shorten vsldoi native code by one instruction, aka. avoid superfluous

shuffle of stale vector value
This commit is contained in:
gbeauche 2005-03-14 00:30:44 +00:00
parent 4918fe8d2d
commit cbf0c3debf

View File

@ -1635,9 +1635,8 @@ void op_sse2_vsldoi_##SH(void) \
"pshufd %3,%%xmm1,%%xmm1\n" \
"pslldq %4,%%xmm0\n" \
"psrldq %5,%%xmm1\n" \
"pshufd %3,%%xmm0,%%xmm0\n" \
"pshufd %3,%%xmm1,%%xmm1\n" \
"por %%xmm1,%%xmm0\n" \
"pshufd %3,%%xmm0,%%xmm0\n" \
"movaps %%xmm0,(%0)\n" \
: : \
"r" (reg_VD), "r" (reg_V0), "r" (reg_V1), \