Section baseline: What are N, M, K? 3 matrices or? Laid out as a flat array, or what? `c[m * N + n] += a[m * K + k] * b[k * N + n];`, ah, apparently a b and c are the matrices? How does this work?
Section body: What is the mathy "C′=αC+A⋅B"? derivative of a constant is the angle times the constant plus the dot product of A and B???
They are laid out, rather than nested arrays, as a single continuous collection of bytes that can be interpreted as having a matrix shape. That's where the `m * N + n` comes from (m rows down and n cols in)
C' = alpha C + A.B
This is the 'generalised matrix-matrix multiplication' (GEMM) operation. It's multiplying the matrices A and B, adding it to a scales version of C and inserting it back into C. Setting alpha to 0 gets you basic matmulJavaScript code is readable, at least.
Not trying to sound dismissive here but the core math the post is working with is actually a pretty straightforward matrix multiplication.
The bulk of the discussion focuses on optimizing the execution of that straightforward multiplication algorithm [triple-nested for loop; O(n^3)] rather than making algorithmic/mathematic optimizations.
And again, specific questions are easier to answer. :)