bm-20241026-3.14.0a1+-f6cc7c8-NOGIL

Results

• fork: python
• version: 3.14.0a1+
• config: NOGIL
• commit hash: f6cc7c8
• commit date: 2024-10-26T23:40:31+02:00
• commit merge base: 26d627779f79d8d5650fe7be348432eccc28f8f9
• ref: f6cc7c8bd01d8468af70

linux aarch64 (arminc)

• GitHub Action run
• cpu model: missing
• platform: Linux-5.15.0-101-generic-aarch64-with-glibc2.35
• raw results

vs. 3.10.4

• Geometric mean: 1.26x slower (HPT: reliability of 100.00%, 1.19x slower at 99th %ile)
• Memory usage: 1.50x
• missing benchmarks: aiohttp, asyncio_tcp, asyncio_tcp_ssl, chameleon, dask, flaskblogging, gunicorn, mypy2, pickle, pickle_dict, pickle_list, sqlite_synth, unpickle, unpickle_list
• new benchmarks: async_tree_cpu_io_mixed_tg, async_tree_io_tg, async_tree_memoization_tg, async_tree_none_tg, bpe_tokeniser, sphinx
• 📄table
• 📈time plot

vs. 3.12.0

• Geometric mean: 1.61x slower (HPT: reliability of 100.00%, 1.39x slower at 99th %ile)
• Memory usage: 1.20x
• missing benchmarks: aiohttp, asyncio_tcp, asyncio_tcp_ssl, chameleon, dask, gunicorn, mypy2, pickle, pickle_dict, pickle_list, sqlite_synth, unpickle, unpickle_list
• new benchmarks: bpe_tokeniser, sphinx
• 📄table
• 📈time plot

vs. 3.13.0

• Geometric mean: 1.63x slower (HPT: reliability of 100.00%, 1.41x slower at 99th %ile)
• Memory usage: 1.31x
• missing benchmarks: chameleon, connected_components, gevent_hub, k_core, mypy2, shortest_path
• new benchmarks: dulwich_log
• 📄table
• 📈time plot

vs. base

• Geometric mean: 1.53x slower (HPT: reliability of 100.00%, 1.40x slower at 99th %ile)
• Memory usage: 1.16x
• 🧠memory plot
• 📄table
• 📈time plot

linux x86_64 (linux)

• GitHub Action run
• cpu model: Intel(R) Xeon(R) W-2255 CPU @ 3.70GHz
• platform: Linux-5.4.0-196-generic-x86_64-with-glibc2.31
• raw results

vs. 3.10.4

• Geometric mean: 1.09x slower (HPT: reliability of 100.00%, 1.04x slower at 99th %ile)
• Memory usage: 1.44x
• missing benchmarks: aiohttp, asyncio_tcp, asyncio_tcp_ssl, chameleon, dask, djangocms, flaskblogging, gunicorn, mypy2, pickle, pickle_dict, pickle_list, sqlite_synth, unpack_sequence, unpickle, unpickle_list
• new benchmarks: async_tree_cpu_io_mixed_tg, async_tree_io_tg, async_tree_memoization_tg, async_tree_none_tg, bpe_tokeniser, sphinx
• 📄table
• 📈time plot

vs. 3.12.0

• Geometric mean: 1.41x slower (HPT: reliability of 100.00%, 1.25x slower at 99th %ile)
• Memory usage: 1.25x
• missing benchmarks: aiohttp, asyncio_tcp, asyncio_tcp_ssl, chameleon, dask, gunicorn, mypy2, pickle, pickle_dict, pickle_list, sqlite_synth, unpack_sequence, unpickle, unpickle_list
• new benchmarks: bpe_tokeniser, genshi_text, genshi_xml, html5lib, pylint, sphinx, thrift
• 📄table
• 📈time plot

vs. 3.13.0

• Geometric mean: 1.50x slower (HPT: reliability of 100.00%, 1.32x slower at 99th %ile)
• Memory usage: 1.29x
• missing benchmarks: chameleon, connected_components, gevent_hub, k_core, mypy2, shortest_path
• 📄table
• 📈time plot

vs. base

• Geometric mean: 1.48x slower (HPT: reliability of 100.00%, 1.32x slower at 99th %ile)
• Memory usage: 1.15x
• 🧠memory plot
• 📄table
• 📈time plot

linux x86_64 (pythonperf2)

