Mercurial > hg
view tests/test-bundle-vs-outgoing.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 | eb586ed5d8ce |
children |
line wrap: on
line source
this structure seems to tickle a bug in bundle's search for changesets, so first we have to recreate it o 8 | | o 7 | | | o 6 |/| o | 5 | | o | 4 | | | o 3 | | | o 2 |/ o 1 | o 0 $ mkrev() > { > revno=$1 > echo "rev $revno" > echo "rev $revno" > foo.txt > hg -q ci -m"rev $revno" > } setup test repo1 $ hg init repo1 $ cd repo1 $ echo "rev 0" > foo.txt $ hg ci -Am"rev 0" adding foo.txt $ mkrev 1 rev 1 first branch $ mkrev 2 rev 2 $ mkrev 3 rev 3 back to rev 1 to create second branch $ hg up -r1 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ mkrev 4 rev 4 $ mkrev 5 rev 5 merge first branch to second branch $ hg up -C -r5 0 files updated, 0 files merged, 0 files removed, 0 files unresolved $ HGMERGE=internal:local hg merge 0 files updated, 1 files merged, 0 files removed, 0 files unresolved (branch merge, don't forget to commit) $ echo "merge rev 5, rev 3" > foo.txt $ hg ci -m"merge first branch to second branch" one more commit following the merge $ mkrev 7 rev 7 back to "second branch" to make another head $ hg up -r5 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ mkrev 8 rev 8 the story so far $ hg log -G --template "{rev}\n" @ 8 | | o 7 | | | o 6 |/| o | 5 | | o | 4 | | | o 3 | | | o 2 |/ o 1 | o 0 check that "hg outgoing" really does the right thing sanity check of outgoing: expect revs 4 5 6 7 8 $ hg clone -r3 . ../repo2 adding changesets adding manifests adding file changes added 4 changesets with 4 changes to 1 files new changesets 6ae4cca4e39a:478f191e53f8 updating to branch default 1 files updated, 0 files merged, 0 files removed, 0 files unresolved this should (and does) report 5 outgoing revisions: 4 5 6 7 8 $ hg outgoing --template "{rev}\n" ../repo2 comparing with ../repo2 searching for changes 4 5 6 7 8 test bundle (destination repo): expect 5 revisions this should bundle the same 5 revisions that outgoing reported, but it actually bundles 7 $ hg bundle foo.bundle ../repo2 searching for changes 5 changesets found test bundle (base revision): expect 5 revisions this should (and does) give exactly the same result as bundle with a destination repo... i.e. it's wrong too $ hg bundle --base 3 foo.bundle 5 changesets found $ cd ..