util: reimplement lrucachedict
As part of attempting to more aggressively use the existing
lrucachedict, collections.deque operations were frequently
showing up in profiling output, negating benefits of caching.
Searching the internet seems to tell me that the most efficient
way to implement an LRU cache in Python is to have a dict indexing
the cached entries and then to use a doubly linked list to track
freshness of each entry. So, this patch replaces our existing
lrucachedict with a version using such a pattern.
The recently introduced perflrucachedict command reveals the
following timings for 10,000 operations for the following cache
sizes for the existing cache:
n=4 init=0.004079 gets=0.003632 sets=0.005188 mixed=0.005402
n=8 init=0.004045 gets=0.003998 sets=0.005064 mixed=0.005328
n=16 init=0.004011 gets=0.004496 sets=0.005021 mixed=0.005555
n=32 init=0.004064 gets=0.005611 sets=0.005188 mixed=0.006189
n=64 init=0.003975 gets=0.007684 sets=0.005178 mixed=0.007245
n=128 init=0.004121 gets=0.012005 sets=0.005422 mixed=0.009471
n=256 init=0.004143 gets=0.020295 sets=0.005227 mixed=0.013612
n=512 init=0.004039 gets=0.036703 sets=0.005243 mixed=0.020685
n=1024 init=0.004193 gets=0.068142 sets=0.005251 mixed=0.033064
n=2048 init=0.004070 gets=0.133383 sets=0.005160 mixed=0.050359
n=4096 init=0.004053 gets=0.265194 sets=0.004868 mixed=0.048352
n=8192 init=0.004087 gets=0.542218 sets=0.004562 mixed=0.032753
n=16384 init=0.004106 gets=1.064055 sets=0.004179 mixed=0.020367
n=32768 init=0.004034 gets=2.097620 sets=0.004260 mixed=0.013031
n=65536 init=0.004108 gets=4.106390 sets=0.004268 mixed=0.010191
As the data shows, the existing cache's retrieval performance
diminishes linearly with cache size. (Keep in mind the microbenchmark
is testing 100% cache hit rate.)
The new cache implementation reveals the following:
n=4 init=0.006665 gets=0.006541 sets=0.005733 mixed=0.006876
n=8 init=0.006649 gets=0.006374 sets=0.005663 mixed=0.006899
n=16 init=0.006570 gets=0.006504 sets=0.005799 mixed=0.007057
n=32 init=0.006854 gets=0.006459 sets=0.005747 mixed=0.007034
n=64 init=0.006580 gets=0.006495 sets=0.005740 mixed=0.006992
n=128 init=0.006534 gets=0.006739 sets=0.005648 mixed=0.007124
n=256 init=0.006669 gets=0.006773 sets=0.005824 mixed=0.007151
n=512 init=0.006701 gets=0.007061 sets=0.006042 mixed=0.007372
n=1024 init=0.006641 gets=0.007620 sets=0.006387 mixed=0.007464
n=2048 init=0.006517 gets=0.008598 sets=0.006871 mixed=0.008077
n=4096 init=0.006720 gets=0.010933 sets=0.007854 mixed=0.008663
n=8192 init=0.007383 gets=0.015969 sets=0.010288 mixed=0.008896
n=16384 init=0.006660 gets=0.025447 sets=0.011208 mixed=0.008826
n=32768 init=0.006658 gets=0.044390 sets=0.011192 mixed=0.008943
n=65536 init=0.006836 gets=0.082736 sets=0.011151 mixed=0.008826
Let's go through the results.
The new cache takes longer to construct. ~6.6ms vs ~4.1ms. However,
this is measuring 10,000 __init__ calls, so the difference is
~0.2us/instance. We currently only create lrucachedict for manifest
instances, so this regression is not likely relevant.
The new cache is slightly slower for retrievals for cache sizes
< 1024. It's worth noting that the only existing use of lurcachedict
is in manifest.py and the default cache size is 4. This regression
is worrisome. However, for n=4, the delta is ~2.9s for 10,000 lookups,
or ~0.29us/op. Again, this is a marginal regression and likely not
relevant in the real world. Timing `hg log -p -l 100` for
mozilla-central reveals that cache lookup times are dominated by
decompression and fulltext resolution (even with lz4 manifests).
The new cache is significantly faster for retrievals at larger
capacities. Whereas the old implementation has retrieval performance
linear with cache capacity, the new cache is constant time until much
larger values. And, when it does start to increase significantly, it
is a few magnitudes faster than the current cache.
The new cache does appear to be slower for sets when capacity is large.
However, performance is similar for smaller capacities. Of course,
caches should generally be optimized for retrieval performance because
if a cache is getting more sets than gets, it doesn't really make
sense to cache. If this regression is worrisome, again, taking the
largest regression at n=65536 of ~6.9ms for 10,000 results in a
regression of ~0.68us/op. This is not significant in the grand scheme
of things.
Overall, the new cache is performant at retrievals at much larger
capacity values which makes it a generally more useful cache backend.
While there are regressions, their absolute value is extremely small.
