Extending Level 2 and Level 3 BLAS routines for linear algebra

This project aims to resolve Issue add blas bindings and implementations for linear algebra by systematically adding the missing Level 1, 2, and 3 BLAS math routines to the @stdlib/blas/base/* namespace. To ensure speed and backward compatibility, every new routine will be split into two parts: a pure, low-level math engine and a simple, user-friendly classic API.This architectural split empowers the underlying math engine to safely handle complex, non-contiguous memory layouts and apply strict cache optimizations using explicit strides. Meanwhile, the classic API acts as a protective shield, automatically calculating the required memory offsets (e.g., negative strides) to deliver a seamless and intuitive developer experience.