wt/json
1
0 Fork 0
mirror of https://github.com/nlohmann/json.git synced 2026-10-21 11:29:53 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
43e97b629bae6cd8614815cac8feabdf883ffaf8
json/tests/benchmarks/README.md
T
Niels Lohmann 122d718330 Add json_document and json_view: node index, parser, and document (#5620) 
* Add json_document and json_view: node index, parser, and document

Add json_document and json_view, a read-only, zero-copy index of a
JSON text, as the first public slice of the zero-copy view (#5295).

A parse produces a flat array of 16-byte nodes in document order,
one per value and one per object key. Strings stay in the source
text; escaped strings are decoded into an arena. Integers are
converted while their digits are in the cache; floats keep only
their digit layout and are converted on read. Containers store the
size of their subtree, so a reader can step over one in constant
time. A document makes a handful of allocations, however many
values it has.

The parser accepts exactly what json::parse accepts, with every
combination of ignore_comments and ignore_trailing_commas, with and
without a trailing NUL, and under JSON_STRICT_NUL_HANDLING. It is
portable C++11 and does not depend on byte order.

basic_json_document adds parse, parse_copy, accept, read (reuses a
document's memory), root, is_discarded, source, owns_source,
node_count, memory_usage, and shrink_to_fit. basic_json_view adds
type, the is_* queries, operator bool, size, empty, materialize,
and source_offset. A parse error throws the same exception
basic_json::parse would throw for the same input, message and
position included.

detail::abi_config keeps JSON_STRICT_NUL_HANDLING and
JSON_USE_LEGACY_DISCARDED_VALUE_COMPARISON readable after json.hpp
undefines them, in the ABI namespace so they always match the
basic_json in use.

A NUL byte that ends a // comment is the end of the input, as in
parse() since #5696.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Fix document_data for old Clang (3.4-3.6)

Drop the empty braced NSDMIs of the std::string members: old Clang
rejects the defaulted constructor when it is used by a member
initializer before the end of the class definition.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Exclude libstdc++'s string_view comparison from the sign-change check

basic_string_view::_S_compare stores the difference of two lengths in a
signed difference_type on purpose, which -fsanitize=integer reports for
every comparison with a shorter view (test-json_view_cpp17).

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Reject integer lengths in json_document::parse

parse(ptr, len) compiled: len converted to allow_exceptions, and ptr was
read as a C string, past the end of a buffer without a terminating NUL.
Delete the overloads of parse, parse_copy, accept, and read that take an
integer other than bool where the flags are expected.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Copy const rvalue strings and document the accepted inputs

json_document::parse(std::move(const_string)) failed to compile with
"no matching read_kind": only a non-const rvalue std::string can be moved
from. Treat a const rvalue as a copied byte container.

parse() does not accept everything BasicJsonType::parse() does: a FILE*
and pointers to or arrays of wide characters are rejected at compile time.
List the supported inputs instead.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Remove the conversion to bool from json_view

explicit operator bool meant "refers to a value", which silently differs
from what a basic_json converts to. Use !v.is_discarded() instead.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Delete json_document::root() on temporary documents

auto v = json_document::parse(text).root() compiled and left the view
dangling. Delete the overload for rvalue documents, take a named document
in the tests, and document the lifetime rule.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Store node integers by halves on big-endian targets

The non-little-endian path of the node writer wrote the integer's native
word over len and next, so len got the high half there. Compose and split
the value explicitly (len is the low half, next the high half); the
little-endian path stays a plain memcpy.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Check the node array size for overflow

reserve(n) and the growth of the index computed n * sizeof(node) without a
check, which wraps around on 32-bit targets for inputs of about 1 GiB and
allocates a too small array. Throw std::bad_alloc for a count beyond the
address space and clamp the growth step to it.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* State the exact input size limit of json_document

The check rejects inputs of 0xFFFFFFF0 bytes or more, but the exception
message and the documentation said 4 GiB. Name the limit once
(max_input_size), and state 4 GiB minus 16 bytes in the message and the
documentation. Test the limit with a container that only claims the size.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Remove the unused legacy_discarded_value_comparison from abi_config

Nothing reads it: the view does not compare values.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Align the banner comment of json_view.hpp

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Fuzz json_document with exact-size buffers and parse options

The fuzzer only parsed a std::string with default options. Also parse an
exact-size byte vector, which has no NUL after its last byte and takes the
bounds-checked path, and derive ignore_comments and ignore_trailing_commas
from the first input byte, comparing against json::parse and json::accept
with the same options.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Give document_data a user-provided constructor

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Fix clang-tidy findings in unit-json_view_builder.cpp

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Annotate intentional patterns in unit-json_view.cpp for clang-tidy

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Suppress Flawfinder's read() finding on the deleted overload

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Check the steps of the deep-nesting test separately

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Make max_input_size a function (GCC -Wunused-const-variable)

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Make json_view.hpp pass include-what-you-use

ci_single_binaries runs IWYU with --error on every header. For json_view.hpp
it suggested adding <array> (std::array is used), the headers that json.hpp
already provides (abi_config, abi_macros, input_adapters, json_pointer,
cpp_future, string_concat, value_t, json_fwd), and <version> for
std::nullptr_t, and removing <cstddef>.

