Mercurial > hg
view tests/test-fileset-generated.t @ 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 | fe0667cc521e |
children | cb70501d8b71 |
line wrap: on
line source
$ hg init Set up history and working copy $ $PYTHON $TESTDIR/generate-working-copy-states.py state 2 1 $ hg addremove -q --similarity 0 $ hg commit -m first $ $PYTHON $TESTDIR/generate-working-copy-states.py state 2 2 $ hg addremove -q --similarity 0 $ hg commit -m second $ $PYTHON $TESTDIR/generate-working-copy-states.py state 2 wc $ hg addremove -q --similarity 0 $ hg forget *_*_*-untracked $ rm *_*_missing-* Test status $ hg st -A 'set:modified()' M content1_content1_content3-tracked M content1_content2_content1-tracked M content1_content2_content3-tracked M missing_content2_content3-tracked $ hg st -A 'set:added()' A content1_missing_content1-tracked A content1_missing_content3-tracked A missing_missing_content3-tracked $ hg st -A 'set:removed()' R content1_content1_content1-untracked R content1_content1_content3-untracked R content1_content1_missing-untracked R content1_content2_content1-untracked R content1_content2_content2-untracked R content1_content2_content3-untracked R content1_content2_missing-untracked R missing_content2_content2-untracked R missing_content2_content3-untracked R missing_content2_missing-untracked $ hg st -A 'set:deleted()' ! content1_content1_missing-tracked ! content1_content2_missing-tracked ! content1_missing_missing-tracked ! missing_content2_missing-tracked ! missing_missing_missing-tracked $ hg st -A 'set:missing()' ! content1_content1_missing-tracked ! content1_content2_missing-tracked ! content1_missing_missing-tracked ! missing_content2_missing-tracked ! missing_missing_missing-tracked $ hg st -A 'set:unknown()' ? content1_missing_content1-untracked ? content1_missing_content3-untracked ? missing_missing_content3-untracked $ hg st -A 'set:clean()' C content1_content1_content1-tracked C content1_content2_content2-tracked C missing_content2_content2-tracked Test log $ hg log -T '{rev}\n' --stat 'set:modified()' 1 content1_content2_content1-tracked | 2 +- content1_content2_content3-tracked | 2 +- missing_content2_content3-tracked | 1 + 3 files changed, 3 insertions(+), 2 deletions(-) 0 content1_content1_content3-tracked | 1 + content1_content2_content1-tracked | 1 + content1_content2_content3-tracked | 1 + 3 files changed, 3 insertions(+), 0 deletions(-) Largefiles doesn't crash $ hg log -T '{rev}\n' --stat 'set:modified()' --config extensions.largefiles= The fsmonitor extension is incompatible with the largefiles extension and has been disabled. (fsmonitor !) 1 content1_content2_content1-tracked | 2 +- content1_content2_content3-tracked | 2 +- missing_content2_content3-tracked | 1 + 3 files changed, 3 insertions(+), 2 deletions(-) 0 content1_content1_content3-tracked | 1 + content1_content2_content1-tracked | 1 + content1_content2_content3-tracked | 1 + 3 files changed, 3 insertions(+), 0 deletions(-) $ hg log -T '{rev}\n' --stat 'set:added()' 1 content1_missing_content1-tracked | 1 - content1_missing_content3-tracked | 1 - 2 files changed, 0 insertions(+), 2 deletions(-) 0 content1_missing_content1-tracked | 1 + content1_missing_content3-tracked | 1 + 2 files changed, 2 insertions(+), 0 deletions(-) $ hg log -T '{rev}\n' --stat 'set:removed()' 1 content1_content2_content1-untracked | 2 +- content1_content2_content2-untracked | 2 +- content1_content2_content3-untracked | 2 +- content1_content2_missing-untracked | 2 +- missing_content2_content2-untracked | 1 + missing_content2_content3-untracked | 1 + missing_content2_missing-untracked | 1 + 7 files changed, 7 insertions(+), 4 deletions(-) 0 content1_content1_content1-untracked | 1 + content1_content1_content3-untracked | 1 + content1_content1_missing-untracked | 1 + content1_content2_content1-untracked | 1 + content1_content2_content2-untracked | 1 + content1_content2_content3-untracked | 1 + content1_content2_missing-untracked | 1 + 7 files changed, 7 insertions(+), 0 deletions(-) $ hg log -T '{rev}\n' --stat 'set:deleted()' 1 content1_content2_missing-tracked | 2 +- content1_missing_missing-tracked | 1 - missing_content2_missing-tracked | 1 + 3 files changed, 2 insertions(+), 2 deletions(-) 0 content1_content1_missing-tracked | 1 + content1_content2_missing-tracked | 1 + content1_missing_missing-tracked | 1 + 3 files changed, 3 insertions(+), 0 deletions(-) $ hg log -T '{rev}\n' --stat 'set:unknown()' 1 content1_missing_content1-untracked | 1 - content1_missing_content3-untracked | 1 - 2 files changed, 0 insertions(+), 2 deletions(-) 0 content1_missing_content1-untracked | 1 + content1_missing_content3-untracked | 1 + 2 files changed, 2 insertions(+), 0 deletions(-) $ hg log -T '{rev}\n' --stat 'set:clean()' 1 content1_content2_content2-tracked | 2 +- missing_content2_content2-tracked | 1 + 2 files changed, 2 insertions(+), 1 deletions(-) 0 content1_content1_content1-tracked | 1 + content1_content2_content2-tracked | 1 + 2 files changed, 2 insertions(+), 0 deletions(-) Test revert $ hg revert 'set:modified()' reverting content1_content1_content3-tracked reverting content1_content2_content1-tracked reverting content1_content2_content3-tracked reverting missing_content2_content3-tracked $ hg revert 'set:added()' forgetting content1_missing_content1-tracked forgetting content1_missing_content3-tracked forgetting missing_missing_content3-tracked $ hg revert 'set:removed()' undeleting content1_content1_content1-untracked undeleting content1_content1_content3-untracked undeleting content1_content1_missing-untracked undeleting content1_content2_content1-untracked undeleting content1_content2_content2-untracked undeleting content1_content2_content3-untracked undeleting content1_content2_missing-untracked undeleting missing_content2_content2-untracked undeleting missing_content2_content3-untracked undeleting missing_content2_missing-untracked $ hg revert 'set:deleted()' reverting content1_content1_missing-tracked reverting content1_content2_missing-tracked forgetting content1_missing_missing-tracked reverting missing_content2_missing-tracked forgetting missing_missing_missing-tracked $ hg revert 'set:unknown()' $ hg revert 'set:clean()'