| Time | Implementation | Compiler | Benchmark date | SUPERCOP version |
| 644 | ref | gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2 | 20161214 | 20161026 |
| 650 | ref | gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3 | 20161214 | 20161026 |
| 658 | ref | gcc -mcpu=native -mfpu=neon-vfpv4 -O3 | 20161214 | 20161026 |
| 659 | ref | gcc -mcpu=native -mfpu=neon-vfpv4 -O2 | 20161214 | 20161026 |
| 705 | ref | gcc -mcpu=native -mfpu=neon-vfpv4 -Os | 20161214 | 20161026 |
| 726 | ref | gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os | 20161214 | 20161026 |
| 858 | armneon2 | gcc -mcpu=native -mfpu=neon-vfpv4 -O2 | 20161214 | 20161026 |
| 860 | armneon2 | gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3 | 20161214 | 20161026 |
| 862 | armneon2 | gcc -mcpu=native -mfpu=neon-vfpv4 -O3 | 20161214 | 20161026 |
| 865 | armneon2 | gcc -mcpu=native -mfpu=neon-vfpv4 -Os | 20161214 | 20161026 |
| 866 | armneon2 | gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os | 20161214 | 20161026 |
| 889 | armneon | gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os | 20161214 | 20161026 |
| 890 | armneon | gcc -mcpu=native -mfpu=neon-vfpv4 -Os | 20161214 | 20161026 |
| 894 | armneon | gcc -mcpu=native -mfpu=neon-vfpv4 -O2 | 20161214 | 20161026 |
| 897 | armneon | gcc -mcpu=native -mfpu=neon-vfpv4 -O3 | 20161214 | 20161026 |
| 903 | armneon | gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3 | 20161214 | 20161026 |
| 905 | armneon | gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2 | 20161214 | 20161026 |
| 906 | armneon2 | gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2 | 20161214 | 20161026 |