Futhark (programming language)
| Futhark | |
|---|---|
| Paradigm | array, functional |
| Family | ML |
| Designed by | Troels Henriksen, Cosmin Oancea, Martin Elsman |
| Developer | University of Copenhagen[1] |
| First appeared | 2014 |
| Typing discipline | inferred, static, strong, Hindley–Milner, uniqueness, dependent |
| OS | cross-platform |
| License | ISC |
| Website | futhark-lang |
| Influenced by | |
| APL, Haskell, NESL, Standard ML | |
Futhark is a multi-paradigm, high-level, functional, data parallel, array programming language. It is a dialect of the language ML, originally developed at UCPH Department of Computer Science (DIKU) as part of the HIPERFIT project.[2] It focuses on enabling data parallel programs written in a functional style to be executed with high performance on massively parallel hardware, especially graphics processing units (GPUs). Futhark is strongly inspired by NESL, and its implementation uses a variant of the flattening transformation[3], but extended to efficiently support the special case of uniform nested data parallelism[4]. The Futhark compiler is free and open-source software released under an ISC license.
Overview
[edit]Futhark is a language in the ML family, with an indentation-insensitive syntax derived from OCaml, Standard ML, and Haskell. The type system is based on a Hindley–Milner type system with a variety of extensions, such as uniqueness types and size-dependent types. Futhark is not intended as a general-purpose programming language for writing full applications, but is instead focused on writing compute kernels (not always the same as a GPU kernel) which are then invoked from applications written in conventional languages.[5]
Futhark is named after the first six letters of the Runic alphabet.[6]: 2
Examples
[edit]Dot product
[edit]The following program computes the dot product of two vectors containing double-precision numbers.
def dotprod xs ys = f64.sum (map2 (*) xs ys))
It can also be equivalently written with explicit type annotations as follows.
def dotprod [n] (xs: [n]f64) (ys: [n]f64) : f64 = f64.sum (map2 (*) xs ys))
This makes the size-dependent types explicit: this function can only be invoked with two arrays of the same size, and the type checker will reject any program where this cannot be statically determined.
Matrix multiplication
[edit]The following program performs matrix multiplication, using the definition of dot product above.
def matmul [n][m][p] (A: [n][m]f64) (B: [m][p]f64) : [n][p]f64 =
map (\A_row ->
map (\B_col -> dotprod A_row B_col)
(transpose B))
A
This shows how the types enforce that the function is only invoked with matrices of compatible size. Also, it is an example of nested data parallelism.
References
[edit]- ↑ "License". futhark-lang.org. Retrieved 2023-03-26.
Developed at DIKU
- ↑ "Home". hiperfit.dk.
- ↑ Henriksen, Troels; Serup, Niels G. W.; Elsman, Martin; Henglein, Fritz; Oancea, Cosmin (2017). "Futhark: Purely Functional GPU-Programming with Nested Parallelism and In-Place Array Updates" (PDF). Proceedings of the 38th ACM SIGPLAN Conference on Programming Language Design and Implementation. PLDI 2017. ACM.
- ↑ "Full flattening of nested data parallelism". The Futhark Programming Language. DIKU. 2026-07-31. Retrieved 2026-11-16.
- ↑ "Futhark User's Guide". futhark.readthedocs.io.
- ↑ Troels, Henriksen (November 2017). Design and Implementation of the Futhark Programming Language (PDF) (PhD thesis). University of Copenhagen. Retrieved 2024-05-25.
- Programming languages
- High-level programming languages
- Functional languages
- Parallel computing
- Array programming languages
- Dependently typed languages
- Dependently typed programming
- Statically typed programming languages
- ML programming language family
- Free and open source compilers
- 2014 software
- Programming languages created in 2014
- Software using the ISC license