Abstract :
The concept underlying the UniBoard was originally proposed by Sergei Pogrebenko, system scientist at the Joint Institute for VLBI in Europe (JIVE, where VLBI stands for Very Long Baseline Interferometry), in the late 90s. As one of the architects of the MarkIV correlator [1], he knew like no other the difficulties and technological pitfalls involved in the construction of such a complex instrument, consisting of many separate custom-made electronic components. A single-board, all-station correlator would do away with the need to transfer large, highly synchronized data streams between all these different components, with their associated complicated messaging systems and timing issues. Such a board should have all the CPU power that could be fitted on, in the form of Field Programmable Gate Arrays (FPGA), and as much I/O capacity as possible. However, in those days FPGA technology was not far enough advanced to make this feasible. When the preparation for the RadioNet FP7 proposal got underway in 2007, the situation had changed completely, with new generations of FPGAs combining massive computing power with ease of programming and fast development. The concept now became the basis for the UniBoard, one of the Joint Research Activities in the RadioNet FP7 programme (EC Contract no. 227290).
Keywords :
astronomy computing; field programmable gate arrays; radioastronomical techniques; radioastronomy; radiowave interferometry; CPU power; FPGA technology; I/O capacity; JIVE; Joint Institute for VLBI in Europe; RadioNet FP7 programme; UniBoard; Very Long Baseline Interferometry; computing power; electronic components; field programmable gate arrays; multipurpose scalable high-performance computing platform; radioastronomical applications; single-board all-station correlator; Backplanes; Correlators; Educational institutions; Field programmable gate arrays; Hardware; Microprogramming; Receivers;