Occam-pi for Programming of Massively Parallel Reconfigurable Architectures

Table 2

Performance and estimated power results of autofocus criterion calculation.

Implementations	Latency (cycles)	Throughput (pixels/sec.)	Speedup throughput	Power (watts)
Sequential on Intel i7 @ 2.67 GHz	—	21,600	1	17.5
Design-I on Ambric @ 300 MHz	16,497	236,386	11	6.52
Design-II on Ambric @ 300 MHz	12,793	486,224	23	9.8