPLEMENTATION
- Automatically select SIMDJSON_BUILTIN_IMPLEMENTATION if
SIMDJSON_IMPLEMENTATION is set
- Move implementation enablement mostly to implementation files
- Make implementation enablement and selection simpler and more robust
- Fix bug where programs linked against simdjson were not passed
SIMDJSON_XXX_IMPLEMENTATION or SIMDJSON_EXCEPTIONS
* Make it possible to check that an implementation is supported at runtime.
* add CI fuzzing on arm 64 bit
This adds fuzzing on drone.io arm64
For some reason, leak detection had to be disabled. If it is enabled, the fuzzer falsely reports a crash at the end of fuzzing.
Closes: #1188
* Guarding the implementation accesses.
* Better doc.
* Updating cxxopts.
* Make it possible to check that an implementation is supported at runtime.
* Guarding the implementation accesses.
* Better doc.
* Updating cxxopts.
* We need to accomodate cxxopts
Co-authored-by: Paul Dreik <github@pauldreik.se>
* Adding new files.
* Better.
* Fixing minifier and adding tests.
* Adding benchmarks.
* Including the array header.
* Replacing old stream-based code by the new code.
* Doubling up the itoa.
* Hidden away to_chars in internal namespace.
* Removing the repetitions.
* Documented the atoi functions.
* Tuning the escape sequences.
* Moving the operators off the main namespace.
* Added more tests.
* Tweaking the implementation so that it works with and without exp.
* The string_builder template and mini_formatter class
are not part of our public API and are subject to change
at any time!
* Adding a benchmark and some optimization.
* Cleaning.
* Strictly speaking, this header is needed.
* This would disable bash scripts under FreeBSD.
* Let us also disable GIT.
* Let us try to just disable GIT
* Nope. We must have both bash and git disabled.
In the parse_many function, we have one thread doing the stage 1, while the main thread does stage 2. So if stage 1 and stage 2 take half the time, the parse_many could run at twice the speed. It is unlikely to do so. Still, we see benefits of about 40% due to threading.
To achieve this interleaving, we load the data in batches (blocks) of some size. In the current code (master), we create a new thread for each batch. Thread creation is expensive so our approach only works over sizeable batches. This PR improves things and makes parse_many faster when using small batches.
This fixes our parse_stream benchmark which is just busted.
This replaces the one-thread per batch routine by a worker object that reuses the same thread. In benchmarks, this allows us to get the same maximal speed, but with smaller processing blocks. It does not help much with larger blocks because the cost of the thread create gets amortized efficiently.
This PR makes parse_many beneficial over small datasets. It also makes us less dependent on the thread creation time.
Unfortunately, it is going to be difficult to say anything definitive in general. The cost of creating a thread varies widely depending on the OS. On some systems, it might be cheap, in others very expensive. It should be expected that the new code will depend less drastically on the performances of the underlying system, since we create juste one thread.
Co-authored-by: John Keiser <john@johnkeiser.com>
Co-authored-by: Daniel Lemire <lemire@gmai.com>
John Keiser
Paul Dreik
friendlyanon
Daniel Lemire
Danila Kutenin
Daniel Lemire