Faster CPython Benchmark Infrastructure

🔒 ▶️ START A BENCHMARK RUN

Results

Here are some recent and important revisions. 👉 Complete list of results.

• Most recent pystats on main (9609574)
• Most recent PYTHON_UOPS pystats on main (9609574)
• Most recent JIT pystats on main (bedaea0)

linux x86_64 (vultr)

date	fork/ref	hash/flags	vs. 3.12.6:	vs. 3.13.0rc2:	vs. base:
2026-01-06	python/b866a1c73f8160647545	b866a1c (NOGIL)	1.030x ↓ 📄📈	1.063x ↓ 📄📈	1.097x ↓ 📄📈🧠
2026-01-06	python/b866a1c73f8160647545	b866a1c	1.069x ↑ 📄📈	1.033x ↑ 📄📈
2026-01-05	python/04ace41fe2cf648be433	04ace41 (NOGIL)	1.032x ↓ 📄📈	1.065x ↓ 📄📈	1.103x ↓ 📄📈🧠
2026-01-05	python/04ace41fe2cf648be433	04ace41	1.074x ↑ 📄📈	1.038x ↑ 📄📈
2026-01-04	python/c99f7667436d8978b407	c99f766 (NOGIL)	1.035x ↓ 📄📈	1.068x ↓ 📄📈	1.101x ↓ 📄📈🧠
2026-01-04	python/c99f7667436d8978b407	c99f766	1.068x ↑ 📄📈	1.032x ↑ 📄📈
2026-01-03	python/9609574e7fd36edfaa8b	9609574 (NOGIL)	1.032x ↓ 📄📈	1.065x ↓ 📄📈	1.104x ↓ 📄📈🧠
2026-01-03	python/9609574e7fd36edfaa8b	9609574 (JIT)	1.124x ↑ 📄📈	1.086x ↑ 📄📈	1.040x ↑ 📄📈🧠
2026-01-03	python/9609574e7fd36edfaa8b	9609574 (CLANG)	1.131x ↑ 📄📈	1.094x ↑ 📄📈	1.051x ↑ 📄📈🧠
2026-01-03	python/9609574e7fd36edfaa8b	9609574	1.074x ↑ 📄📈	1.039x ↑ 📄📈

darwin arm64 (macm4pro)

date	fork/ref	hash/flags	vs. 3.12.6:	vs. 3.13.0rc2:	vs. base:
2026-01-06	python/b866a1c73f8160647545	b866a1c (NOGIL)	1.082x ↑ 📄📈	1.004x ↑ 📄📈	1.067x ↓ 📄📈🧠
2026-01-06	python/b866a1c73f8160647545	b866a1c	1.159x ↑ 📄📈	1.075x ↑ 📄📈
2026-01-05	python/04ace41fe2cf648be433	04ace41 (NOGIL)	1.083x ↑ 📄📈	1.005x ↑ 📄📈	1.065x ↓ 📄📈🧠
2026-01-05	python/04ace41fe2cf648be433	04ace41	1.158x ↑ 📄📈	1.074x ↑ 📄📈
2026-01-04	python/c99f7667436d8978b407	c99f766 (NOGIL)	1.085x ↑ 📄📈	1.006x ↑ 📄📈	1.061x ↓ 📄📈🧠
2026-01-04	python/c99f7667436d8978b407	c99f766	1.153x ↑ 📄📈	1.070x ↑ 📄📈
2026-01-03	python/9609574e7fd36edfaa8b	9609574 (NOGIL)	1.087x ↑ 📄📈	1.007x ↑ 📄📈	1.069x ↓ 📄📈🧠
2026-01-03	python/9609574e7fd36edfaa8b	9609574 (JIT)	1.250x ↑ 📄📈	1.159x ↑ 📄📈	1.072x ↑ 📄📈🧠
2026-01-03	python/9609574e7fd36edfaa8b	9609574 (CLANG)	1.175x ↑ 📄📈	1.088x ↑ 📄📈	1.007x ↑ 📄📈🧠
2026-01-03	python/9609574e7fd36edfaa8b	9609574	1.163x ↑ 📄📈	1.079x ↑ 📄📈

* indicates that the exact same versions of pyperformance was not used.

Improvement of the geometric mean of key merged benchmarks, computed with pyperf compare. The results have a resolution of 0.01 (1%).

Documentation

Running benchmarks from the GitHub web UI

Visit the 🔒 benchmark action and click the "Run Workflow" button.

The available parameters are:

• fork: The fork of CPython to benchmark. If benchmarking a pull request, this would normally be your GitHub username.
• ref: The branch, tag or commit SHA to benchmark. If a SHA, it must be the full SHA, since finding it by a prefix is not supported.
• machine: The machine to run on. One of linux-amd64 (default), windows-amd64, darwin-arm64 or all.
• benchmark_base: If checked, the base of the selected branch will also be benchmarked. The base is determined by running git merge-base upstream/main $ref.
• pystats: If checked, collect the pystats from running the benchmarks.

To watch the progress of the benchmark, select it from the 🔒 benchmark action page. It may be canceled from there as well. To show only your benchmark workflows, select your GitHub ID from the "Actor" dropdown.

When the benchmarking is complete, the results are published to this repository and will appear in the complete table. Each set of benchmarks will have:

• The raw .json results from pyperformance.
• Comparisons against important reference releases, as well as the merge base of the branch if benchmark_base was selected. These include
  • A markdown table produced by pyperf compare_to.
  • A set of "violin" plots showing the distribution of results for each benchmark.

The most convenient way to get results locally is to clone this repo and git pull from it.

Running benchmarks from the GitHub CLI

To automate benchmarking runs, it may be more convenient to use the GitHub CLI. Once you have gh installed and configured, you can run benchmarks by cloning this repository and then from inside it:

gh workflow run benchmark.yml -f fork=me -f ref=my_branch

Any of the parameters described above are available at the commandline using the -f key=value syntax.

Collecting Linux perf profiling data

To collect Linux perf sampling profile data for a benchmarking run, run the _benchmark action and check the perf checkbox. Follow this by a run of the _generate action to regenerate the plots.

License

This repo is licensed under the BSD 3-Clause License, as found in the LICENSE file.