[Idea]: develop C implementations for base special mathematical functions

[kgryte]

Idea

This idea builds on the work outlined in stdlib-js/stdlib#649. Namely, implementing base special mathematical functions in C. Currently, all special mathematical functions have JavaScript implementations, which are often ports from other languages.

The goal of this idea is to port all JavaScript implementations to C. Having such implementations will allow stdlib to provide Node.js native add-ons for higher performance ndarray computation and is more generally necessary for achieving NumPy/SciPy parity.

Expected Outcomes

Users will be able to leverage C implementations for use in Node.js native add-ons, and stdlib will be able to expose element-wise APIs for evaluating base special math functions over ndarrays.

Involved Software

No other software is necessary.

Prerequisite Knowledge

C, JavaScript, Node.js.

Difficulty

Intermediate. Familiarity with C is beneficial. This idea mainly involves porting existing implementations (many of which are written in C/C++) and doing so in a manner which conforms with stdlib conventions.

Project Length

90/175/350 hours. Can be scoped accordingly. Scope can be expanded to implement new special mathematical functions.

Potential Mentors

@kgryte @Planeshifter @steff456 @rreusser @Pranavchiku @czgdp1807

[kgryte]

If you are interested in working on this idea, you are strongly advised to see stdlib-js/stdlib#649 and determine what work remains. You should also spend time reading the source code of any upstream reference implementations (e.g., msun, openlibm, cephes, boost, etc). Some implementations are likely to be fairly involved, so you need to have a good understanding and familiarity with implementing C mathematical routines.

[kgryte]

Additionally, you are encouraged to work on C implementations of math functions in advance of submitting your proposal in order to demonstrate that you can successfully execute on this project.