eSSENCE@LU 5:10 - Programmable hardware platform for scientific software

As detectors and sensor technologies evolve, X-ray laboratories and their users are faced with the problem to efficiently process very large data volumes. With high data rates, traditional methods of storing data first, and processing later are a serious bottleneck. And so increasing attention is given to a real-time data analysis.

Massively parallel computing architectures as graphical processing units have demonstrated to be very efficient in processing large datasets. Recent advances in reconfigurable hardware as field-programmable gate arrays (FPGAs) and evolution in high level languages for their programming brought these hardware accelerators to computing centers. They already proved to be very reliable and energy efficient there. Their big potential in bridging high data rate acquisition and high performance computing should be utilized for moderating data flows from fast X-ray cameras.

In order to research capabilities and advantages of the new generation reconfigurable hardware in connection with high performance computing environment we are developing a software toolbox for processing X-ray synchrotron data on them.