Welcome to pytest-benchmark’s documentation!¶

This plugin provides a benchmark fixture. This fixture is a callable object that will benchmark any function passed to it.

Notable features and goals:

Sensible defaults and automatic calibration for microbenchmarks
Good integration with pytest
Comparison and regression tracking
Exhausive statistics
JSON export

Examples:

def something(duration=0.000001):
    """
    Function that needs some serious benchmarking.
    """
    time.sleep(duration)
    # You may return anything you want, like the result of a computation
    return 123

def test_my_stuff(benchmark):
    # benchmark something
    result = benchmark(something)

    # Extra code, to verify that the run completed correctly.
    # Sometimes you may want to check the result, fast functions
    # are no good if they return incorrect results :-)
    assert result == 123

def test_my_stuff_different_arg(benchmark):
    # benchmark something, but add some arguments
    result = benchmark(something, 0.001)
    assert result == 123

Screenshots¶

Normal run:

Compare mode (--benchmark-compare):

Screenshot of py.test summary in compare mode

Histogram (--benchmark-histogram):

Also, it has nice tooltips.

User guide¶

Installation¶

At the command line:

pip install pytest-benchmark

Usage¶

This plugin provides a benchmark fixture. This fixture is a callable object that will benchmark any function passed to it.

Example:

def something(duration=0.000001):
    """
    Function that needs some serious benchmarking.
    """
    time.sleep(duration)
    # You may return anything you want, like the result of a computation
    return 123

def test_my_stuff(benchmark):
    # benchmark something
    result = benchmark(something)

    # Extra code, to verify that the run completed correctly.
    # Sometimes you may want to check the result, fast functions
    # are no good if they return incorrect results :-)
    assert result == 123

You can also pass extra arguments:

def test_my_stuff(benchmark):
    benchmark(time.sleep, 0.02)

Or even keyword arguments:

def test_my_stuff(benchmark):
    benchmark(time.sleep, duration=0.02)

Another pattern seen in the wild, that is not recommended for micro-benchmarks (very fast code) but may be convenient:

def test_my_stuff(benchmark):
    @benchmark
    def something():  # unnecessary function call
        time.sleep(0.000001)

A better way is to just benchmark the final function:

def test_my_stuff(benchmark):
    benchmark(time.sleep, 0.000001)  # way more accurate results!

If you need to do fine control over how the benchmark is run (like a setup function, exact control of iterations and rounds) there’s a special mode - pedantic:

def my_special_setup():
    ...

def test_with_setup(benchmark):
    benchmark.pedantic(something, setup=my_special_setup, args=(1, 2, 3), kwargs={'foo': 'bar'}, iterations=10, rounds=100)

Commandline options¶

py.test command-line options:

--benchmark-min-time=SECONDS

Minimum time per round in seconds. Default: ‘0.000005’

--benchmark-max-time=SECONDS

Maximum run time per test - it will be repeated until this total time is reached. It may be exceeded if test function is very slow or –benchmark-min-rounds is large (it takes precedence). Default: ‘1.0’

--benchmark-min-rounds=NUM

Minimum rounds, even if total time would exceed –max-time. Default: 5

--benchmark-sort=COL

Column to sort on. Can be one of: ‘min’, ‘max’, ‘mean’ or ‘stddev’. Default: ‘min’

--benchmark-group-by=LABEL

How to group tests. Can be one of: ‘group’, ‘name’, ‘fullname’, ‘func’, ‘fullfunc’ or ‘param’. Default: ‘group’

--benchmark-timer=FUNC

Timer to use when measuring time. Default: ‘time.perf_counter’

--benchmark-calibration-precision=NUM

Precision to use when calibrating number of iterations. Precision of 10 will make the timer look 10 times more accurate, at a cost of less precise measure of deviations. Default: 10

--benchmark-warmup=KIND

Activates warmup. Will run the test function up to number of times in the calibration phase. See –benchmark-warmup-iterations. Note: Even the warmup phase obeys –benchmark-max-time. Available KIND: ‘auto’, ‘off’, ‘on’. Default: ‘auto’ (automatically activate on PyPy).

--benchmark-warmup-iterations=NUM

Max number of iterations to run in the warmup phase. Default: 100000

--benchmark-verbose

Dump diagnostic and progress information.

--benchmark-disable-gc

Disable GC during benchmarks.

--benchmark-skip

Skip running any benchmarks.

--benchmark-only

Only run benchmarks.