- include <array>
- keep <cstddef> (nullptr_t, size_t; IWYU attributes them to <version> and <cstring>)
- tell IWYU not to suggest the headers that json.hpp provides: the amalgamated
  json_view.hpp only includes json.hpp, so including them here would duplicate
  their definitions
- export json.hpp, keep macro_unscope.hpp, and drop the unused forward declaration of the document class

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Read the node array base after emit() in the view builder's open()

emit() moves the node array when it grows, and the subtraction read base
in the same expression, so the order was unspecified. MSVC Release builds
without forced inlining read the old base; the container index then
pointed outside the array and close() wrote out of bounds.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Silence a clang-tidy use-after-move finding in the json_view test

const_text is const, so std::move does not move from it; the second call
is part of the const rvalue test.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Regenerate the amalgamated headers

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Address the review of the view builder and document

Changes promised in the replies to the review of #5620:

- read() reuses the owned copy: a copied input (an rvalue byte container,
  a stream, FILE*, wide characters, an iterator range) is appended to the
  cleared owned buffer instead of a fresh std::string being move-assigned
  over it. A moved-in std::string still replaces it.
- The error replay runs the library parser with a SAX handler that stores
  nothing and throws the reported exception, instead of
  BasicJsonType::parse(), which built a DOM up to the error.
- Pointers to and arrays of wide characters, and FILE*, take the input
  adapter path (converted to UTF-8 into an owned buffer) instead of
  failing a static_assert. Wide pointers are NUL-terminated.
- Streams are read in blocks through rdbuf() into the buffer, instead of
  character by character through the input adapter; eofbit is set at the
  end, as by the library's stream adapter.
- tape/tape_size/tape_cap are now nodes/node_count/node_capacity.
- The builder's members get descriptive, m_-prefixed names.
- The shallow frame stack is a std::array (in a wrapper with a
  user-provided constructor: -Weffc++ wants the member initialized, and
  value-initializing the 64 frames costs up to 30 % on small documents).
- The value dispatch macro has no NEXT parameter; the switch is left with
  break, and the macro is undefined after its last use.
- literal() and digits_exceed() deduce the length from the string literal;
  each literal has one helper; the two `if (!is_float)` blocks are merged.
- arena_ensure() sits right before the writes; the decode arena is a
  std::vector<char>.
- The builder records the deepest nesting, and materialize() sizes its
  stack from it once (growing it only if a subtree is deeper).
- Comment fixes (hex4, slow_string, the NUL rule, put_utf8, 0 - m, the
  arena resize, grow()) and formatting (macro backslashes, digit cases,
  one power of ten per line, value_t assertion next to is_container()).
- build_with's template parameters follow the builder's order.

Greg's idea of keeping the source offset in `off` and the arena offset in
`next` for escaped strings was measured and not adopted: string access got
7-18 % slower with a select, and 3.5-18 % slower with a branch-free field
pick (Apple M-series, clang -O3), parsing was unchanged.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Apply the review's wording fixes to the json_view documentation

- Copying is disabled because of its cost, not because of two documents
  borrowing the same buffer; the unique_ptr follows from move-only.
- Inputs: streams, FILE*, and wide characters (strings, arrays, pointers)
  are all supported and owned; point to parse_copy() for owned copies.
- The error replay only runs for failing input, builds no values, and is
  skipped with allow_exceptions = false.
- read() writes copied inputs into the previous owned copy.
- Views can dangle like std::string_view, and why there is no
  shared/weak pointer design.
- The 4 GiB limit (16-byte nodes, 16-byte scanner steps), streams being
  read to the end first, which macros apply, why materialize() has no
  source positions, and wording nits (accept() vs basic_json::accept(),
  is_binary, is_discarded, see-also lists, size table, "back into").

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Say that json_view converts integers while parsing

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Address the review: null inline nodes, trait spelling, and max() in json_view

- inline_nodes is null when the document has no inline nodes, instead of
  pointing past the end of the allocation; shrink_to_fit() no longer
  passes it to memcpy() for an empty node array
- is_byte_char and is_wide_char use conjunction/disjunction, which
  astyle leaves unspaced
- append_stream() uses (std::max) for the count read

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Include materialize.hpp in the view builder test

The test calls detail::view::materialize(), which only the single header
pulled in; the multiple-header build (JSON_TEST_USING_MULTIPLE_HEADERS)
failed with "'materialize' is not a member of detail::view".

