Add AVX int32_t pdiv

Reference issue

What does this implement/fix?

Enables vectorized int32_t division by internally casting to double and truncating the result. Approximately 2x throughput compared to native integer division.

Additional information