The Co-dfns project aims to provide a high-performance, high-reliability compiler for a parallel extension of the Dyalog dfns programming language. The dfns language is a functionally oriented, lexically scoped dialect of APL. The Co-dfns language extends the dfns language to include explicit task parallelism with implicit structures for synchronization and determinism. The language is designed to enable rigorous formal analysis of programs to aid in compiler optimization and programmer productivity, as well as in the general reliability of the code itself.
Our mission is to deliver scalable APL programming to information and domain experts across many fields, expanding the scope and capabilities of what you can effectively accomplish with APL.
We are seeking to create an open funding model for Co-dfns research through user and patron contributions. You can support the project by contributing code, feedback, benchmarks, and so forth, but you can also directly support the Co-dfns project by funding the author:
https://www.patreon.com/arcfide
There are a number of related initiatives that are based on the Co-dfns technology:
- Mystika: a high-level, high-performance cryptographic stack
- apixlib: programmable, easy to use image processing
Live Reading/Writing Co-dfns, The Way of APL
Does APL Need a Type System? (FnConf '18)
Array-oriented Functional Programming
Tree Manipulation Workshop and Dyalog '18 Talk
APL Patterns vs. Anti-Patterns @ FunctionalConf 2017
APL Style: Patterns and Anti-patterns
Co-dfns Compiler Architecture and Design (Video)
The Key to a Data Parallel Compiler
Accelerating Information Experts Through Compiler Design
Co-dfns: Ancient Language, Modern Compiler
U11: Using Co-dfns to Accelerate APL Code
U07: Co-dfns Report: Performance and Reliability Prototyping