avx: add _mm256_div_pd, _mm256_div_ps

[vbarrielle]

It looks like there is no llvm compiler intrinsic for these isntructions, so I mimicked the implementation in https://github.com/llvm-mirror/clang/blob/master/lib/Headers/avxintrin.h#L166-L201

I have inspected the resulting binary with objdump -d, and I get

0000000000028eb0 <_ZN7stdsimd3x863avx13_mm256_div_pd17hfb622e668096388cE>:
   28eb0:       55                      push   %rbp
   28eb1:       48 89 e5                mov    %rsp,%rbp
   28eb4:       c5 fd 5e c1             vdivpd %ymm1,%ymm0,%ymm0
   28eb8:       5d                      pop    %rbp
   28eb9:       c3                      retq   
   28eba:       66 0f 1f 44 00 00       nopw   0x0(%rax,%rax,1)

which seems ok.

[BurntSushi]

Poifect!