Mercurial > hg
view tests/test-linerange.py @ 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 | 6939b6ac960a |
children | 2372284d9457 |
line wrap: on
line source
from __future__ import absolute_import import unittest from mercurial import error, mdiff # for readability, line numbers are 0-origin text1 = b''' 00 at OLD 01 at OLD 02 at OLD 02 at NEW, 03 at OLD 03 at NEW, 04 at OLD 04 at NEW, 05 at OLD 05 at NEW, 06 at OLD 07 at OLD 08 at OLD 09 at OLD 10 at OLD 11 at OLD '''[1:] # strip initial LF text2 = b''' 00 at NEW 01 at NEW 02 at NEW, 03 at OLD 03 at NEW, 04 at OLD 04 at NEW, 05 at OLD 05 at NEW, 06 at OLD 06 at NEW 07 at NEW 08 at NEW 09 at NEW 10 at NEW 11 at NEW '''[1:] # strip initial LF def filteredblocks(blocks, rangeb): """return `rangea` extracted from `blocks` coming from `mdiff.blocksinrange` along with the mask of blocks within rangeb. """ filtered, rangea = mdiff.blocksinrange(blocks, rangeb) skipped = [b not in filtered for b in blocks] return rangea, skipped class blocksinrangetests(unittest.TestCase): def setUp(self): self.blocks = list(mdiff.allblocks(text1, text2)) assert self.blocks == [ ([0, 3, 0, 2], b'!'), ((3, 7, 2, 6), b'='), ([7, 12, 6, 12], b'!'), ((12, 12, 12, 12), b'='), ], self.blocks def testWithinEqual(self): """linerange within an "=" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^^ linerange2 = (3, 5) linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (4, 6)) self.assertEqual(skipped, [True, False, True, True]) def testWithinEqualStrictly(self): """linerange matching exactly an "=" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^^^^ linerange2 = (2, 6) linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (3, 7)) self.assertEqual(skipped, [True, False, True, True]) def testWithinEqualLowerbound(self): """linerange at beginning of an "=" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^^ linerange2 = (2, 4) linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (3, 5)) self.assertEqual(skipped, [True, False, True, True]) def testWithinEqualLowerboundOneline(self): """oneline-linerange at beginning of an "=" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^ linerange2 = (2, 3) linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (3, 4)) self.assertEqual(skipped, [True, False, True, True]) def testWithinEqualUpperbound(self): """linerange at end of an "=" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^^^ linerange2 = (3, 6) linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (4, 7)) self.assertEqual(skipped, [True, False, True, True]) def testWithinEqualUpperboundOneLine(self): """oneline-linerange at end of an "=" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^ linerange2 = (5, 6) linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (6, 7)) self.assertEqual(skipped, [True, False, True, True]) def testWithinFirstBlockNeq(self): """linerange within the first "!" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^ # | (empty) # ^ # ^^ for linerange2 in [ (0, 1), (1, 1), (1, 2), (0, 2), ]: linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (0, 3)) self.assertEqual(skipped, [False, True, True, True]) def testWithinLastBlockNeq(self): """linerange within the last "!" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^ # ^ # | (empty) # ^^^^^^ # ^ for linerange2 in [ (6, 7), (7, 8), (7, 7), (6, 12), (11, 12), ]: linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (7, 12)) self.assertEqual(skipped, [True, True, False, True]) def testAccrossTwoBlocks(self): """linerange accross two blocks""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^^^^ linerange2 = (1, 5) linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (0, 6)) self.assertEqual(skipped, [False, False, True, True]) def testCrossingSeveralBlocks(self): """linerange accross three blocks""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^^^^^^^ linerange2 = (1, 8) linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, (0, 12)) self.assertEqual(skipped, [False, False, False, True]) def testStartInEqBlock(self): """linerange starting in an "=" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^^^^ # ^^^^^^^ for linerange2, expectedlinerange1 in [ ((5, 9), (6, 12)), ((4, 11), (5, 12)), ]: linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, expectedlinerange1) self.assertEqual(skipped, [True, False, False, True]) def testEndInEqBlock(self): """linerange ending in an "=" block""" # IDX 0 1 # 012345678901 # SRC NNOOOONNNNNN (New/Old) # ^^ # ^^^^^ for linerange2, expectedlinerange1 in [ ((1, 3), (0, 4)), ((0, 4), (0, 5)), ]: linerange1, skipped = filteredblocks(self.blocks, linerange2) self.assertEqual(linerange1, expectedlinerange1) self.assertEqual(skipped, [False, False, True, True]) def testOutOfRange(self): """linerange exceeding file size""" exctype = error.Abort for linerange2 in [ (0, 34), (15, 12), ]: # Could be `with self.assertRaises(error.Abort)` but python2.6 # does not have assertRaises context manager. try: mdiff.blocksinrange(self.blocks, linerange2) except exctype as exc: self.assertTrue('line range exceeds file size' in str(exc)) else: self.fail('%s not raised' % exctype.__name__) if __name__ == '__main__': import silenttestrunner silenttestrunner.main(__name__)