cheukwing / hashkell

Semi-automatic Parallelisation of Timed Functional Programs

hashkell

Parser for a simple subset of Haskell with time complexity annotations and dependency graphs.

Compilation, running, and testing

This project uses Stack:

stack build # to build

stack run -- <args> # or ./hashkell-exe <args>, to run with <args>

stack test # to test

Usage

To parallelise an input file code.hs with the func method:

stack run -- parallelise func code.hs

To graph all the functions in an input file code.hs:

stack run -- graph -A code.hs

For all options, refer to the help manual:

stack run -- --help

Example of Supported Code

fib ## 2^x;
fib :: Int -> Int;
fib x =
  if x < 1
    then 0
    else if x < 2
           then 1
           else fib (x - 1) + fib (x - 2)

Generated Dependency Graph

Time Complexity Annotations

Time complexity annotations are used by the compiler to generate branching paths for parallelisation. Annotations are written using ##, as so:

<function-name> ## <time-complexity>;

The following annotations are supported:

-- Constant time
func ## 1;

-- Polynomial time
func ## n;

func ## n^2;

func ## n^100;

-- Exponential time
func ## 2^n;

func ## 100^n;

-- Logarithmic time
func ## log n;

-- Factorial time
func ## fac n;

The given identifier name in the annotation must correspond to an argument in the function declaration, and only one such name is allowed in any annotation.