The Joint Python Accelerators project gathers developers of Transonic, Cython, Pythran and Numba namely:
- Pierre Augier (Transonic)
- Serge Guelton (Pythran)
- Stefan Behnel (Cython)
- Stanley Seibert (Numba)
This project has been initiated by core-developers of scikit-image (Juan Nunez-Iglesias, Emmanuelle Gouillart and Stéfan van der Walt).
The joint Python Accelerators project gathers developers of Transonic, Cython, Pythran, Numba and scikit-image. Our collaboration aims at improving the state of Python acceleration so that clean and modern Python code can easily be accelerated to reach high performance.
We propose to improve interoperability and compatibility between existing accelerators. We will improve the integration and feature coverage of the Pythran support in Cython, which will impact much existing Cython code.
We will base our work on Transonic, a new package that accelerates the same code with different accelerators, using just-in-time and ahead-of-time compilations. We propose to work on usage of Transonic in scikit-image, which is a good example of a widely used library relying a lot on Cython while its developers would love to be able to write simpler Python and to also use Numba and Pythran as accelerators.
While working on scikit-image code, we will improve its maintainability and performance. We will also improve the accelerators by fixing bugs, implementing missing features and increasing performance. This will have a direct impact on the life sciences through the improvement of many downstream Python packages.
More generally, this project will greatly improve the state of the scientific Python ecosystem. There will be one tool adapted for both developers of fundamental libraries (like scikit-image) and end users (such as scientists and students), with a clean API and good documentation. Moreover, this project will launch long-term dynamics on performance for scientific Python, based on compatibility, interoperability and gentle competition between the accelerators.
See also the detailed work plan.
Here is the call for proposals for the CZI program:
https://chanzuckerberg.com/rfa/essential-open-source-software-for-science/
Lots of info on that site. There is also a SciPy 2019 talk about the program by Jeremy Freeman, who leads the computational biology division at CZI:
https://www.youtube.com/watch?v=Zu_MmxdEW0M
For administrative and programmatic inquiries pertaining to this RFA, please contact sciencegrants@chanzuckerberg.com.
This proposal has been submitted on Feb. 4 2020 (application ID number EOSS2-0000000137).
Portal opening date Mid-December 2019
Application due date February 3, 2020
Notification of decisions Late April 2020
Earliest start date of grant June 1, 2020
The Chan Zuckerberg Initiative uses SurveyMonkey Apply (SMApply) as its grants management portal. All applications must be submitted through this portal (https://apply.chanzuckerberg.com). SMApply is configured to work best using the Google Chrome browser. It is recommended that you familiarize yourself with this portal well in advance of any deadlines. Deadline extensions will not be granted.
For the purpose of the application, we will use the following terms:
-
Applicant: The person submitting the application materials on behalf of the software project(s)
-
Organization: The organization directly receiving and distributing funding
-
Software Project(s): The software project(s) that will be supported by the funding
-
Proposal: The proposed use of funding
-
Key Personnel: People involved in the software project(s) described in the proposal and supported by the funding
The application consists of the following sections (called tasks in the grants portal): Applicant Details Part I, Applicant Details Part 2, Organization Details, Proposal Details, Optional Attachments, CV of Applicant, Budget Description, Number of Open Source Software Project(s), Open Source Software Project Details