--benchmark-save=NAME

Save the current run into ‘STORAGE-PATH/counter- NAME.json’. Default: ‘<commitid>_<date>_<time>_<isdirty>’, example: ‘e689af57e7439b9005749d806248897ad550eab5_20150811_041632_uncommitted-changes’.

--benchmark-autosave

Autosave the current run into ‘STORAGE-PATH/<counter>_<commitid>_<date>_<time>_<isdirty>’, example: ‘STORAGE-PATH/0123_525685bcd6a51d1ade0be75e2892e713e02dfd19_20151028_221708_uncommitted-changes.json’

--benchmark-save-data

Use this to make –benchmark-save and –benchmark- autosave include all the timing data, not just the stats.

–benchmark-compare=NUM|ID

Compare the current run against run NUM (or prefix of _id in elasticsearch) or the latest saved run if unspecified.

--benchmark-compare-fail=EXPR

Fail test if performance regresses according to given EXPR (eg: min:5% or mean:0.001 for number of seconds). Can be used multiple times.

--benchmark-storage=STORAGE-PATH

Specify a different path to store the runs (when –benchmark-save or –benchmark-autosave are used). Default: ‘./.benchmarks/<os>-<pyimplementation>-<pyversion>-<arch>bit’, example: ‘Linux-CPython-2.7-64bit’.

--benchmark-cprofile=COLUMN

If specified measure one run with cProfile and stores 10 top functions. Argument is a column to sort by. Available columns: ‘ncallls_recursion’, ‘ncalls’, ‘tottime’, ‘tottime_per’, ‘cumtime’, ‘cumtime_per’, ‘function_name’.

--benchmark-storage=URI

Specify a path to store the runs as uri in form file://path or elasticsearch+http[s]://host1,host2/index/doctype (when –benchmark-save or –benchmark-autosave are used). Default: ‘file://./.benchmarks‘.

--benchmark-histogram=FILENAME-PREFIX

Plot graphs of min/max/avg/stddev over time in FILENAME-PREFIX-test_name.svg. If FILENAME-PREFIX contains slashes (‘/’) then directories will be created. Default: ‘benchmark_<date>_<time>’, example: ‘benchmark_20150811_041632’.

--benchmark-json=PATH

Dump a JSON report into PATH. Note that this will include the complete data (all the timings, not just the stats).

Markers¶

You can set per-test options with the benchmark marker:

@pytest.mark.benchmark(
    group="group-name",
    min_time=0.1,
    max_time=0.5,
    min_rounds=5,
    timer=time.time,
    disable_gc=True,
    warmup=False
)
def test_my_stuff(benchmark):
    @benchmark
    def result():
        # Code to be measured
        return time.sleep(0.000001)

    # Extra code, to verify that the run
    # completed correctly.
    # Note: this code is not measured.
    assert result is None

Extra info¶

You can set arbirary values in the benchmark.extra_info dictionary, which will be saved in the JSON if you use --benchmark-autosave or similar:

def test_my_stuff(benchmark):
    benchmark.extra_info['foo'] = 'bar'
    benchmark(time.sleep, 0.02)

Patch utilities¶

Suppose you want to benchmark an internal function from a class:

class Foo(object):
    def __init__(self, arg=0.01):
        self.arg = arg

    def run(self):
        self.internal(self.arg)

    def internal(self, duration):
        time.sleep(duration)

With the benchmark fixture this is quite hard to test if you don’t control the Foo code or it has very complicated construction.

For this there’s an experimental benchmark_weave fixture that can patch stuff using aspectlib (make sure you pip install aspectlib or pip install pytest-benchmark[aspect]):

def test_foo(benchmark):
    benchmark.weave(Foo.internal, lazy=True):
    f = Foo()
    f.run()

Calibration¶

pytest-benchmark will run your function multiple times between measurements. A round is that set of runs done between measurements. This is quite similar to the builtin timeit module but it’s more robust.

The problem with measuring single runs appears when you have very fast code. To illustrate:

Diagram illustrating issues with measuring very fast code

In other words, a round is a set of runs that are averaged together, those resulting numbers are then used to compute the result tables. The default settings will try to keep the round small enough (so that you get to see variance), but not too small, because then you have the timer calibration issues illustrated above (your test function is faster than or as fast as the resolution of the timer).

