Bottleneck vs Pythran in 2026

Bottleneck and Pythran compared with PyOverdrive on purpose, pricing, setup, daily use and questions to ask.

By SayDeploy · 2026-10-11

SayDeploy is published by LunarWerx, which makes some of the tools compared here.

Bottleneck's own page calls it a set of fast array functions for NumPy, written in C ^[1]. Pythran is an ahead of time compiler for a subset of Python with a focus on scientific computing ^[2]. PyOverdrive patches supported NumPy functions so existing code uses faster paths, while unsupported calls run on stock NumPy ^[3]. Bottleneck's sources describe accelerated int32, int64, float32 and float64 arrays ^[1]. Pythran's sources describe use of multi-cores and SIMD instruction units ^[2]. PyOverdrive's sources describe self checks, calibration and controls for its fast paths ^[3] ^[3] ^[3].

Option	License	Price	Setup
Bottleneck	Simplified BSD license ^[1]	not published	Binary wheels on PyPI for the most common platforms ^[1]; source install requires Python >3.9 and NumPy 1.16.0+ ^[1]
Pythran	BSD-3-Clause license ^[2]	not published	Supports Python 3.7 and upward ^[2]
PyOverdrive	Open source under the MIT License ^[3]	not published	Install with pip straight from the GitHub repository ^[3]; needs Python 3.12 or newer ^[3]

The short answer

Bottleneck is a collection of fast NumPy array functions written in C ^[1]. Pythran is an ahead of time compiler for a subset of the Python language, with a focus on scientific computing ^[2]. PyOverdrive lets existing code use faster paths after calling pyoverdrive.enable(), while calls it does not support run on stock NumPy ^[3].

Bottleneck's sources say only int32, int64, float32 and float64 arrays are accelerated ^[1]. Pythran's sources say it supports Python 3.7 and upward ^[2]. PyOverdrive's sources describe Python 3.12 or newer ^[3], an OS independent package and a test suite run on Windows and Linux x86-64 ^[3].

What Bottleneck, Pythran is built for

Bottleneck's own page describes it as a collection of fast NumPy array functions written in C ^[1]. Its page says only arrays with data type int32, int64, float32 and float64 are accelerated ^[1]. It also says other dtypes use slower unaccelerated functions ^[1]. Bottleneck is distributed under a Simplified BSD license ^[1].

Pythran's own page describes it as an ahead of time compiler for a subset of the Python language, with a focus on scientific computing ^[2]. Its page says Pythran takes advantage of multi-cores and SIMD instruction units ^[2]. Pythran is BSD-3-Clause licensed ^[2].

What PyOverdrive is built for

PyOverdrive is made by the team that publishes this site.

PyOverdrive lets a developer call pyoverdrive.enable() to patch supported NumPy functions so existing code uses faster paths, and unsupported calls run on stock NumPy ^[3]. A fast path runs only when a check confirms the input is in its measured range, and anything else or a fast path that raises an error goes to stock NumPy ^[3].

PyOverdrive can compare every fast path with stock NumPy on the developer's own machine through python -m pyoverdrive --selfcheck ^[3]. A developer can run python -m pyoverdrive --calibrate to retime threaded paths at their thresholds on the CPU and switch off paths that do not pay there ^[3].

PyOverdrive can show which path a call would take and why without running the call through pyoverdrive.explain() ^[3]. A developer can turn off one fast path with disable_path() or PYOVERDRIVE_DISABLE, or restore original NumPy functions with pyoverdrive.disable() ^[3].

Pricing and fees

Bottleneck is distributed under a Simplified BSD license ^[1]. Pythran is BSD-3-Clause licensed ^[2]. The facts provided for Bottleneck and Pythran do not state paid plans or prices, so prices are not published here.

PyOverdrive is open source under the MIT License ^[3]. The product claims provided here do not state a paid plan or price, so a paid price is not published here.

Setting it up and daily use

Bottleneck provides binary wheels on PyPI for the most common platforms ^[1].

Pythran's page says it supports Python 3.7 and upward ^[2]. Its page says it takes advantage of multi-cores and SIMD instruction units ^[2]. The sources here say nothing more about its setup or routine use.

PyOverdrive is installed with pip straight from the GitHub repository ^[3]. It needs Python 3.12 or newer ^[3]. A developer can run python -m pyoverdrive --demo to time headline operations with stock NumPy and PyOverdrive on the developer's machine, in about 20 seconds ^[3].

Things to check before you choose

Check which kind of NumPy work needs support and whether the tool's stated supported cases match the code. Bottleneck's facts describe supported array dtypes ^[1]. PyOverdrive's claims describe supported paths and fallback behavior ^[3].

Check the Python version and installation method required by the tool. PyOverdrive needs Python 3.12 or newer ^[3] and installs with pip from its GitHub repository ^[3].

Check which pricing and account details matter for the project, because the provided facts do not state paid plans or account requirements for these tools.

How to choose

If the need is a collection of fast NumPy array functions written in C, Bottleneck's page describes that purpose ^[1].

If the need is an ahead of time compiler for a subset of Python focused on scientific computing, Pythran's page describes that purpose ^[2].

If the need is to patch supported NumPy functions while keeping unsupported calls on stock NumPy, PyOverdrive's claims describe that workflow ^[3]. If the need is to inspect or control fast paths, PyOverdrive's claims describe explain, calibration and disabling options ^[3] ^[3] ^[3].

Frequently asked questions

Is Bottleneck or Pythran better for NumPy acceleration?

Bottleneck's page describes a collection of fast NumPy array functions written in C ^[1]. Pythran's page describes an ahead of time compiler for a subset of Python with a focus on scientific computing ^[2]. The right choice depends on whether the described purpose matches the code and workflow being considered.

What does PyOverdrive do for NumPy code?

PyOverdrive patches supported NumPy functions after calling pyoverdrive.enable(), so existing code uses faster paths and unsupported calls run on stock NumPy ^[3]. It can also explain which path a call would take and why without running the call ^[3].

Does Bottleneck support every NumPy array type?

Bottleneck's page says only arrays with data type int32, int64, float32 and float64 are accelerated ^[1]. Its page says other dtypes use slower unaccelerated functions ^[1]. Check the array types used by a project against these stated types.

How do you install PyOverdrive?

PyOverdrive is installed with pip straight from the GitHub repository ^[3]. It needs Python 3.12 or newer ^[3]. The package is marked OS independent, and its test suite has been run on Windows and Linux x86-64 ^[3].

Can PyOverdrive disable an acceleration path?

PyOverdrive lets a developer turn off one fast path with disable_path() or PYOVERDRIVE_DISABLE, or call pyoverdrive.disable() to restore the original NumPy functions ^[3]. Its sources also describe checks that send unsupported cases back to stock NumPy ^[3].

It is open source under the MIT License. ^[3]

Try PyOverdrive