| Time | Implementation | Compiler | Benchmark date | SUPERCOP version |
| 51229 | faster | gcc -funroll-loops -mcpu=marvell-pj4 -O3 | 20161211 | 20161026 |
| 53152 | faster | gcc -mcpu=marvell-pj4 -O2 | 20161211 | 20161026 |
| 54044 | faster | gcc -funroll-loops -mcpu=marvell-pj4 -O2 | 20161211 | 20161026 |
| 57256 | faster | gcc -mcpu=marvell-pj4 -O3 | 20161211 | 20161026 |
| 58472 | faster | gcc -funroll-loops -mcpu=marvell-pj4 -Os | 20161211 | 20161026 |
| 58692 | faster | gcc -mcpu=marvell-pj4 -Os | 20161211 | 20161026 |
| 65618 | ref | gcc -funroll-loops -mcpu=marvell-pj4 -O3 | 20161211 | 20161026 |
| 70656 | ref | gcc -mcpu=marvell-pj4 -O3 | 20161211 | 20161026 |
| 81092 | ref | gcc -funroll-loops -mcpu=marvell-pj4 -O2 | 20161211 | 20161026 |
| 110758 | ref | gcc -mcpu=marvell-pj4 -O2 | 20161211 | 20161026 |
| 117560 | ref | gcc -funroll-loops -mcpu=marvell-pj4 -Os | 20161211 | 20161026 |
| 118920 | ref | gcc -mcpu=marvell-pj4 -Os | 20161211 | 20161026 |