Since we aren't using lrucachedict aggressively today, these
regressions should not be relevant. The improved scalability of
lrucachedict should enable us to more aggressively utilize
lrucachedict for more granular caching (read: higher capacity caches)
in the near future. The impetus for this patch is to establish a cache
of decompressed revlog revisions, notably manifest revisions. And since
delta chains can grow to >10,000 and cache hit rate can be high, the
improved retrieval performance of lrucachedict should be relevant.
setup repo
$ hg init t
$ cd t
$ echo a > a
$ hg commit -Am'add a'
adding a
$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
1 files, 1 changesets, 1 total revisions
$ hg parents
changeset: 0:1f0dee641bb7
tag: tip
user: test
date: Thu Jan 01 00:00:00 1970 +0000
summary: add a
rollback to null revision
$ hg status
$ hg rollback
repository tip rolled back to revision -1 (undo commit)
working directory now based on revision -1
$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
0 files, 0 changesets, 0 total revisions
$ hg parents
$ hg status
A a
Two changesets this time so we rollback to a real changeset
$ hg commit -m'add a again'
$ echo a >> a
$ hg commit -m'modify a'
Test issue 902 (current branch is preserved)
$ hg branch test
marked working directory as branch test
(branches are permanent and global, did you want a bookmark?)
$ hg rollback
repository tip rolled back to revision 0 (undo commit)
working directory now based on revision 0
$ hg branch
default
Test issue 1635 (commit message saved)
$ cat .hg/last-message.txt ; echo
modify a
Test rollback of hg before issue 902 was fixed
$ hg commit -m "test3"
$ hg branch test
marked working directory as branch test
(branches are permanent and global, did you want a bookmark?)
$ rm .hg/undo.branch
$ hg rollback
repository tip rolled back to revision 0 (undo commit)
named branch could not be reset: current branch is still 'test'
working directory now based on revision 0
$ hg branch
test
working dir unaffected by rollback: do not restore dirstate et. al.
$ hg log --template '{rev} {branch} {desc|firstline}\n'
0 default add a again
$ hg status
M a
$ hg bookmark foo
$ hg commit -m'modify a again'
$ echo b > b
$ hg bookmark bar -r default #making bar active, before the transaction
$ hg commit -Am'add b'
adding b
$ hg log --template '{rev} {branch} {desc|firstline}\n'
2 test add b
1 test modify a again
0 default add a again
$ hg update bar
1 files updated, 0 files merged, 1 files removed, 0 files unresolved
(activating bookmark bar)
$ cat .hg/undo.branch ; echo
test
$ hg rollback -f
repository tip rolled back to revision 1 (undo commit)
$ hg id -n
0
$ hg branch
default
$ cat .hg/bookmarks.current ; echo
bar
$ hg bookmark --delete foo bar
rollback by pretxncommit saves commit message (issue1635)
$ echo a >> a
$ hg --config hooks.pretxncommit=false commit -m"precious commit message"
transaction abort!
rollback completed
abort: pretxncommit hook exited with status * (glob)
[255]
$ cat .hg/last-message.txt ; echo
precious commit message
same thing, but run $EDITOR
$ cat > editor.sh << '__EOF__'
> echo "another precious commit message" > "$1"
> __EOF__
$ HGEDITOR="\"sh\" \"`pwd`/editor.sh\"" hg --config hooks.pretxncommit=false commit 2>&1
note: commit message saved in .hg/last-message.txt
transaction abort!
rollback completed
abort: pretxncommit hook exited with status * (glob)
[255]
$ cat .hg/last-message.txt
another precious commit message
test rollback on served repository
#if serve
$ hg commit -m "precious commit message"
$ hg serve -p $HGPORT -d --pid-file=hg.pid -A access.log -E errors.log
$ cat hg.pid >> $DAEMON_PIDS
$ cd ..
$ hg clone http://localhost:$HGPORT u
requesting all changes
adding changesets
adding manifests
adding file changes
added 3 changesets with 2 changes to 1 files (+1 heads)
updating to branch default
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd u
$ hg id default
068774709090
now rollback and observe that 'hg serve' reloads the repository and
presents the correct tip changeset:
$ hg -R ../t rollback
repository tip rolled back to revision 1 (undo commit)
working directory now based on revision 0
$ hg id default
791dd2169706
#endif
update to older changeset and then refuse rollback, because
that would lose data (issue2998)
$ cd ../t
$ hg -q update
$ rm `hg status -un`
$ template='{rev}:{node|short} [{branch}] {desc|firstline}\n'
$ echo 'valuable new file' > b
$ echo 'valuable modification' >> a
$ hg commit -A -m'a valuable change'
adding b
$ hg update 0
1 files updated, 0 files merged, 1 files removed, 0 files unresolved
$ hg rollback
abort: rollback of last commit while not checked out may lose data
(use -f to force)
[255]
$ hg tip -q
2:4d9cd3795eea
$ hg rollback -f
repository tip rolled back to revision 1 (undo commit)
$ hg status
$ hg log --removed b # yep, it's gone
same again, but emulate an old client that doesn't write undo.desc
$ hg -q update
$ echo 'valuable modification redux' >> a
$ hg commit -m'a valuable change redux'
$ rm .hg/undo.desc
$ hg update 0
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg rollback
rolling back unknown transaction
$ cat a
a
corrupt journal test
$ echo "foo" > .hg/store/journal
$ hg recover
rolling back interrupted transaction
couldn't read journal entry 'foo\n'!
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
1 files, 2 changesets, 2 total revisions