By default pytest-benchmark will try to run your function as many times needed to fit a 10 x TIMER_RESOLUTION period. You can fine tune this with the --benchmark-min-time and --benchmark-calibration-precision options.

Pedantic mode¶

pytest-benchmark allows a special mode that doesn’t do any automatic calibration. To make it clear it’s only for people that know exactly what they need it’s called “pedantic”.

def test_with_setup(benchmark):
    benchmark.pedantic(stuff, args=(1, 2, 3), kwargs={'foo': 'bar'}, iterations=10, rounds=100)

Reference¶

benchmark.pedantic(target, args=(), kwargs=None, setup=None, rounds=1, warmup_rounds=0, iterations=1)¶

Parameters:

target (callable) – Function to benchmark.
args (list or tuple) – Positional arguments to the target function.
kwargs (dict) – Named arguments to the target function.
setup (callable) –
A function to call right before calling the target function.

The setup function can also return the arguments for the function (in case you need to create new arguments every time).
```
def test_with_setup(benchmark):
    def setup():
        # can optionally return a (args, kwargs) tuple
        return (1, 2, 3), {'foo': 'bar'}
    benchmark.pedantic(stuff, setup=setup, rounds=100)
```
Note

if you use a setup function then you cannot use the args, kwargs and iterations options.
rounds (int) – Number of rounds to run.
iterations (int) –
Number of iterations.

In the non-pedantic mode (eg: benchmark(stuff, 1, 2, 3, foo='bar')) the iterations is automatically chosen depending on what timer you have. In other words, be careful in what you chose for this option.

The default value (1) is unsafe for benchmarking very fast functions that take under 100μs (100 microseconds).
warmup_rounds (int) –
Number of warmup rounds.

Set to non-zero to enable warmup. Warmup will run with the same number of iterations.

Example: if you have iteration=5, warmup_rounds=10 then your function will be called 50 times.

Comparing past runs¶

Before comparing different runs it’s ideal to make your tests as consistent as possible, see Frequently Asked Questions for more details.

pytest-benchmark has support for storing stats and data for the previous runs.

To store a run just add --benchmark-autosave or --benchmark-save=some-name to your pytest arguments. All the files are saved in a path like .benchmarks/Linux-CPython-3.4-64bit.

--benchmark-autosave saves a file like 0001_c9cca5de6a4c7eb2_20150815_215724.json where:
- 0001 is an automatically incremented id, much like how django migrations have a number.
- c9cca5de6a4c7eb2 is the commit id (if you use Git or Mercurial)
- 20150815_215724 is the current time
You should add --benchmark-autosave to addopts in you pytest configuration so you dont have to specify it all the time.
--benchmark-name=foobar works similarly, but saves a file like 0001_foobar.json. It’s there in case you want to give specific name to the run.

After you have saved your first run you can compare against it with --benchmark-compare=0001. You will get an additional row for each test in the result table, showing the differences.

You can also make the suite fail with --benchmark-compare-fail=<stat>:<num>% or --benchmark-compare-fail=<stat>:<num>. Examples:

--benchmark-compare-fail=min:5% will make the suite fail if Min is 5% slower for any test.
--benchmark-compare-fail=mean:0.001 will make the suite fail if Mean is 0.001 seconds slower for any test.

Plotting¶

Note

To use plotting you need to pip install pygal pygaljs or pip install pytest-benchmark[histogram].

You can also get a nice plot with --benchmark-histogram. The result is a modified Tukey box and wiskers plot where the outliers (the small bullets) are Min and Max.

Example output:

Hooks¶

Hooks for customizing various parts of pytest-benchmark.

pytest_benchmark.hookspec.pytest_benchmark_compare_machine_info(config, benchmarksession, machine_info, compared_benchmark)[source]¶

You may want to use this hook to implement custom checks or abort execution. pytest-benchmark builtin hook does this:

def pytest_benchmark_compare_machine_info(config, benchmarksession, machine_info, compared_benchmark):
    if compared_benchmark["machine_info"] != machine_info:
        benchmarksession.logger.warn(
            "Benchmark machine_info is different. Current: %s VS saved: %s." % (
                format_dict(machine_info),
                format_dict(compared_benchmark["machine_info"]),
            )
    )

pytest_benchmark.hookspec.pytest_benchmark_generate_commit_info(config)[source]¶

To completely replace the generated commit_info do something like this:

def pytest_benchmark_generate_commit_info(config):
    return {'id': subprocess.check_output(['svnversion']).strip()}

pytest_benchmark.hookspec.pytest_benchmark_generate_json(config, benchmarks, include_data, machine_info, commit_info)[source]¶

You should read pytest-benchmark’s code if you really need to wholly customize the json.

