Skip to content

understanding

Thomas Herault edited this page Feb 7, 2017 · 3 revisions

Understanding the compilation/runtime process of PaRSEC

The goal of this page is to give the PaRSEC-Enabled Operation Developper (someone who codes her operation using PaRSEC), or the PaRSEC user (someone who develops an application that includes one or more PaRSEC-Enabled Operations), a global view of the different steps. They are summarized by the image below:

Compilation process

The compilation process starts either with a JDF, or with a Serial code.

Using the PaRSEC compiler, the developper of a PaRSEC-Enabled operation can generate the JDF for a Serial Code. This process is described in How to write a PaRSEC-Enabled operation using the PaRSEC compiler.

Once a JDF is obtained, it can be edited by the PaRSEC-Enabled Operation Developper, and it passes through the JDF translator to build a C file, see How to write a PaRSEC enabled operation by editing the JDF file directly, section "Compiling the JDF".

This C file defines a set of functions. The PaRSEC-Enabled Operation User links her code with the PaRSEC runtime library and the compiled version of this file, to be enable to call the PaRSEC-Enabled Operation. See How to write a program that uses a PaRSEC-Enabled Operation to study this is part is done.