Table 5 Summary of synthesis results of original and optimized DCTs
From: A novel method for the approximation of multiplierless constant matrix vector multiplication
Objective | Optimization of the number of operations | ||||||||
DCTs | oper | step | ASIC | FPGA | |||||
A | D | P | LUTs | slices | D | P | |||
Original | 56 | 8 | 16.8 | 5.7 | 31.0 | 626 | 179 | 14.5 | 2201 |
Optimized | 54 | 6 | 16.0 | 5.4 | 27.6 | 595 | 170 | 14.1 | 2149 |
ε 1=ε ∞ =4 | |||||||||
Optimized | 48 | 6 | 13.8 | 4.7 | 21.1 | 525 | 151 | 13.3 | 2107 |
ε 1=ε ∞ =8 | |||||||||
Optimized | 46 | 6 | 13.1 | 4.7 | 19.5 | 504 | 144 | 13.1 | 1988 |
ε 1=ε ∞ =16 | |||||||||
Objective | Optimization under a delay constraint | ||||||||
DCTs | oper | step | ASIC | FPGA | |||||
A | D | P | LUTs | slices | D | P | |||
Original | 58 | 5 | 17.7 | 5.2 | 31.9 | 701 | 201 | 13.4 | 2079 |
Optimized | 54 | 5 | 16.4 | 5.1 | 28.8 | 612 | 177 | 13.3 | 2058 |
ε 1=ε ∞ =4 | |||||||||
Optimized | 48 | 5 | 14.1 | 4.7 | 22.9 | 542 | 156 | 12.8 | 2088 |
ε 1=ε ∞ =8 | |||||||||
Optimized | 46 | 5 | 13.1 | 4.5 | 19.5 | 508 | 145 | 12.2 | 1874 |
ε 1=ε ∞ =16 | |||||||||