
Multicore processors will not execute your application any faster than single core processors
unless your application is parallelized for multicores. Parallelization is typically a manual process that has proven to be a tough job in many cases. It is often difficult to figure out where to parallelize and which dependencies are present. Making mistakes leads to disappointing speedups or incorrect results.
Biface introduces Parallelization Assistant that helps programmers with parallelization. It identifies where parallelization is likely to be most beneficial and which dependencies are present. Dependencies are clearly marked in the source code. With Parallelization Assistant programmers are able to parallelize code in less time or to deliver better performance results.
Parallelization Assistant is based on source-to-source compiler technology that instruments the source code so that certain information gets collected when the application is executed. The collected information is analyzed and presented to the user via a graphical user interface. With the presented information the user gets the information that he needs to parallelize the application.
Parallelization Assistant supports C with many gcc language extensions. Target platforms are Linux on ARM Cortex-A as well as Android on ARM Cortex-A.
Parallelization Assistant does not do the parallelization for you. It does it with you!