Pyston Joins Anaconda to Expand Open-Source Python-based Project Development

Leading open-source data science innovation company Anaconda has tapped the entire team of Pyston to boost its open-source development of its Python ecosystem. If you are looking to enhance your Python, you usually do it using Pyston, the fastest way to speed up any Python code. In short, Pyston is a powerful and highly compatible fork that is used in the faster implantation of the Python programming language. As part of the deal, Anaconda will onboard Kevin Modzelewski and ‪Marius Wachtler from Pyston where the two professionals would continue to work independently on the Pyston-Numba projects and grow the open-source community. They will join Anaconda’s team of funded OSS developers, alongside other Anaconda contributors to Dask, Numba, Bokeh, and many other projects. Anaconda would maintain Pyston, as a stakeholder and as a sponsor.

Read Also: Don’t Let the Supply Chain Silence Your Marketing

At the time of this announcement, Anaconda confirmed in their official blog that Pyston is seen as a very powerful OS platform to quickly bring Python development to the “mainstream audience”. Pyston’s role as a fork of standard Python 3.8 / CPython interpreter could not only accelerate Python’s development in the open-source community but also improve the execution performance of all the current and future Python programs.

How Does Pyston Improve Python’s Execution Performance?

Pyston improves the Python program’s execution performance through various techniques. These include:

• Quickening-based Optimization: In-line caching via quickening to optimize call instructions provided to Python interpreter.
• Reference Counting
• Dynamic Assembler (DynASM) for code generation engines, and so on.

Pyston 2.x series is an entirely new, reworked from scratch version applied from a fork of CPython 3.8. The older version, Pyston 1.0 project was last used at Dropbox. Currently, Pyston 2.x series delivers speeds of up to 2x times on a wide range of Python programs developed in the open-source.

Top Blog: NVIDIA Turbocharges Extreme-Scale AI For Argonne National Laboratory’s Polaris Supercomputer

Pyston developers can install Pyston either through Anaconda or through Conda.

By tapping Pyston’s coding resources and DevOps team, Anaconda intends to grow its own offering to Python optimization across all Python use cases. Anaconda’s Numba compiler, its oldest Python project based on JIT compiler helps programmers to run numerical Python functions on a CPU / GPU. Pyston’s arrival will make this whole project faster and more efficient.

In fact, Numba and Pyston complement each other even as Pyston developers bring in a totally new way to look at problems associated with Python compilation for numerical uses cases. So, Numba and Pyston DevOps teams would work in sync to improve the numerical use cases of Python programming language specific to CPython.

For the open source developer community, the Anaconda – Pyston combination means one thing – it will add Anaconda’s expertise in data science engineering and infrastructure building to Pyston’s legacy of creating fast-paced Python use cases with superior testing, distribution and sharing opportunities in the OS. It is likely that Anaconda would use Pyston as an independent open source project to improve its own specific Python projects, though it has been confirmed that Pyston and Anaconda projects would be clearly distinguished from each other in the open-source market.

Register Now: Everything You Want To Know About Salesforce’s Dreamforce 2021

Currently, there are over 25 million Anaconda users who account for 2.4 billion package downloads in 2019. Founded in 2012 when Python was still in its infancy, the Anaconda team was quickly amped up within the next 2 years with new releases – Conda, PyData, NumFOCUS, Numba, Bokeh and Blaze. Today, Anaconda Enterprise is a highly revered data science platform that supports many Python projects built to rewrite the history of AI Machine Learning and Deep Learning applications within the enterprise architecture.

Source: Anaconda

[To share your insights with us, please write to sghosh@martechseries.com]