Mercurial > hg
view tests/helpers-testrepo.sh @ 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 | 152f1b47e0ad |
children |
line wrap: on
line source
# In most cases, the mercurial repository can be read by the bundled hg, but # that isn't always true because third-party extensions may change the store # format, for example. In which case, the system hg installation is used. # # We want to use the hg version being tested when interacting with the test # repository, and the system hg when interacting with the mercurial source code # repository. # # The mercurial source repository was typically orignally cloned with the # system mercurial installation, and may require extensions or settings from # the system installation. if [ -n "$HGTESTEXTRAEXTENSIONS" ]; then for extension in $HGTESTEXTRAEXTENSIONS; do extraoptions="$extraoptions --config extensions.$extension=!" done fi syshg () { ( syshgenv exec hg "$@" ) } # Revert the environment so that running "hg" runs the system hg # rather than the test hg installation. syshgenv () { . "$HGTEST_RESTOREENV" HGPLAIN=1 export HGPLAIN } # The test-repo is a live hg repository which may have evolution markers # created, e.g. when a ~/.hgrc enabled evolution. # # Tests may be run using a custom HGRCPATH, which do not enable evolution # markers by default. # # If test-repo includes evolution markers, and we do not enable evolution # markers, hg will occasionally complain when it notices them, which disrupts # tests resulting in sporadic failures. # # Since we aren't performing any write operations on the test-repo, there's # no harm in telling hg that we support evolution markers, which is what the # following lines for the hgrc file do: cat >> "$HGRCPATH" << EOF [experimental] evolution = createmarkers EOF # Use the system hg command if the bundled hg can't read the repository with # no warning nor error. if [ -n "`hg id -R "$TESTDIR/.." 2>&1 >/dev/null`" ]; then alias testrepohg=syshg alias testrepohgenv=syshgenv else alias testrepohg="hg $extraoptions" alias testrepohgenv=: fi