Parallware is a source-to-source parallelizing compiler for sequential scientific programs written in the C programming language. ParallWare automatically discovers the parallelism available in the input sequential C code, and automatically generates parallel-equivalent C source code annotated with OpenMP compiler directives.
How does it work?
In order to maximize the performance of the OpenMP parallel program, focus on the part of the source code that takes most of the execution time of the whole program. First, isolate such hotspot in a single C function and, next, use Parallware to generate an equivalent parallel C function annotated with OpenMP pragmas. Check a live demo of usage of Parallware in Appentra’s youtube channel. In the technical blog, there are additional articles about automatic parallelization of scientific programs using linear algebra routines and well-known numerical simulation methods.
How can I try it?
Contact us and ask for a 30-day evaluation license of Parallware.
How much does it cost?
Contact us and we will help you to evaluate how Parallware can meet your HPC needs.