| Compiler | Implementations |
| gcc -funroll-loops -mcpu=marvell-pj4 -O2 | optimized_nonSSE ref |
| gcc -funroll-loops -mcpu=marvell-pj4 -O3 | optimized_nonSSE ref |
| gcc -funroll-loops -mcpu=marvell-pj4 -Os | optimized_nonSSE ref |
| gcc -mcpu=marvell-pj4 -O2 | optimized_nonSSE ref |
| gcc -mcpu=marvell-pj4 -O3 | optimized_nonSSE ref |
| gcc -mcpu=marvell-pj4 -Os | optimized_nonSSE ref |