aling 16
lab_1:
prefetchnta [esi+320]
movaps xmm0, [esi+ 0]
movaps xmm1, [esi+16]
movntps [edi+ 0], xmm0
movntps [edi+ 16], xmm1
movaps xmm2, [esi+32]
movaps xmm3, [esi+48]
movntps [edi+ 32], xmm2
movntps [edi+ 48], xmm3
add esi, 64
add edi, 64
dec eax
jne lab_1

こっちの方がよくないか?というかプリフェッチのページおしえてくれ