| Compiler | Implementations |
| gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os | optimized_nonSSE |
| gcc -mcpu=native -mfpu=neon-vfpv4 -Os | optimized_nonSSE |
| Compiler | Implementations |
| gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2 | optimized_nonSSE |
| gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3 | optimized_nonSSE ref |
| gcc -mcpu=native -mfpu=neon-vfpv4 -O2 | optimized_nonSSE ref |
| gcc -mcpu=native -mfpu=neon-vfpv4 -O3 | ref |