Appentra is a startup promoted by PhD Manuel Arenaz after ten years of research in the area of parallelizing compilers as a member of the Computer Architecture Group of the University of Coruña (UDC). A technology transfer agreement with the UDC enabled the creation of Appentra in July 2012. The Appentra´s team wants to celebrate the milestones achieved during its first two years of activity, and the best way to do it is to show the timeline of these milestones.
During this 2-year period, we have also received some awards that recognize the business-orientation and technology-disruptiveness of the project. These awards are “Best Young Company” in the XIII Edition of the Innovative Business Project Competition at University of Santiago de Compostela (Uniemprende 2013), “Accesit to innovation” in the XIV Premio AMJE Emprende 2013 from AMJE Coruña (december 2013), and “Technology Transfer Award” from the European Network on High Performance and Embedded Architecture and Compilation (december 2013).
Now we are very motivated to continue learning about our target market, and we have already begun to make prospections to find our first customers. Thus, we have recently presented Parallware in the HPC Advisory Council Spain Conference 2014. From now on, we will be presenting the competitive advantages of our tool in world-wide events specialized in the Oil&Gas industry and in computational electromagnetics (CEM).
Here are some testimonials reported about Parallware:
We were pretty impressed by the tool, because parallelism extraction is not a simple problem to solve.
I do not see automatic parallelization of the outer loop in the MOM matrix fill as a trivial task. Thus, for a compiler to do this automatically is quite impressive.
With Parallware I can develop multiple versions of complex C source codes for real electromagnetic simulations, and in each new functionality or updating I do not have to worry about the tedious, error-prone manual parallelization.
Parallware automatically extracts the parallelism implicit in most of interest cases in CEM. In those situations where the current version of Parallware does not extract the parallelism yet, the Appentra team makes a great work in tuning the code for Parallware or in the manual parallelization of the most complex parts of the code. They also work hard in improving Parallware technology with more sophisticated techniques for automatic detection of parallelism.
Finally, we would like to thank to all the people we have met during these two years. They have all provided us with very valuable feedback about the project and, overall, they have all made possible for us to reach this point.
Leave a Reply