• GitHub Action run
• cpu model: 12th Gen Intel(R) Core(TM) i9-12900
• platform: Linux-5.15.0-105-generic-x86_64-with-glibc2.35
• raw results

vs. 3.10.4

• Geometric mean: 1.15x slower (HPT: reliability of 100.00%, 1.07x slower at 99th %ile)
• Memory usage: 1.47x
• missing benchmarks: aiohttp, asyncio_tcp, asyncio_tcp_ssl, chameleon, dask, flaskblogging, gunicorn, mypy2, pickle, pickle_dict, pickle_list, sqlite_synth, unpack_sequence, unpickle, unpickle_list
• new benchmarks: async_tree_cpu_io_mixed_tg, async_tree_io_tg, async_tree_memoization_tg, async_tree_none_tg, bpe_tokeniser, sphinx
• 📄table
• 📈time plot

vs. 3.12.0

• Geometric mean: 1.48x slower (HPT: reliability of 100.00%, 1.27x slower at 99th %ile)
• Memory usage: 1.21x
• missing benchmarks: aiohttp, asyncio_tcp, asyncio_tcp_ssl, chameleon, dask, gunicorn, mypy2, pickle, pickle_dict, pickle_list, sqlite_synth, unpack_sequence, unpickle, unpickle_list
• new benchmarks: bpe_tokeniser, genshi_text, genshi_xml, html5lib, pylint, sphinx, thrift
• 📄table
• 📈time plot

vs. 3.13.0

• Geometric mean: 1.48x slower (HPT: reliability of 100.00%, 1.29x slower at 99th %ile)
• Memory usage: 1.31x
• missing benchmarks: chameleon, connected_components, gevent_hub, k_core, mypy2, shortest_path
• 📄table
• 📈time plot

vs. base

• Geometric mean: 1.39x slower (HPT: reliability of 100.00%, 1.33x slower at 99th %ile)
• Memory usage: 1.17x
• 🧠memory plot
• 📄table
• 📈time plot

darwin arm64 (darwin)

• GitHub Action run
• cpu model: missing
• platform: macOS-15.0.1-arm64-arm-64bit-Mach-O
• raw results

vs. 3.10.4

• Geometric mean: 1.13x slower (HPT: reliability of 100.00%, 1.10x slower at 99th %ile)
• Memory usage: 1.45x
• missing benchmarks: aiohttp, asyncio_tcp, asyncio_tcp_ssl, chameleon, dask, flaskblogging, gunicorn, mypy2, pickle, pickle_dict, pickle_list, sqlite_synth, unpack_sequence, unpickle, unpickle_list
• new benchmarks: async_tree_cpu_io_mixed_tg, async_tree_eager, async_tree_eager_cpu_io_mixed, async_tree_eager_cpu_io_mixed_tg, async_tree_eager_io, async_tree_eager_io_tg, async_tree_eager_memoization, async_tree_eager_memoization_tg, async_tree_eager_tg, async_tree_io_tg, async_tree_memoization_tg, async_tree_none_tg, bpe_tokeniser, sphinx
• 📄table
• 📈time plot

vs. 3.12.0

• Geometric mean: 1.33x slower (HPT: reliability of 100.00%, 1.16x slower at 99th %ile)
• Memory usage: 1.29x
• missing benchmarks: aiohttp, asyncio_tcp, asyncio_tcp_ssl, chameleon, dask, gunicorn, mypy2, pickle, pickle_dict, pickle_list, sqlite_synth, unpack_sequence, unpickle, unpickle_list
• new benchmarks: async_tree_eager, async_tree_eager_cpu_io_mixed, async_tree_eager_cpu_io_mixed_tg, async_tree_eager_io, async_tree_eager_io_tg, async_tree_eager_memoization, async_tree_eager_memoization_tg, async_tree_eager_tg, bpe_tokeniser, genshi_text, genshi_xml, html5lib, pylint, sphinx, thrift
• 📄table
• 📈time plot

vs. 3.13.0

• Geometric mean: 1.30x slower (HPT: reliability of 100.00%, 1.16x slower at 99th %ile)
• Memory usage: 6.64x
• missing benchmarks: chameleon, connected_components, dask, gevent_hub, k_core, mypy2, shortest_path
• 📄table
• 📈time plot

vs. base

• Geometric mean: 1.40x slower (HPT: reliability of 100.00%, 1.23x slower at 99th %ile)
• Memory usage: 1.11x
• 🧠memory plot
• 📄table
• 📈time plot