Mercurial > hg
view tests/test-convert-p4.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 | 78ac8acfc4bd |
children |
line wrap: on
line source
#require p4 $ echo "[extensions]" >> $HGRCPATH $ echo "convert = " >> $HGRCPATH create p4 depot $ P4ROOT=`pwd`/depot; export P4ROOT $ P4AUDIT=$P4ROOT/audit; export P4AUDIT $ P4JOURNAL=$P4ROOT/journal; export P4JOURNAL $ P4LOG=$P4ROOT/log; export P4LOG $ P4PORT=localhost:$HGPORT; export P4PORT $ P4DEBUG=1; export P4DEBUG start the p4 server $ [ ! -d $P4ROOT ] && mkdir $P4ROOT $ p4d -f -J off >$P4ROOT/stdout 2>$P4ROOT/stderr & $ echo $! >> $DAEMON_PIDS $ trap "echo stopping the p4 server ; p4 admin stop" EXIT $ # wait for the server to initialize $ while ! p4 ; do > sleep 1 > done >/dev/null 2>/dev/null create a client spec $ P4CLIENT=hg-p4-import; export P4CLIENT $ DEPOTPATH=//depot/test-mercurial-import/... $ p4 client -o | sed '/^View:/,$ d' >p4client $ echo View: >>p4client $ echo " $DEPOTPATH //$P4CLIENT/..." >>p4client $ p4 client -i <p4client Client hg-p4-import saved. populate the depot $ echo a > a $ mkdir b $ echo c > b/c $ p4 add a b/c //depot/test-mercurial-import/a#1 - opened for add //depot/test-mercurial-import/b/c#1 - opened for add $ p4 submit -d initial Submitting change 1. Locking 2 files ... add //depot/test-mercurial-import/a#1 add //depot/test-mercurial-import/b/c#1 Change 1 submitted. change some files $ p4 edit a //depot/test-mercurial-import/a#1 - opened for edit $ echo aa >> a $ p4 submit -d "change a" Submitting change 2. Locking 1 files ... edit //depot/test-mercurial-import/a#2 Change 2 submitted. $ p4 edit b/c //depot/test-mercurial-import/b/c#1 - opened for edit $ echo cc >> b/c $ p4 submit -d "change b/c" Submitting change 3. Locking 1 files ... edit //depot/test-mercurial-import/b/c#2 Change 3 submitted. convert $ hg convert -s p4 $DEPOTPATH dst initializing destination dst repository scanning source... reading p4 views collecting p4 changelists 1 initial 2 change a 3 change b/c sorting... converting... 2 initial 1 change a 0 change b/c $ hg -R dst log --template 'rev={rev} desc="{desc}" tags="{tags}" files="{files}"\n' rev=2 desc="change b/c" tags="tip" files="b/c" rev=1 desc="change a" tags="" files="a" rev=0 desc="initial" tags="" files="a b/c" change some files $ p4 edit a b/c //depot/test-mercurial-import/a#2 - opened for edit //depot/test-mercurial-import/b/c#2 - opened for edit $ echo aaa >> a $ echo ccc >> b/c $ p4 submit -d "change a b/c" Submitting change 4. Locking 2 files ... edit //depot/test-mercurial-import/a#3 edit //depot/test-mercurial-import/b/c#3 Change 4 submitted. convert again $ hg convert -s p4 $DEPOTPATH dst scanning source... reading p4 views collecting p4 changelists 4 change a b/c sorting... converting... 0 change a b/c $ hg -R dst log --template 'rev={rev} desc="{desc}" tags="{tags}" files="{files}"\n' rev=3 desc="change a b/c" tags="tip" files="a b/c" rev=2 desc="change b/c" tags="" files="b/c" rev=1 desc="change a" tags="" files="a" rev=0 desc="initial" tags="" files="a b/c" interesting names $ echo dddd > "d d" $ mkdir " e" $ echo fff >" e/ f" $ p4 add "d d" " e/ f" //depot/test-mercurial-import/d d#1 - opened for add //depot/test-mercurial-import/ e/ f#1 - opened for add $ p4 submit -d "add d e f" Submitting change 5. Locking 2 files ... add //depot/test-mercurial-import/ e/ f#1 add //depot/test-mercurial-import/d d#1 Change 5 submitted. convert again $ hg convert -s p4 $DEPOTPATH dst scanning source... reading p4 views collecting p4 changelists 5 add d e f sorting... converting... 0 add d e f $ hg -R dst log --template 'rev={rev} desc="{desc}" tags="{tags}" files="{files}"\n' rev=4 desc="add d e f" tags="tip" files=" e/ f d d" rev=3 desc="change a b/c" tags="" files="a b/c" rev=2 desc="change b/c" tags="" files="b/c" rev=1 desc="change a" tags="" files="a" rev=0 desc="initial" tags="" files="a b/c" empty commit message $ p4 edit a //depot/test-mercurial-import/a#3 - opened for edit $ echo aaaaa >> a $ p4 submit -d "" Submitting change 6. Locking 1 files ... edit //depot/test-mercurial-import/a#4 Change 6 submitted. $ hg convert -s p4 $DEPOTPATH dst scanning source... reading p4 views collecting p4 changelists 6 **empty changelist description** sorting... converting... 0 exit trap: stopping the p4 server