Project developers Pyston offering high-performance implementation of Python language using modern Jit compilation technologies, pyston 2.2 And announced the return of the project to the number of open. The implementation is aimed at achieving high performance close to the performance of traditional systemic languages, such as C ++. Pyston 2 branch code Published on GitHub under the PSFL license (Python Software Foundation License), similar to CPYthon license.
Recall that earlier the project Pyston oversaw Dropbox, which in 2017 ceased to finance the development. The Pyston developers founded their company and released a substantially recycled Pyston 2 branch, which was declared stable and ready for widespread use. At the same time, the developers stopped publishing the source texts and moved to the provision of only binary assemblies. Now Pyston decided to make an open project again, and transfer to the business model associated with the development of open software. Moreover, the possibility of transferring optimization from Pyston to a regular CPYTHON.
It is noted that Pyston 2.2 in performance tests that assess the loads inherent applications for the Web server are faster than regular Python by 30%. There is also a significant increase in Pyston 2.2 performance compared to previous issues, which managed to achieve mainly due to the addition of optimization for new areas, as well as improved JIT and caching mechanisms.
In addition to optimizing performance, the new release is also interesting for the transfer of changes from the CPYthon 3.8.8 branch. From the point of view of compatibility with Python, the Pyston project is presented as the most compatible CPYthon alternative implementation, since Pyston is a branch from the main CPYTHON code base. All CPYTHON features are supported in Pyston, including the C API to develop extensions in the SI language. Among the main differences between Pyston from CPYTHON highlights the use of dynasm jit , Inline caching and common optimizations.
From changes to Pyston 2.2 is also mentioned cleaning the code base from many debugging capabilities of CPYthon, which negatively affect performance, but at the same time developers are almost not claimed. Statistics are given in accordance with which the removal of debugging tools leads to acceleration of operation by 2%, while only about 2% of developers use these functions.