Chatroom: https://gitter.im/ocramz/sparse-linear-algebra
Issues: https://github.com/ocramz/sparse-linear-algebra/issues
sparse-linear-algebra is a young and experimental library, but it aims to provide high-performance and high-level functionality.
A combination of approaches is necessary to meet these ambitious goals:
- Performance
- Ergonomics
- Applications
This is the trickiest bit (especially in Haskell). The initial version was definitely not performant, which motivated a
complete rewrite and reorganization of the backend (which will be split into sparse-linear-algebra-accelerate and possibly also
sparse-linear-algebra-vector).
Got an idea for code generation of a certain method? Would you like to have a repa backend or perhaps have something to say
about how a method is implemented? Don't esitate to join the discussion.
One of the first aims of this library was to bridge the gap between scientific computing textbooks and application code. We often think in math, but must talk to computers; the translation doesn't need to be a traumatic experience! To all prospective users and contributors: do try the library out! Is there something missing, or does something feel wrong or clunky? Please get in touch and share all your thoughts !
This library aims to being used in the "real world", so all feedback from developers of applications and wrapping libraries is more than welcome. Anything from performance regressions, or un-exposed functionality such as exceptions etc. Shoot us a line, or even better a pull request!