Google Summer of Code is an annual global program focused on bringing more student developers into open source software development. Students work with an open source organization on a 3 month programming project.
I applied to the program under the SciRuby organization with the project titled “Improving NMatrix: Adding features to NMatrix core”. I am thankful to SciRuby for giving me the opportunity to work on this project during summer as part of GSoC. I am also thankful to my mentor Prasun Anand for guiding me so far and for helping me complete my proposal in time.
Detailed proposal can be found here.
Brief introduction to NMatrix
NMatrix is a fast numerical linear algebra library for Ruby, with dense and sparse matrices, written mostly in C and C++ (and with experimental JRuby support). It supports dense matrices as well as two types of sparse (linked-list-based and Yale/CSR). NMatrix currently relies on ATLAS/CBLAS/CLAPACK and standard LAPACK for several of its linear algebra operations.
NMatrix’s backend is currently written in C and C++ with some parts of the backend in Ruby too. This has made NMatrix to be a bit slower for some operations and also some parts of code base are not much readable. The project is being re-implemented from scratch by SciRuby contributors here at https://github.com/prasunanand/nmatrix_reloaded with backend written completely in C. This will make the project more efficient and the code more uniform and readable.
Summary of my project outcomes:
- Adding support for sparse matrices (COO, CSR, CSC, Dia).
- Adding N-dimensional matrix support which includes indexing, iterating, slicing and broadcasting.
- Adding linear algebra support which includes implementation of BLAS and LAPACK routines.
- Integration with iruby notebooks which includes writing methods such as pretty_print and taking care of some use cases.
This is my first blog regarding my GSoC 2019 progress. I’ll be writing more such posts on this blog in near future. A blog post will be written each week giving a brief overview of the work done during that week.