I've been working with a system that uses a P2020. We began with only data collection but then added embedded processing, requiring double-precision floating-point calculations. Things are working but it's slow, and we would like to upgrade the system. The company that builds the system has suggested switching to a board that uses the P4080.
I've seen the FPU on the P4080 described as "classic," whatever that means. Can anyone give me a rough idea of what kind of performance difference I would see on the DP FP calculations? We're running LInux with RT patches. The main program is multi-threaded. Most of the FP work is done in a series of separate (single-threaded) OS processes.