Warning

Improperly customizing this may cause breakage if --benchmark-compare or --benchmark-histogram are used.

By default, pytest_benchmark_generate_json strips benchmarks that have errors from the output. To prevent this, implement the hook like this:

@pytest.mark.hookwrapper
def pytest_benchmark_generate_json(config, benchmarks, include_data, machine_info, commit_info):
    for bench in benchmarks:
        bench.has_error = False
    yield

pytest_benchmark.hookspec.pytest_benchmark_generate_machine_info(config)[source]¶

To completely replace the generated machine_info do something like this:

def pytest_benchmark_generate_machine_info(config):
    return {'user': getpass.getuser()}

pytest_benchmark.hookspec.pytest_benchmark_group_stats(config, benchmarks, group_by)[source]¶

You may perform grouping customization here, in case the builtin grouping doesn’t suit you.

Example:

@pytest.mark.hookwrapper
def pytest_benchmark_group_stats(config, benchmarks, group_by):
    outcome = yield
    if group_by == "special":  # when you use --benchmark-group-by=special
        result = defaultdict(list)
        for bench in benchmarks:
            # `bench.special` doesn't exist, replace with whatever you need
            result[bench.special].append(bench)
        outcome.force_result(result.items())

pytest_benchmark.hookspec.pytest_benchmark_update_commit_info(config, commit_info)[source]¶

To add something into the commit_info, like the commit message do something like this:

def pytest_benchmark_update_commit_info(config, commit_info):
    commit_info['message'] = subprocess.check_output(['git', 'log', '-1', '--pretty=%B']).strip()

pytest_benchmark.hookspec.pytest_benchmark_update_json(config, benchmarks, output_json)[source]¶

Use this to add custom fields in the output JSON.

Example:

def pytest_benchmark_update_json(config, benchmarks, output_json):
    output_json['foo'] = 'bar'

pytest_benchmark.hookspec.pytest_benchmark_update_machine_info(config, machine_info)[source]¶

If benchmarks are compared and machine_info is different then warnings will be shown.

To add the current user to the commit info override the hook in your conftest.py like this:

def pytest_benchmark_update_machine_info(config, machine_info):
    machine_info['user'] = getpass.getuser()

Frequently Asked Questions¶

Why is my StdDev so high?

There can be few causes for this:

Bad isolation. You run other services in your machine that eat up your cpu or you run in a VM and that makes machine performance inconsistent. Ideally you’d avoid such setups, stop all services and applications and use bare metal machines.
Bad tests or too much complexity. The function you’re testing is doing I/O, using external resources, has side-effects or doing other non-deterministic things. Ideally you’d avoid testing huge chunks of code.

One special situation is PyPy: it’s GC and JIT can add unpredictable overhead - you’ll see it as huge spikes all over the place. You should make sure that you have a good amount of warmup (using --benchmark-warmup and --benchmark-warmup-iterations) to prime the JIT as much as possible. Unfortunately not much can be done about GC overhead.

If you cannot make your tests more predictable and remove overhead you should look at different stats like: IQR and Median. IQR is often better than StdDev.

My is my Min way lower than Q1-1.5IQR?

You may see this issue in the histogram plot. This is another instance of bad isolation.

For example, Intel CPUs has a feature called Turbo Boost wich overclocks your CPU depending how many cores you have at that time and hot your CPU is. If your CPU is too hot you get no Turbo Boost. If you get Turbo Boost active then the CPU quickly gets hot. You can see how this won’t work for sustained workloads.

When Turbo Boost kicks in you may see “speed spikes” - and you’d get this strange outlier Min.

When you have other programs running on your machine you may also see the “speed spikes” - the other programs idle for a brief moment and that allows your function to run way faster in that brief moment.

I can’t avoid using VMs or running other programs. What can I do?

As a last ditch effort pytest-benchmark allows you to plugin in custom timers (--benchmark-timer). You could use something like time.process_time (Python 3.3+ only) as the timer. Process time doesn’t include sleeping or waiting for I/O.

The histogram doesn’t show Max time. What gives?!

The height of the plot is limited to Q3+1.5IQR because Max has the nasty tendency to be way higher and making everything else small and undiscerning. For this reason Max is plotted outside.

