3 ms·FPGA-accelerated ML inference as a service for particle physics computing3 points by matt_d 7y ago