Module guff_matrix::x86

x86_64-specific SIMD

This module contains architecture-specific code to implement the “Wrap-Around Read Matrix” multiply algorithm for the x86_64 architecture.

At present a fixed polynomial 0x11b is used to implement calculations in GF(2⁸). This may change later to include a choice of different polynomials and field sizes.

Structs

X86SimpleMatrix

Matrix storage type using direct memory accesses to implement wrap-around.

X86u8x16Long0x11b

Newtype for working with __m128i vector type.

Functions

vector_cube_p8x16^⚠

A function to help examine whether/how inlining is done. Returns the cube of each element

vmul_p8_buffer^⚠

16-way SIMD multiplication of buffers in GF(2⁸) with poly 0x11b

vmul_p8x16^⚠

16-way SIMD multiplication of elements in GF(2⁸) with poly 0x11b