Most people don’t care about Max at all so this is fine.

Glossary¶

Iteration

A single run of your benchmarked function.

Round

A set of iterations. The size of a round is computed in the calibration phase.

Stats are computed with rounds, not with iterations. The duration for a round is an average of all the iterations in that round.

See: Calibration for an explanation of why it’s like this.

Mean

TODO

Median

TODO

IQR

InterQuertile Range. This is a different way to measure variance. Good explanation here

StdDev

TODO: Standard Deviation

Outliers

TODO

Contributing¶

Contributions are welcome, and they are greatly appreciated! Every little bit helps, and credit will always be given.

Bug reports¶

When reporting a bug please include:

Your operating system name and version.

Any details about your local setup that might be helpful in troubleshooting.

Detailed steps to reproduce the bug.

Documentation improvements¶

pytest-benchmark could always use more documentation, whether as part of the official pytest-benchmark docs, in docstrings, or even on the web in blog posts, articles, and such.

Feature requests and feedback¶

The best way to send feedback is to file an issue at https://github.com/ionelmc/pytest-benchmark/issues.

If you are proposing a feature:

Explain in detail how it would work.
Keep the scope as narrow as possible, to make it easier to implement.
Remember that this is a volunteer-driven project, and that code contributions are welcome :)

Development¶

To set up pytest-benchmark for local development:

Fork pytest-benchmark (look for the “Fork” button).

Clone your fork locally:

git clone git@github.com:your_name_here/pytest-benchmark.git

Create a branch for local development:
```
git checkout -b name-of-your-bugfix-or-feature
```
Now you can make your changes locally.
When you’re done making changes, run all the checks, doc builder and spell checker with tox one command:
```
tox
```

Commit your changes and push your branch to GitHub:

git add .
git commit -m "Your detailed description of your changes."
git push origin name-of-your-bugfix-or-feature

Submit a pull request through the GitHub website.

Pull Request Guidelines¶

If you need some code review or feedback while you’re developing the code just make the pull request.

For merging, you should:

Include passing tests (run tox) [1].
Update documentation when there’s new API, functionality etc.
Add a note to CHANGELOG.rst about the changes.
Add yourself to AUTHORS.rst.

[1]	If you don’t have all the necessary python versions available locally you can rely on Travis - it will run the tests for each change you add in the pull request. It will be slower though ...

Tips¶

To run a subset of tests:

tox -e envname -- py.test -k test_myfeature

To run all the test environments in parallel (you need to pip install detox):

detox

Authors¶

Ionel Cristian Mărieș - https://blog.ionelmc.ro
Marc Abramowitz - http://marc-abramowitz.com
Dave Collins - https://github.com/thedavecollins
Stefan Krastanov - http://blog.krastanov.org/
Thomas Waldmann - https://github.com/ThomasWaldmann
Antonio Cuni - http://antocuni.eu/en/
Petr Šebek - https://github.com/Artimi
Swen Kooij - https://github.com/Photonios
“varac” - https://github.com/varac
Andre Bianchi - https://github.com/drebs
Jeremy Dobbins-Bucklad - https://github.com/jad-b
Alexey Popravka - https://github.com/popravich

Changelog¶

3.1.0 (2017-07-21)¶

Added “operations per second” (ops field in Stats) metric – shows the call rate of code being tested. Contributed by Alexey Popravka in #78.
Added a time field in commit_info. Contributed by “varac” in #71.
Added a author_time field in commit_info. Contributed by “varac” in #75.
Fixed the leaking of credentials by masking the URL printed when storing data to elasticsearch.
Added a –benchmark-netrc option to use credentials from a netrc file when storing data to elasticsearch. Both contributed by Andre Bianchi in #73.
Fixed docs on hooks. Contributed by Andre Bianchi in #74.
Remove git and hg as system dependencies when guessing the project name.

3.1.0a2 (2017-03-27)¶

machine_info now contains more detailed information about the CPU, in particular the exact model. Contributed by Antonio Cuni in #61.
Added benchmark.extra_info, which you can use to save arbitrary stuff in the JSON. Contributed by Antonio Cuni in the same PR as above.
Fix support for latest PyGal version (histograms). Contributed by Swen Kooij in #68.
Added support for getting commit_info when not running in the root of the repository. Contributed by Vara Canero in #69 <https://github.com/ionelmc/pytest-benchmark/pull/69> _.
Added short form for --storage/--verbose options in CLI.
Added an alternate pytest-benchmark CLI bin (in addition to py.test-benchmark) to match the madness in pytest.
Fix some issues with –help` in CLI.
Improved git remote parsing (for commit_info in JSON outputs).
Fixed default value for –benchmark-columns`.
Fixed comparison mode (loading was done too late).
Remove the project name from the autosave name. This will get the old brief naming from 3.0 back.

