[Oberon] PO2013 - Real time measurement
skulski at pas.rochester.edu
Mon Mar 25 03:09:07 CET 2019
> But of course the whole point of using a PIC, apart from adding the
> analogue interface(s) that simply don't exist on the FPGA, is to offload
> some of the processing (which is exactly what the PIC was designed for
> in the first place) so polling may well be fine as a solution.
Another solution may be adding more RISC5 processors to the same FPGA. Not possible with the original Spartan-3 200, but it is certainly possible with larger FPGAs. The extra CPUs do not need to be full RISC5, actually. They can be modified RISC5's without floating point. Adding one or a few PicoBlaze processors is another option.
A multicore design can be fun, I suppose. The communication among the cores was worked out by Professor Wirth in one of his papers available from his web site. A simple case can be handled by a shared register or a shared BRAM, depending on the amount of data to be exchanged. Polling is definitely an option here.
Concerning analog interfaces, these can be added with PMOD boards which are available from Digilent. They are simple slave boards interfaced with SPI and/or GPIO.
More information about the Oberon