Mercurial > hg
view tests/test-pull-branch.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 | d0abd7949ea3 |
children |
line wrap: on
line source
$ hg init t $ cd t $ echo 1 > foo $ hg ci -Am1 # 0 adding foo $ hg branch branchA marked working directory as branch branchA (branches are permanent and global, did you want a bookmark?) $ echo a1 > foo $ hg ci -ma1 # 1 $ cd .. $ hg init tt $ cd tt $ hg pull ../t pulling from ../t requesting all changes adding changesets adding manifests adding file changes added 2 changesets with 2 changes to 1 files new changesets 495a0ec48aaf:50e089d141b7 (run 'hg update' to get a working copy) $ hg up branchA 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ cd ../t $ echo a2 > foo $ hg ci -ma2 # 2 Create branch B: $ hg up 0 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ hg branch branchB marked working directory as branch branchB $ echo b1 > foo $ hg ci -mb1 # 3 $ cd ../tt A new branch is there $ hg pull -u ../t pulling from ../t searching for changes adding changesets adding manifests adding file changes added 2 changesets with 2 changes to 1 files (+1 heads) new changesets 9f878dea0b96:5be59ce5067b 1 files updated, 0 files merged, 0 files removed, 0 files unresolved Develop both branches: $ cd ../t $ hg up branchA 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo a3 > foo $ hg ci -ma3 # 4 $ hg up branchB 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo b2 > foo $ hg ci -mb2 # 5 $ cd ../tt Should succeed, no new heads: $ hg pull -u ../t pulling from ../t searching for changes adding changesets adding manifests adding file changes added 2 changesets with 2 changes to 1 files new changesets 7c8fe7e20c32:453e93fa00a5 1 files updated, 0 files merged, 0 files removed, 0 files unresolved Add a head on other branch: $ cd ../t $ hg up branchA 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo a4 > foo $ hg ci -ma4 # 6 $ hg up branchB 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo b3.1 > foo $ hg ci -m b3.1 # 7 $ hg up 5 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo b3.2 > foo $ hg ci -m b3.2 # 8 created new head $ cd ../tt Should succeed because there is only one head on our branch: $ hg pull -u ../t pulling from ../t searching for changes adding changesets adding manifests adding file changes added 3 changesets with 3 changes to 1 files (+1 heads) new changesets da3a8a0161c6:b61cab8fe4e8 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ cd ../t $ hg up -C branchA 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo a5.1 > foo $ hg ci -ma5.1 # 9 $ hg up 6 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo a5.2 > foo $ hg ci -ma5.2 # 10 created new head $ hg up 7 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo b4.1 > foo $ hg ci -m b4.1 # 11 $ hg up -C 8 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo b4.2 > foo $ hg ci -m b4.2 # 12 $ cd ../tt $ hg pull -u ../t pulling from ../t searching for changes adding changesets adding manifests adding file changes added 4 changesets with 4 changes to 1 files (+1 heads) new changesets 0c4d148ae29e:ecfc3f4a6fd9 1 files updated, 0 files merged, 0 files removed, 0 files unresolved updated to "d740e1a584e7: a5.2" 1 other heads for branch "branchA" Make changes on new branch on tt $ hg up 6 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ hg branch branchC marked working directory as branch branchC $ echo b1 > bar $ hg ci -Am "commit on branchC on tt" adding bar Make changes on default branch on t $ cd ../t $ hg up -C default 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo a1 > bar $ hg ci -Am "commit on default on t" adding bar Pull branchC from tt $ hg pull ../tt pulling from ../tt searching for changes adding changesets adding manifests adding file changes added 1 changesets with 1 changes to 1 files (+1 heads) new changesets 7d8ffa4c0b22 13 local changesets published (run 'hg heads' to see heads) Make changes on default and branchC on tt $ cd ../tt $ hg pull ../t pulling from ../t searching for changes adding changesets adding manifests adding file changes added 1 changesets with 1 changes to 1 files (+1 heads) new changesets 2b94b54b6b5f 1 local changesets published (run 'hg heads' to see heads) $ hg up -C default 2 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo a1 > bar1 $ hg ci -Am "commit on default on tt" adding bar1 $ hg up branchC 2 files updated, 0 files merged, 1 files removed, 0 files unresolved $ echo a1 > bar2 $ hg ci -Am "commit on branchC on tt" adding bar2 Make changes on default and branchC on t $ cd ../t $ hg up default 0 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo a1 > bar3 $ hg ci -Am "commit on default on t" adding bar3 $ hg up branchC 2 files updated, 0 files merged, 1 files removed, 0 files unresolved $ echo a1 > bar4 $ hg ci -Am "commit on branchC on tt" adding bar4 Pull from tt $ hg pull ../tt pulling from ../tt searching for changes adding changesets adding manifests adding file changes added 2 changesets with 2 changes to 2 files (+2 heads) new changesets eed40c14b407:e634733b0309 1 local changesets published (run 'hg heads .' to see heads, 'hg merge' to merge) $ cd ..