3.1.0a1 (2016-10-29)¶

Added --benchmark-colums command line option. It selects what columns are displayed in the result table. Contributed by Antonio Cuni in #34.
Added support for grouping by specific test parametrization (--benchmark-group-by=param:NAME where NAME is your param name). Contributed by Antonio Cuni in #37.
Added support for name or fullname in --benchmark-sort. Contributed by Antonio Cuni in #37.
Changed signature for pytest_benchmark_generate_json hook to take 2 new arguments: machine_info and commit_info.
Changed –benchmark-histogram` to plot groups instead of name-matching runs.
Changed –benchmark-histogram` to plot exactly what you compared against. Now it’s 1:1 with the compare feature.
Changed –benchmark-compare` to allow globs. You can compare against all the previous runs now.
Changed –benchmark-group-by` to allow multiple values separated by comma. Example: --benchmark-group-by=param:foo,param:bar
Added a command line tool to compare previous data: py.test-benchmark. It has two commands:
- list - Lists all the available files.
- compare - Displays result tables. Takes optional arguments:
  - --sort=COL
  - --group-by=LABEL
  - --columns=LABELS
  - --histogram=[FILENAME-PREFIX]
Added --benchmark-cprofile that profiles last run of benchmarked function. Contributed by Petr Šebek.
Changed --benchmark-storage so it now allows elasticsearch storage. It allows to store data to elasticsearch instead to json files. Contributed by Petr Šebek in #58.

3.0.0 (2015-11-08)¶

Improved --help text for --benchmark-histogram, --benchmark-save and --benchmark-autosave.
Benchmarks that raised exceptions during test now have special highlighting in result table (red background).
Benchmarks that raised exceptions are not included in the saved data anymore (you can still get the old behavior back by implementing pytest_benchmark_generate_json in your conftest.py).
The plugin will use pytest’s warning system for warnings. There are 2 categories: WBENCHMARK-C (compare mode issues) and WBENCHMARK-U (usage issues).
The red warnings are only shown if --benchmark-verbose is used. They still will be always be shown in the pytest-warnings section.
Using the benchmark fixture more than one time is disallowed (will raise exception).
Not using the benchmark fixture (but requiring it) will issue a warning (WBENCHMARK-U1).

3.0.0rc1 (2015-10-25)¶

Changed --benchmark-warmup to take optional value and automatically activate on PyPy (default value is auto). MAY BE BACKWARDS INCOMPATIBLE
Removed the version check in compare mode (previously there was a warning if current version is lower than what’s in the file).

3.0.0b3 (2015-10-22)¶

Changed how comparison is displayed in the result table. Now previous runs are shown as normal runs and names get a special suffix indicating the origin. Eg: “test_foobar (NOW)” or “test_foobar (0123)”.
Fixed sorting in the result table. Now rows are sorted by the sort column, and then by name.
Show the plugin version in the header section.
Moved the display of default options in the header section.

3.0.0b2 (2015-10-17)¶

Add a --benchmark-disable option. It’s automatically activated when xdist is on
When xdist is on or statistics can’t be imported then --benchmark-disable is automatically activated (instead of --benchmark-skip). BACKWARDS INCOMPATIBLE
Replace the deprecated __multicall__ with the new hookwrapper system.
Improved description for --benchmark-max-time.

3.0.0b1 (2015-10-13)¶

Tests are sorted alphabetically in the results table.
Failing to import statistics doesn’t create hard failures anymore. Benchmarks are automatically skipped if import failure occurs. This would happen on Python 3.2 (or earlier Python 3).

3.0.0a4 (2015-10-08)¶

Changed how failures to get commit info are handled: now they are soft failures. Previously it made the whole test suite fail, just because you didn’t have git/hg installed.

3.0.0a3 (2015-10-02)¶

Added progress indication when computing stats.

3.0.0a2 (2015-09-30)¶

Fixed accidental output capturing caused by capturemanager misuse.