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Rename emit() in the view builder for Qt compatibility

Qt defines emit as an empty macro, so a Qt header included before
json_view.hpp broke the declaration of the builder's emit(), see
issue #5865. It is now called push_node().

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Do not copy an empty run into an unallocated arena

memcpy() must not be passed a null pointer even for zero bytes, which
arena_run() did for an empty run while the arena was still empty
(ci_test_clang_sanitizer: "null pointer passed as argument 1" in
builder.hpp). Only the slow path can see such a run.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Exclude the view builder test from Infer's reports

Infer's std::unique_ptr model confuses the values that the test's build()
helper returns by value and reports USE_AFTER_DELETE for one of them after
the destructor of an unrelated temporary (ci_infer, 12 reports, all in
tests/src/unit-json_view_builder.cpp). AddressSanitizer is clean on the
same tests.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

* Skip the deleted-overload checks of json_view on MSVC 2015

MSVC 2015 does not treat selecting a deleted function in decltype as a
substitution failure, so is_detected reports the deleted
parse/parse_copy/accept/read(ptr, len) overloads as callable and the
static_asserts in "integer arguments do not compile" fail (AppVeyor
VS2015 jobs). Calling them still fails to compile there. Guard the
negative checks as unit-delete_deprecated_functions.cpp does.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

