std::vector.find simd version

std::find doesn't use simd intrinsics. ( check https://gms.tf/stdfind-and-memchr-optimizations.html )

So i thought simd can make std::find faster for some types.

Feature

• use avx2 intrinsic
• You don't need worry about alignment problem. ( this library handle it )

Requirement

• std::vector's value type must be arithmetic type or pointer type
• require avx2 option
• require c++ 11

Usage

std::vector<short> a;
a.reserve(1001);
for (unsigned short i = 0; i <= 1000; i++)
{
  a.push_back(i);
}

for (short i = 0; i < 1000; i++)
{
  fast_vector_find::find_simd(a, (short)i);
}

To develop

• implement partial template specialization of std::find ( Is this possible? )
• support for avx512 intrinsic ( ex) _mm256_cmpeq_epi8_mask )
• more optimization ( https://www.reddit.com/r/simd/comments/r3upl8/comment/hmd5uw9/?utm_source=share&utm_medium=web2x&context=3 )

Benchmark

tested on msvc, x64, windows