3.0.0a1 (2015-09-13)¶

Added JSON report saving (the --benchmark-json command line arguments). Based on initial work from Dave Collins in #8.
Added benchmark data storage(the --benchmark-save and --benchmark-autosave command line arguments).
Added comparison to previous runs (the --benchmark-compare command line argument).
Added performance regression checks (the --benchmark-compare-fail command line argument).
Added possibility to group by various parts of test name (the –benchmark-compare-group-by` command line argument).
Added historical plotting (the --benchmark-histogram command line argument).
Added option to fine tune the calibration (the --benchmark-calibration-precision command line argument and calibration_precision marker option).
Changed benchmark_weave to no longer be a context manager. Cleanup is performed automatically. BACKWARDS INCOMPATIBLE
Added benchmark.weave method (alternative to benchmark_weave fixture).
Added new hooks to allow customization:
- pytest_benchmark_generate_machine_info(config)
- pytest_benchmark_update_machine_info(config, info)
- pytest_benchmark_generate_commit_info(config)
- pytest_benchmark_update_commit_info(config, info)
- pytest_benchmark_group_stats(config, benchmarks, group_by)
- pytest_benchmark_generate_json(config, benchmarks, include_data)
- pytest_benchmark_update_json(config, benchmarks, output_json)
- pytest_benchmark_compare_machine_info(config, benchmarksession, machine_info, compared_benchmark)
Changed the timing code to:
- Tracers are automatically disabled when running the test function (like coverage tracers).
- Fixed an issue with calibration code getting stuck.
Added pedantic mode via benchmark.pedantic(). This mode disables calibration and allows a setup function.

2.5.0 (2015-06-20)¶

Improved test suite a bit (not using cram anymore).
Improved help text on the --benchmark-warmup option.
Made warmup_iterations available as a marker argument (eg: @pytest.mark.benchmark(warmup_iterations=1234)).
Fixed --benchmark-verbose‘s printouts to work properly with output capturing.
Changed how warmup iterations are computed (now number of total iterations is used, instead of just the rounds).
Fixed a bug where calibration would run forever.
Disabled red/green coloring (it was kinda random) when there’s a single test in the results table.

2.4.1 (2015-03-16)¶

Fix regression, plugin was raising ValueError: no option named 'dist' when xdist wasn’t installed.

2.4.0 (2015-03-12)¶

Add a benchmark_weave experimental fixture.
Fix internal failures when xdist plugin is active.
Automatically disable benchmarks if xdist is active.

2.3.0 (2014-12-27)¶

Moved the warmup in the calibration phase. Solves issues with benchmarking on PyPy.

Added a --benchmark-warmup-iterations option to fine-tune that.

2.2.0 (2014-12-26)¶

Make the default rounds smaller (so that variance is more accurate).
Show the defaults in the --help section.

2.1.0 (2014-12-20)¶

Simplify the calibration code so that the round is smaller.
Add diagnostic output for calibration code (--benchmark-verbose).

2.0.0 (2014-12-19)¶

Replace the context-manager based API with a simple callback interface. BACKWARDS INCOMPATIBLE
Implement timer calibration for precise measurements.

1.0.0 (2014-12-15)¶

Use a precise default timer for PyPy.

? (?)¶

Readme and styling fixes. Contributed by Marc Abramowitz in #4.
Lots of wild changes.

