[llvm] [X86] Vectorize non-power-of-two integer division (PR #215076)

[via llvm-commits]

Andarwinux wrote:

see https://github.com/llvm/llvm-project/pull/205263#discussion_r3749540232
please add this to tests
```llvm
define void @div_extract(ptr %0, ptr %1, ptr %2) nounwind {
  %4 = load <2 x i8>, ptr %1
  %5 = load <2 x i8>, ptr %2
  %6 = udiv <2 x i8> %4, %5
  %7 = extractelement <2 x i8> %6, i64 0
  store i8 %7, ptr %0, align 1
  %8 = getelementptr inbounds nuw i8, ptr %0, i64 1
  %9 = extractelement <2 x i8> %6, i64 1
  store i8 %9, ptr %8, align 1
  ret void
}
```

https://github.com/llvm/llvm-project/pull/215076