view mercurial/similar.py @ 35793:4fb2bb61597c

bundle2: increase payload part chunk size to 32kb

Bundle2 payload parts are framed chunks. Esentially, we obtain
data in equal size chunks of size `preferedchunksize` and emit those
to a generator. That generator is fed into a compressor (which can
be the no-op compressor, which just re-emits the generator). And
the output from the compressor likely goes to a file descriptor
or socket.

What this means is that small chunk sizes create more Python objects
and Python function calls than larger chunk sizes. And as we know,
Python object and function call overhead in performance sensitive
code matters (at least with CPython).

This commit increases the bundle2 part payload chunk size from 4k
to 32k. Practically speaking, this means that the chunks we feed
into a compressor (implemented in C code) or feed directly into a
file handle or socket write() are larger. It's possible the chunks
might be larger than what the receiver can handle in one logical
operation. But at that point, we're in C code, which is much more
efficient at dealing with splitting up the chunk and making multiple
function calls than Python is.

A downside to larger chunks is that the receiver has to wait for that
much data to arrive (either raw or from a decompressor) before it
can process the chunk. But 32kb still feels like a small buffer to
have to wait for. And in many cases, the client will convert from
8 read(4096) to 1 read(32768). That's happening in Python land. So
we cut down on the number of Python objects and function calls,
making the client faster as well. I don't think there are any
significant concerns to increasing the payload chunk size to 32kb.

The impact of this change on performance significant. Using `curl`
to obtain a stream clone bundle2 payload from a server on localhost
serving the mozilla-unified repository:

before: 20.78 user; 7.71 system;  80.5 MB/s
after:  13.90 user; 3.51 system; 132 MB/s
legacy:  9.72 user; 8.16 system; 132 MB/s

bundle2 stream clone generation is still more resource intensive than
legacy stream clone (that's likely because of the use of a
util.chunkbuffer). But the throughput is the same. We might
be in territory we're this is effectively a benchmark of the
networking stack or Python's syscall throughput.

From the client perspective, `hg clone -U --stream`:

before: 33.50 user; 7.95 system; 53.3 MB/s
after:  22.82 user; 7.33 system; 72.7 MB/s
legacy: 29.96 user; 7.94 system; 58.0 MB/s

And for `hg clone --stream` with a working directory update of
~230k files:

after:  119.55 user; 26.47 system; 0:57.08 wall
legacy: 126.98 user; 26.94 system; 1:05.56 wall

So, it appears that bundle2's stream clone is now definitively faster
than legacy stream clone!

Differential Revision: https://phab.mercurial-scm.org/D1932

# similar.py - mechanisms for finding similar files
#
# Copyright 2005-2007 Matt Mackall <[e-mail]>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.

from __future__ import absolute_import

from .i18n import _
from . import (
    mdiff,
)

def _findexactmatches(repo, added, removed):
    '''find renamed files that have no changes

    Takes a list of new filectxs and a list of removed filectxs, and yields
    (before, after) tuples of exact matches.
    '''
    numfiles = len(added) + len(removed)

    # Build table of removed files: {hash(fctx.data()): [fctx, ...]}.
    # We use hash() to discard fctx.data() from memory.
    hashes = {}
    for i, fctx in enumerate(removed):
        repo.ui.progress(_('searching for exact renames'), i, total=numfiles,
                         unit=_('files'))
        h = hash(fctx.data())
        if h not in hashes:
            hashes[h] = [fctx]
        else:
            hashes[h].append(fctx)

    # For each added file, see if it corresponds to a removed file.
    for i, fctx in enumerate(added):
        repo.ui.progress(_('searching for exact renames'), i + len(removed),
                total=numfiles, unit=_('files'))
        adata = fctx.data()
        h = hash(adata)
        for rfctx in hashes.get(h, []):
            # compare between actual file contents for exact identity
            if adata == rfctx.data():
                yield (rfctx, fctx)
                break

    # Done
    repo.ui.progress(_('searching for exact renames'), None)

def _ctxdata(fctx):
    # lazily load text
    orig = fctx.data()
    return orig, mdiff.splitnewlines(orig)

def _score(fctx, otherdata):
    orig, lines = otherdata
    text = fctx.data()
    # mdiff.blocks() returns blocks of matching lines
    # count the number of bytes in each
    equal = 0
    matches = mdiff.blocks(text, orig)
    for x1, x2, y1, y2 in matches:
        for line in lines[y1:y2]:
            equal += len(line)

    lengths = len(text) + len(orig)
    return equal * 2.0 / lengths

def score(fctx1, fctx2):
    return _score(fctx1, _ctxdata(fctx2))

def _findsimilarmatches(repo, added, removed, threshold):
    '''find potentially renamed files based on similar file content

    Takes a list of new filectxs and a list of removed filectxs, and yields
    (before, after, score) tuples of partial matches.
    '''
    copies = {}
    for i, r in enumerate(removed):
        repo.ui.progress(_('searching for similar files'), i,
                         total=len(removed), unit=_('files'))

        data = None
        for a in added:
            bestscore = copies.get(a, (None, threshold))[1]
            if data is None:
                data = _ctxdata(r)
            myscore = _score(a, data)
            if myscore > bestscore:
                copies[a] = (r, myscore)
    repo.ui.progress(_('searching'), None)

    for dest, v in copies.iteritems():
        source, bscore = v
        yield source, dest, bscore

def _dropempty(fctxs):
    return [x for x in fctxs if x.size() > 0]

def findrenames(repo, added, removed, threshold):
    '''find renamed files -- yields (before, after, score) tuples'''
    wctx = repo[None]
    pctx = wctx.p1()

    # Zero length files will be frequently unrelated to each other, and
    # tracking the deletion/addition of such a file will probably cause more
    # harm than good. We strip them out here to avoid matching them later on.
    addedfiles = _dropempty(wctx[fp] for fp in sorted(added))
    removedfiles = _dropempty(pctx[fp] for fp in sorted(removed) if fp in pctx)

    # Find exact matches.
    matchedfiles = set()
    for (a, b) in _findexactmatches(repo, addedfiles, removedfiles):
        matchedfiles.add(b)
        yield (a.path(), b.path(), 1.0)

    # If the user requested similar files to be matched, search for them also.
    if threshold < 1.0:
        addedfiles = [x for x in addedfiles if x not in matchedfiles]
        for (a, b, score) in _findsimilarmatches(repo, addedfiles,
                                                 removedfiles, threshold):
            yield (a.path(), b.path(), score)