`--benchmark-min-time=SECONDS`
	Minimum time per round in seconds. Default: ‘0.000005’
`--benchmark-max-time=SECONDS`
	Maximum run time per test - it will be repeated until this total time is reached. It may be exceeded if test function is very slow or –benchmark-min-rounds is large (it takes precedence). Default: ‘1.0’
`--benchmark-min-rounds=NUM`
	Minimum rounds, even if total time would exceed –max-time. Default: 5
`--benchmark-sort=COL`
	Column to sort on. Can be one of: ‘min’, ‘max’, ‘mean’ or ‘stddev’. Default: ‘min’
`--benchmark-group-by=LABEL`
	How to group tests. Can be one of: ‘group’, ‘name’, ‘fullname’, ‘func’, ‘fullfunc’ or ‘param’. Default: ‘group’
`--benchmark-timer=FUNC`
	Timer to use when measuring time. Default: ‘time.perf_counter’
`--benchmark-calibration-precision=NUM`
	Precision to use when calibrating number of iterations. Precision of 10 will make the timer look 10 times more accurate, at a cost of less precise measure of deviations. Default: 10
`--benchmark-warmup=KIND`
	Activates warmup. Will run the test function up to number of times in the calibration phase. See –benchmark-warmup-iterations. Note: Even the warmup phase obeys –benchmark-max-time. Available KIND: ‘auto’, ‘off’, ‘on’. Default: ‘auto’ (automatically activate on PyPy).
`--benchmark-warmup-iterations=NUM`
	Max number of iterations to run in the warmup phase. Default: 100000
`--benchmark-verbose`
	Dump diagnostic and progress information.
`--benchmark-disable-gc`
	Disable GC during benchmarks.
`--benchmark-skip`
	Skip running any benchmarks.
`--benchmark-only`
	Only run benchmarks.
`--benchmark-save=NAME`
	Save the current run into ‘STORAGE-PATH/counter- NAME.json’. Default: ‘<commitid>_<date>_<time>_<isdirty>’, example: ‘e689af57e7439b9005749d806248897ad550eab5_20150811_041632_uncommitted-changes’.
`--benchmark-autosave`
	Autosave the current run into ‘STORAGE-PATH/<counter>_<commitid>_<date>_<time>_<isdirty>’, example: ‘STORAGE-PATH/0123_525685bcd6a51d1ade0be75e2892e713e02dfd19_20151028_221708_uncommitted-changes.json’
`--benchmark-save-data`
	Use this to make –benchmark-save and –benchmark- autosave include all the timing data, not just the stats.

`--benchmark-compare-fail=EXPR`
	Fail test if performance regresses according to given EXPR (eg: min:5% or mean:0.001 for number of seconds). Can be used multiple times.
`--benchmark-storage=STORAGE-PATH`
	Specify a different path to store the runs (when –benchmark-save or –benchmark-autosave are used). Default: ‘./.benchmarks/<os>-<pyimplementation>-<pyversion>-<arch>bit’, example: ‘Linux-CPython-2.7-64bit’.
`--benchmark-cprofile=COLUMN`
	If specified measure one run with cProfile and stores 10 top functions. Argument is a column to sort by. Available columns: ‘ncallls_recursion’, ‘ncalls’, ‘tottime’, ‘tottime_per’, ‘cumtime’, ‘cumtime_per’, ‘function_name’.
`--benchmark-storage=URI`
	Specify a path to store the runs as uri in form file://path or elasticsearch+http[s]://host1,host2/index/doctype (when –benchmark-save or –benchmark-autosave are used). Default: ‘file://./.benchmarks‘.
`--benchmark-histogram=FILENAME-PREFIX`
	Plot graphs of min/max/avg/stddev over time in FILENAME-PREFIX-test_name.svg. If FILENAME-PREFIX contains slashes (‘/’) then directories will be created. Default: ‘benchmark_<date>_<time>’, example: ‘benchmark_20150811_041632’.
`--benchmark-json=PATH`
	Dump a JSON report into PATH. Note that this will include the complete data (all the timings, not just the stats).

Navigation

Welcome to pytest-benchmark’s documentation!¶

Screenshots¶

User guide¶

Installation¶

Usage¶

Commandline options¶

Markers¶

Extra info¶

Patch utilities¶

Calibration¶

Pedantic mode¶

Reference¶

Comparing past runs¶

Plotting¶

Hooks¶

Frequently Asked Questions¶

Glossary¶

Contributing¶

Bug reports¶

Documentation improvements¶

Feature requests and feedback¶

Development¶

Pull Request Guidelines¶

Tips¶

Authors¶

Changelog¶

3.1.0 (2017-07-21)¶

3.1.0a2 (2017-03-27)¶

3.1.0a1 (2016-10-29)¶

3.0.0 (2015-11-08)¶

3.0.0rc1 (2015-10-25)¶

3.0.0b3 (2015-10-22)¶

3.0.0b2 (2015-10-17)¶

3.0.0b1 (2015-10-13)¶

3.0.0a4 (2015-10-08)¶

3.0.0a3 (2015-10-02)¶

3.0.0a2 (2015-09-30)¶

3.0.0a1 (2015-09-13)¶

2.5.0 (2015-06-20)¶

2.4.1 (2015-03-16)¶

2.4.0 (2015-03-12)¶

2.3.0 (2014-12-27)¶

2.2.0 (2014-12-26)¶

2.1.0 (2014-12-20)¶

2.0.0 (2014-12-19)¶

1.0.0 (2014-12-15)¶

? (?)¶

Indices and tables¶

Navigation