contrib/perf-utils/compare-discovery-case

stream-clone: disable gc for the initial section for the v3 format

The number of small container created turn Python in a gc-frenzy that seriously
impact performance.

This significantly boost performance. The following number comes from a large
private repository using perf::stream-locked-section:

base-line:   35.04 seconds
prev-change: 24.51 seconds (-30%)
prev-change: 20.88 seconds (-40%)
this-change: 14.22 seconds (-60% from baseline; -31% from prev)

#!/usr/bin/env python3
# compare various algorithm variants for a given case
#
#  search-discovery-case REPO LOCAL_CASE REMOTE_CASE
#
# The description for the case input uses the same format as the ouput of
# search-discovery-case

import json
import os
import subprocess
import sys

this_script = os.path.abspath(sys.argv[0])
script_name = os.path.basename(this_script)
this_dir = os.path.dirname(this_script)
hg_dir = os.path.join(this_dir, '..', '..')
HG_REPO = os.path.normpath(hg_dir)
HG_BIN = os.path.join(HG_REPO, 'hg')


SUBSET_PATH = os.path.join(HG_REPO, 'contrib', 'perf-utils', 'subsetmaker.py')

CMD_BASE = (
    HG_BIN,
    'debugdiscovery',
    '--template',
    'json',
    '--config',
    'extensions.subset=%s' % SUBSET_PATH,
)

# --old
# --nonheads
#
# devel.discovery.exchange-heads=True
# devel.discovery.grow-sample=True
# devel.discovery.grow-sample.dynamic=True

VARIANTS = {
    'tree-discovery': ('--old',),
    'set-discovery-basic': (
        '--config',
        'devel.discovery.exchange-heads=no',
        '--config',
        'devel.discovery.grow-sample=no',
        '--config',
        'devel.discovery.grow-sample.dynamic=no',
        '--config',
        'devel.discovery.randomize=yes',
    ),
    'set-discovery-heads': (
        '--config',
        'devel.discovery.exchange-heads=yes',
        '--config',
        'devel.discovery.grow-sample=no',
        '--config',
        'devel.discovery.grow-sample.dynamic=no',
        '--config',
        'devel.discovery.randomize=yes',
    ),
    'set-discovery-grow-sample': (
        '--config',
        'devel.discovery.exchange-heads=yes',
        '--config',
        'devel.discovery.grow-sample=yes',
        '--config',
        'devel.discovery.grow-sample.dynamic=no',
        '--config',
        'devel.discovery.randomize=yes',
    ),
    'set-discovery-dynamic-sample': (
        '--config',
        'devel.discovery.exchange-heads=yes',
        '--config',
        'devel.discovery.grow-sample=yes',
        '--config',
        'devel.discovery.grow-sample.dynamic=yes',
        '--config',
        'devel.discovery.randomize=yes',
    ),
    'set-discovery-default': (
        '--config',
        'devel.discovery.randomize=yes',
    ),
}

VARIANTS_KEYS = [
    'tree-discovery',
    'set-discovery-basic',
    'set-discovery-heads',
    'set-discovery-grow-sample',
    'set-discovery-dynamic-sample',
    'set-discovery-default',
]

assert set(VARIANTS.keys()) == set(VARIANTS_KEYS)


def parse_case(case):
    case_type, case_args = case.split('-', 1)
    if case_type == 'file':
        case_args = (case_args,)
    else:
        case_args = tuple(int(x) for x in case_args.split('-'))
    case = (case_type,) + case_args
    return case


def format_case(case):
    return '-'.join(str(s) for s in case)


def to_revsets(case):
    t = case[0]
    if t == 'scratch':
        return 'not scratch(all(), %d, "%d")' % (case[1], case[2])
    elif t == 'randomantichain':
        return '::randomantichain(all(), "%d")' % case[1]
    elif t == 'rev':
        return '::%d' % case[1]
    elif t == 'file':
        return '::nodefromfile("%s")' % case[1]
    else:
        assert False


def compare(
    repo,
    local_case,
    remote_case,
    display_header=True,
    display_case=True,
):
    case = (repo, local_case, remote_case)
    if display_header:
        pieces = ['#']
        if display_case:
            pieces += [
                "repo",
                "local-subset",
                "remote-subset",
            ]

        pieces += [
            "discovery-variant",
            "roundtrips",
            "queries",
            "revs",
            "local-heads",
            "common-heads",
            "undecided-initial",
            "undecided-common",
            "undecided-missing",
        ]
        print(*pieces)
    for variant in VARIANTS_KEYS:
        res = process(case, VARIANTS[variant])
        revs = res["nb-revs"]
        local_heads = res["nb-head-local"]
        common_heads = res["nb-common-heads"]
        roundtrips = res["total-roundtrips"]
        queries = res["total-queries"]
        pieces = []
        if display_case:
            pieces += [
                repo,
                format_case(local_case),
                format_case(remote_case),
            ]
        pieces += [
            variant,
            roundtrips,
            queries,
            revs,
            local_heads,
            common_heads,
        ]
        if 'tree-discovery' not in variant:
            undecided_common = res["nb-ini_und-common"]
            undecided_missing = res["nb-ini_und-missing"]
            undecided = undecided_common + undecided_missing
            pieces += [
                undecided,
                undecided_common,
                undecided_missing,
            ]
        print(*pieces)
    return 0


def process(case, variant):
    (repo, left, right) = case
    cmd = list(CMD_BASE)
    cmd.append('-R')
    cmd.append(repo)
    cmd.append('--local-as-revs')
    cmd.append(to_revsets(left))
    cmd.append('--remote-as-revs')
    cmd.append(to_revsets(right))
    cmd.extend(variant)
    s = subprocess.Popen(cmd, stdout=subprocess.PIPE)
    out, err = s.communicate()
    return json.loads(out)[0]


if __name__ == '__main__':

    argv = sys.argv[:]

    kwargs = {}
    # primitive arg parsing
    if '--no-header' in argv:
        kwargs['display_header'] = False
        argv = [a for a in argv if a != '--no-header']
    if '--no-case' in argv:
        kwargs['display_case'] = False
        argv = [a for a in argv if a != '--no-case']

    if len(argv) != 4:
        usage = f'USAGE: {script_name} REPO LOCAL_CASE REMOTE_CASE'
        print(usage, file=sys.stderr)
        sys.exit(128)
    repo = argv[1]
    local_case = parse_case(argv[2])
    remote_case = parse_case(argv[3])
    sys.exit(compare(repo, local_case, remote_case, **kwargs))