---------

Signed-off-by: Niels Lohmann <mail@nlohmann.me>
2026-10-21 10:21:27 +02:00

146 lines
8.0 KiB
Markdown
Raw Blame History

# Benchmarks
Micro-benchmarks for parsing, serialization and the binary formats, written with
[Google Benchmark](https://github.com/google/benchmark). They are not run by CI; see
[When to run them](#when-to-run-them).
## What is measured
| benchmark | what it does |
|---|---|
| `ParseFile`, `ParseString` | parse JSON from a file stream or a string |
| `Accept` | validate JSON from a string (`json::accept`) |
| `ParseIndented` | parse the large files re-indented by 4 spaces, for the lexer's whitespace handling |
| `Dump` | serialize, compact (`-`) and indented (`4`) |
| `ToCbor`, `BinaryToCbor` | write CBOR; `BinaryToCbor` writes binary values of growing size |
| `FromMsgpack` | read MessagePack; unchanged over the years, so its numbers stay comparable across releases |
| `FromBinaryBuffer`, `FromBinaryFile` | read CBOR, MessagePack, UBJSON, BJData and BSON from a buffer or a `FILE*` |
| `FromBinaryShape` | read deeply nested, container-heavy and scalar-heavy documents in every binary format |
| `FromCborChunkedString` | read CBOR strings split into indefinite-length chunks |
| `ViewParse`, `ViewRead` | parse with [`json_document`](https://json.nlohmann.me/features/json_view/) into a new document, or into one that is reused; compare with `ParseString` |
| `ViewParseIndented` | as `ParseIndented`, with a reused `json_document` |
| `ViewAccept` | validate with `json_document::accept`; compare with `Accept` |
| `ViewMaterialize` | convert a parsed `json_document` into a `json` value |
The input files are those of [nativejson-benchmark](https://github.com/miloyip/nativejson-benchmark) (`canada`,
`citm_catalog`, `twitter`), a large `jeopardy` file, and number-heavy files (`floats`, `signed_ints`, ...).
`bytes_per_second` counts the bytes read or written: the JSON text when parsing, the output when serializing.
## Requirements
- CMake 3.14 or later, a C++11 compiler, and Ninja for the `make` target.
- Network access on the first configure: CMake downloads Google Benchmark and the
[test data](https://github.com/nlohmann/json_test_data) into the build directory. To reuse a download of the test
data, pass `-DJSON_TestDataDirectory=<build directory>/test_files`.
- Google Benchmark is fetched as a fixed release with a checksum, so that every build uses the same
code and a download cannot change under you. It is updated like any other dependency when a new release comes out:
change `JSON_GOOGLE_BENCHMARK_VERSION` and the archive's `URL_HASH` in `CMakeLists.txt` together. Results are kept
comparable by building both sides of a comparison with the same version, not by keeping that version forever (see
[Comparing two versions](#comparing-two-versions)).
- The benchmarks include `single_include/nlohmann/json.hpp`, so run `make amalgamate` after changing anything in
`include/`.
GCC and Clang builds use `-O3 -flto -DNDEBUG`.
## Running them
From the repository root, this builds everything from scratch in `cmake-build-benchmarks` and runs all benchmarks:
```sh
make run_benchmarks
```
To build once and run selectively:
```sh
cmake -S tests/benchmarks -B build-benchmarks -G Ninja -DCMAKE_BUILD_TYPE=Release
cmake --build build-benchmarks
build-benchmarks/json_benchmarks --benchmark_filter='ParseString|Dump'
```
Useful options of `json_benchmarks`:
| option | effect |
|---|---|
| `--benchmark_list_tests` | list the benchmarks instead of running them |
| `--benchmark_filter=<regex>` | run only the benchmarks whose names match |
| `--benchmark_repetitions=<n>` | run every benchmark `n` times and add mean, median, standard deviation and coefficient of variation |
| `--benchmark_enable_random_interleaving=true` | run the repetitions in random order, which spreads out drifts such as thermal throttling |
| `--benchmark_min_time=<seconds>s` | run each benchmark at least this long (e.g. `2s`) |
| `--benchmark_out=<file> --benchmark_out_format=json` | also write the results to a file, e.g. for `compare.py` |
## Reading the output
Each line shows the wall-clock `Time` and the `CPU` time per iteration, the number of `Iterations` Google Benchmark
chose, and the throughput in `bytes_per_second`. With repetitions, the lines ending in `_median` are the ones to
compare. A `_cv` (coefficient of variation) above a few percent means the machine was too noisy for small
differences to mean anything.
## Comparing two versions
To see what a change or a release did, build the same benchmarks twice: once against the header of the version to
compare with, and once against the current one. `JSON_BENCHMARK_INCLUDE_DIR` names the directory holding the
`nlohmann/json.hpp` to benchmark. For example, to compare the current checkout with 3.12.0:
```sh
# the header of the version to compare with
mkdir -p build-baseline-header/nlohmann
git show v3.12.0:single_include/nlohmann/json.hpp > build-baseline-header/nlohmann/json.hpp
# the same benchmarks, built against either header
cmake -S tests/benchmarks -B build-baseline -G Ninja -DCMAKE_BUILD_TYPE=Release \
-DJSON_BENCHMARK_INCLUDE_DIR="$PWD/build-baseline-header"
cmake -S tests/benchmarks -B build-current -G Ninja -DCMAKE_BUILD_TYPE=Release
cmake --build build-baseline
cmake --build build-current
# run both, back to back
build-baseline/json_benchmarks --benchmark_repetitions=10 --benchmark_enable_random_interleaving=true \
--benchmark_out=build-baseline/results.json --benchmark_out_format=json
build-current/json_benchmarks --benchmark_repetitions=10 --benchmark_enable_random_interleaving=true \
--benchmark_out=build-current/results.json --benchmark_out_format=json
```
Google Benchmark ships a tool to compare the two result files. It needs NumPy and SciPy:
```sh
python3 -m venv build-venv
build-venv/bin/pip install numpy scipy
build-venv/bin/python build-current/_deps/benchmark-src/tools/compare.py -a benchmarks build-baseline/results.json build-current/results.json
```
The tool's own `tools/requirements.txt` pins NumPy and SciPy versions that need Python 3.11 or later; with an older
Python, unpinned versions work as well. In its output:
- the `Time` and `CPU` columns are relative changes: `-0.35` means 35% faster, `+0.10` means 10% slower;
- `_pvalue` lines report a Mann-Whitney U test of whether the two versions differ. It needs at least 9
repetitions, and a p-value below 0.05 means the difference is unlikely to be noise;
- `OVERALL_GEOMEAN` summarizes all benchmarks;
- `-a` shows only the aggregates, not every repetition.
Both builds use the same Google Benchmark, the one `CMakeLists.txt` names, so only the header differs. For the same
reason, do not compare against result files from an earlier checkout: a different Google Benchmark, compiler or
machine state can shift the numbers as much as the change under test. Rerun the baseline instead.
The header you compare with must support everything the benchmarks use. The current benchmarks build against 3.12.0;
the `View*` benchmarks (in `src/benchmarks_view.cpp`) are only built if the directory also holds
`nlohmann/json_view.hpp`.
Only benchmarks present in both result files are compared, so for older releases, either filter the benchmarks or
build that release's own `tests/benchmarks` against its own header.
## Getting stable numbers
- Build and run both versions on the same machine with the same compiler and Google Benchmark, one right after the
other.
- Keep the machine otherwise idle: no builds, no browser, and a laptop plugged in.
- On Linux, set the CPU frequency governor to `performance`, e.g. `sudo cpupower frequency-set --governor performance`.
Google Benchmark prints a warning when frequency scaling is enabled. Pinning the process to a core
(`taskset -c 2 ...`) helps as well.
- Use 10 or more repetitions with random interleaving, compare medians, and treat changes within the `_cv` as noise.
## When to run them
They are a manual step, not part of CI: shared CI runners vary more between runs than most of the effects measured.
Run the comparison above before a release, comparing the previous release tag with `develop`, and for pull requests
that claim to change performance.
Reference in new issue View Git Blame Copy Permalink