Using lustre-mt to parallelize a Lustre program

This tutorial will illustrate :

1. How to create a task
2. How to use the tool lustre-mt to parallelize these tasks

First, we need a Lustre V6 program with a node call.

function A(in:int) returns (o:int);
let
  o = in*2;
tel

node para(x, y:int) returns (o:int);
var a, b:int;
let
  a = A(x);
  b = A(y);
  o = a+b;
tel

In order to parallelize tasks, we first need to declare them. We use the pragma %MT:task1% in front of a node call in order to do this.

function A(in:int) returns (o:int);
let
  o = in*2;
tel

node para(x, y:int) returns (o:int);
var a, b:int;
let
  a = A(x);
  b = %MT:task1% A(y);
  o = a+b;
tel

Now that we have declared a task, we need to compile with the --2c-multi-task (or -2cmt for short) option in order to generate C files with this task.

lv6  para_task.lus -n para --2c-multi-task

Compiling with the -2cmt option will also generate a Yaml file. The C files will not compile, and will require to generate another file using the external tool lustre-mt and this Yaml file.

lustre-mt para_task_para.yml

lustre-mt generates 2 files, called wraptask.c and wraptask.h, which contains functions to parallelize tasks. Now we can compile with these new files :

gcc -lpthread -o para.exec *.c

And finally execute the program with parallel tasks :

./para.exec