Mercurial > hg
view tests/svn/startrev.svndump @ 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 | 90d8dfb481e7 |
children |
line wrap: on
line source
SVN-fs-dump-format-version: 2 UUID: c731c652-65e9-4325-a17e-fed96a319f22 Revision-number: 0 Prop-content-length: 56 Content-length: 56 K 8 svn:date V 27 2008-12-06T13:44:21.642421Z PROPS-END Revision-number: 1 Prop-content-length: 112 Content-length: 112 K 7 svn:log V 10 init projA K 10 svn:author V 7 pmezard K 8 svn:date V 27 2008-12-06T13:44:21.759281Z PROPS-END Node-path: branches Node-kind: dir Node-action: add Prop-content-length: 10 Content-length: 10 PROPS-END Node-path: tags Node-kind: dir Node-action: add Prop-content-length: 10 Content-length: 10 PROPS-END Node-path: trunk Node-kind: dir Node-action: add Prop-content-length: 10 Content-length: 10 PROPS-END Revision-number: 2 Prop-content-length: 109 Content-length: 109 K 7 svn:log V 8 createab K 10 svn:author V 7 pmezard K 8 svn:date V 27 2008-12-06T13:44:22.179257Z PROPS-END Node-path: trunk/a Node-kind: file Node-action: add Prop-content-length: 10 Text-content-length: 2 Text-content-md5: 60b725f10c9c85c70d97880dfe8191b3 Content-length: 12 PROPS-END a Node-path: trunk/b Node-kind: file Node-action: add Prop-content-length: 10 Text-content-length: 2 Text-content-md5: 3b5d5c3712955042212316173ccf37be Content-length: 12 PROPS-END b Revision-number: 3 Prop-content-length: 108 Content-length: 108 K 7 svn:log V 7 removeb K 10 svn:author V 7 pmezard K 8 svn:date V 27 2008-12-06T13:44:23.176546Z PROPS-END Node-path: trunk/b Node-action: delete Revision-number: 4 Prop-content-length: 109 Content-length: 109 K 7 svn:log V 8 changeaa K 10 svn:author V 7 pmezard K 8 svn:date V 27 2008-12-06T13:44:25.147151Z PROPS-END Node-path: trunk/a Node-kind: file Node-action: change Text-content-length: 4 Text-content-md5: 0d227f1abf8c2932d342e9b99cc957eb Content-length: 4 a a Revision-number: 5 Prop-content-length: 119 Content-length: 119 K 7 svn:log V 17 branch, changeaaa K 10 svn:author V 7 pmezard K 8 svn:date V 27 2008-12-06T13:44:28.158475Z PROPS-END Node-path: branches/branch1 Node-kind: dir Node-action: add Node-copyfrom-rev: 4 Node-copyfrom-path: trunk Prop-content-length: 34 Content-length: 34 K 13 svn:mergeinfo V 0 PROPS-END Node-path: branches/branch1/a Node-kind: file Node-action: change Text-content-length: 6 Text-content-md5: 7d4ebf8f298d22fc349a91725b00af1c Content-length: 6 a a a Revision-number: 6 Prop-content-length: 117 Content-length: 117 K 7 svn:log V 15 addc,changeaaaa K 10 svn:author V 7 pmezard K 8 svn:date V 27 2008-12-06T13:44:29.180655Z PROPS-END Node-path: branches/branch1/a Node-kind: file Node-action: change Text-content-length: 8 Text-content-md5: d12178e74d8774e34361e0a08d1fd2b7 Content-length: 8 a a a a Node-path: branches/branch1/c Node-kind: file Node-action: add Prop-content-length: 10 Text-content-length: 2 Text-content-md5: 2cd6ee2c70b0bde53fbe6cac3c8b8bb1 Content-length: 12 PROPS-END c