FazBrowse GitHub Viewer | Trending |
URL:
| Home
Tools: [Download Repo ZIP]   [Original HTTPS Page]

Make benchmarking 25% faster by parallelizing benchmarking venv setup by Fidget-Spinner · Pull Request #375 · python/pyperformance · GitHub

  1. python
  2. pyperformance

Repository navigation

Make benchmarking 25% faster by parallelizing benchmarking venv setup #375

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
Fidget-Spinner wants to merge 4 commits into from
Conversation 6 Commits 4 Checks 0 Files changed
Closed

Make benchmarking 25% faster by parallelizing benchmarking venv setup #375

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
ecbfbf0
Parallelize benchmarking venv setup
Fidget-Spinner Dec 16, 2024
db29d34
fix bug
Fidget-Spinner Dec 16, 2024
aefde59
Use threads because windows is unhappy
Fidget-Spinner Dec 16, 2024
0cfc548
switch back to processpoolexecutor
Fidget-Spinner Dec 16, 2024
File filter

Filter by extension

Filter by extension .py  (1) All 1 file type selected
Viewed files
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Unified
Split
Hide whitespace
Diff view
Unified
Split
Hide whitespace
85 changes: 56 additions & 29 deletions pyperformance/run.py
Show comments
View file
Open in desktop
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters. Learn more about bidirectional Unicode characters
Show hidden characters
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,9 @@
import sys
import time
import traceback
import concurrent.futures
import io
import contextlib

import pyperformance
from . import _utils, _python, _pythoninfo
Expand Down Expand Up @@ -67,47 +70,40 @@ def get_loops_from_file(filename):
return loops


def run_benchmarks(should_run, python, options):
if options.same_loops is not None:
loops = get_loops_from_file(options.same_loops)
else:
loops = {}

to_run = sorted(should_run)
def setup_single_venv(args):
(i, num_benchmarks, python, options, bench) = args

info = _pythoninfo.get_info(python)
runid = get_run_id(info)
stdout = io.StringIO()
with contextlib.redirect_stdout(stdout):
info = _pythoninfo.get_info(python)
runid = get_run_id(info)

unique = getattr(options, 'unique_venvs', False)
if not unique:
common = VenvForBenchmarks.ensure(
_venv.get_venv_root(runid.name, python=info),
info,
upgrade='oncreate',
inherit_environ=options.inherit_environ,
)

benchmarks = {}
venvs = set()
for i, bench in enumerate(to_run):
unique = getattr(options, 'unique_venvs', False)
if not unique:
common = VenvForBenchmarks.ensure(
_venv.get_venv_root(runid.name, python=info),
info,
upgrade='oncreate',
inherit_environ=options.inherit_environ,
)
bench_runid = runid._replace(bench=bench)
assert bench_runid.name, (bench, bench_runid)
name = bench_runid.name
venv_root = _venv.get_venv_root(name, python=info)
bench_status = f'({i+1:>2}/{num_benchmarks})'
print()
print('='*50)
print(f'({i+1:>2}/{len(to_run)}) creating venv for benchmark ({bench.name})')
print(f'{bench_status} creating venv for benchmark ({bench.name})')
print()
if not unique:
print('(trying common venv first)')
print(f'{bench_status} (trying common venv first)')
# Try the common venv first.
try:
common.ensure_reqs(bench)
except _venv.RequirementsInstallationFailedError:
print('(falling back to unique venv)')
print(f'{bench_status} (falling back to unique venv)')
else:
benchmarks[bench] = (common, bench_runid)
continue
return(bench, None, common, bench_runid, stdout.getvalue())
try:
venv = VenvForBenchmarks.ensure(
venv_root,
Expand All @@ -118,12 +114,43 @@ def run_benchmarks(should_run, python, options):
# XXX Do not override when there is a requirements collision.
venv.ensure_reqs(bench)
except _venv.RequirementsInstallationFailedError:
print('(benchmark will be skipped)')
print(f'{bench_status} (benchmark will be skipped)')
print()
venv = None
print(f'{bench_status} done')
return (bench, venv_root, venv, bench_runid, stdout.getvalue())

def run_benchmarks(should_run, python, options):
if options.same_loops is not None:
loops = get_loops_from_file(options.same_loops)
else:
loops = {}

to_run = sorted(should_run)
benchmarks = {}
venvs = set()

# Setup a first venv on its own first to create common
# requirements without threading issues.
bench, venv_root, venv, bench_runid, cons_output = setup_single_venv((0, len(to_run), python, options, to_run[0]))
if venv_root is not None:
venvs.add(venv_root)
benchmarks[bench] = (venv, bench_runid)
print()
benchmarks[bench] = (venv, bench_runid)
print(cons_output)

# Parallelise the rest.
executor_input = [(i+1, len(to_run), python, options, bench)
for i, bench in enumerate(to_run[1:])]
# It's fine to set a higher worker count, because this is IO-bound anyways.
with concurrent.futures.ProcessPoolExecutor() as executor:
for bench, venv_root, venv, bench_runid, cons_output in list(executor.map(setup_single_venv, executor_input)):
if venv_root is not None:
venvs.add(venv_root)
benchmarks[bench] = (venv, bench_runid)
print(cons_output)

print("Completed venv installation. Now sleeping 15s to stabilize thermals.")
time.sleep(15)

suite = None
run_count = str(len(to_run))
Expand Down
Back | FazBrowse Home | New Git URL