Mercurial > hg
view tests/fsmonitor-run-tests.py @ 38732:be4984261611
merge: mark file gets as not thread safe (issue5933)
In default installs, this has the effect of disabling the thread-based
worker on Windows when manifesting files in the working directory. My
measurements have shown that with revlog-based repositories, Mercurial
spends a lot of CPU time in revlog code resolving file data. This ends
up incurring a lot of context switching across threads and slows down
`hg update` operations when going from an empty working directory to
the tip of the repo.
On mozilla-unified (246,351 files) on an i7-6700K (4+4 CPUs):
before: 487s wall
after: 360s wall (equivalent to worker.enabled=false)
cpus=2: 379s wall
Even with only 2 threads, the thread pool is still slower.
The introduction of the thread-based worker (02b36e860e0b) states that
it resulted in a "~50%" speedup for `hg sparse --enable-profile` and
`hg sparse --disable-profile`. This disagrees with my measurement
above. I theorize a few reasons for this:
1) Removal of files from the working directory is I/O - not CPU - bound
and should benefit from a thread pool (unless I/O is insanely fast
and the GIL release is near instantaneous). So tests like `hg sparse
--enable-profile` may exercise deletion throughput and aren't good
benchmarks for worker tasks that are CPU heavy.
2) The patch was authored by someone at Facebook. The results were
likely measured against a repository using remotefilelog. And I
believe that revision retrieval during working directory updates with
remotefilelog will often use a remote store, thus being I/O and not
CPU bound. This probably resulted in an overstated performance gain.
Since there appears to be a need to enable the thread-based worker with
some stores, I've made the flagging of file gets as thread safe
configurable. I've made it experimental because I don't want to formalize
a boolean flag for this option and because this attribute is best
captured against the store implementation. But we don't have a proper
store API for this yet. I'd rather cross this bridge later.
It is possible there are revlog-based repositories that do benefit from
a thread-based worker. I didn't do very comprehensive testing. If there
are, we may want to devise a more proper algorithm for whether to use
the thread-based worker, including possibly config options to limit the
number of threads to use. But until I see evidence that justifies
complexity, simplicity wins.
Differential Revision: https://phab.mercurial-scm.org/D3963
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Wed, 18 Jul 2018 09:49:34 -0700 |
parents | efd6e941e933 |
children | b7ba1cfba174 |
line wrap: on
line source
#!/usr/bin/env python # fsmonitor-run-tests.py - Run Mercurial tests with fsmonitor enabled # # Copyright 2017 Facebook, Inc. # # This software may be used and distributed according to the terms of the # GNU General Public License version 2 or any later version. # # This is a wrapper around run-tests.py that spins up an isolated instance of # Watchman and runs the Mercurial tests against it. This ensures that the global # version of Watchman isn't affected by anything this test does. from __future__ import absolute_import from __future__ import print_function import argparse import contextlib import json import os import shutil import subprocess import sys import tempfile import uuid osenvironb = getattr(os, 'environb', os.environ) if sys.version_info > (3, 5, 0): PYTHON3 = True xrange = range # we use xrange in one place, and we'd rather not use range def _bytespath(p): return p.encode('utf-8') elif sys.version_info >= (3, 0, 0): print('%s is only supported on Python 3.5+ and 2.7, not %s' % (sys.argv[0], '.'.join(str(v) for v in sys.version_info[:3]))) sys.exit(70) # EX_SOFTWARE from `man 3 sysexit` else: PYTHON3 = False # In python 2.x, path operations are generally done using # bytestrings by default, so we don't have to do any extra # fiddling there. We define the wrapper functions anyway just to # help keep code consistent between platforms. def _bytespath(p): return p def getparser(): """Obtain the argument parser used by the CLI.""" parser = argparse.ArgumentParser( description='Run tests with fsmonitor enabled.', epilog='Unrecognized options are passed to run-tests.py.') # - keep these sorted # - none of these options should conflict with any in run-tests.py parser.add_argument('--keep-fsmonitor-tmpdir', action='store_true', help='keep temporary directory with fsmonitor state') parser.add_argument('--watchman', help='location of watchman binary (default: watchman in PATH)', default='watchman') return parser @contextlib.contextmanager def watchman(args): basedir = tempfile.mkdtemp(prefix='hg-fsmonitor') try: # Much of this configuration is borrowed from Watchman's test harness. cfgfile = os.path.join(basedir, 'config.json') # TODO: allow setting a config with open(cfgfile, 'w') as f: f.write(json.dumps({})) logfile = os.path.join(basedir, 'log') clilogfile = os.path.join(basedir, 'cli-log') if os.name == 'nt': sockfile = '\\\\.\\pipe\\watchman-test-%s' % uuid.uuid4().hex else: sockfile = os.path.join(basedir, 'sock') pidfile = os.path.join(basedir, 'pid') statefile = os.path.join(basedir, 'state') argv = [ args.watchman, '--sockname', sockfile, '--logfile', logfile, '--pidfile', pidfile, '--statefile', statefile, '--foreground', '--log-level=2', # debug logging for watchman ] envb = osenvironb.copy() envb[b'WATCHMAN_CONFIG_FILE'] = _bytespath(cfgfile) with open(clilogfile, 'wb') as f: proc = subprocess.Popen( argv, env=envb, stdin=None, stdout=f, stderr=f) try: yield sockfile finally: proc.terminate() proc.kill() finally: if args.keep_fsmonitor_tmpdir: print('fsmonitor dir available at %s' % basedir) else: shutil.rmtree(basedir, ignore_errors=True) def run(): parser = getparser() args, runtestsargv = parser.parse_known_args() with watchman(args) as sockfile: osenvironb[b'WATCHMAN_SOCK'] = _bytespath(sockfile) # Indicate to hghave that we're running with fsmonitor enabled. osenvironb[b'HGFSMONITOR_TESTS'] = b'1' runtestdir = os.path.dirname(__file__) runtests = os.path.join(runtestdir, 'run-tests.py') blacklist = os.path.join(runtestdir, 'blacklists', 'fsmonitor') runtestsargv.insert(0, runtests) runtestsargv.extend([ '--extra-config', 'extensions.fsmonitor=', '--blacklist', blacklist, ]) return subprocess.call(runtestsargv) if __name__ == '__main__': sys.exit(run())