Enum GfBackend 

#[repr(u8)]
pub enum GfBackend {
    Scalar = 0,
    Ssse3 = 1,
    Avx2 = 2,
    Avx512Pshufb = 3,
    Gfni256 = 4,
    Gfni512 = 5,
    AffineScalar = 6,
    Neon = 7,
}

GF(256) multiply-add backend, exposed so an A/B bench can compare every SIMD rung against the scalar baseline and so the GFNI / AVX-512 logic can be validated on a host without the silicon via the bit-exact software affine emulation. The production hot loop (gf_mul_add) auto-selects the fastest rung this host can run; the ladder always bottoms out at Scalar.

Variants

Scalar = 0

Portable table-lookup multiply. Always available; the fallback floor.

Ssse3 = 1

SSSE3 PSHUFB nibble-table multiply, 16 bytes per op.

Avx2 = 2

AVX2 PSHUFB nibble-table multiply, 32 bytes per op.

Avx512Pshufb = 3

AVX-512BW PSHUFB nibble-table multiply, 64 bytes per op (no GFNI).

Gfni256 = 4

GFNI affine GF(2^8) multiply on 256-bit lanes: a hardware field multiply (no table), broad consumer reach (gfni + avx2, no AVX-512 needed).

Gfni512 = 5

GFNI affine GF(2^8) multiply on 512-bit lanes: 64 bytes per op in one hardware instruction.

AffineScalar = 6

Software emulation of the GFNI affine transform: bit-exact to the GFNI hardware path, runnable on any host. Validates the GFNI logic where the silicon is absent, the same emulate-to-verify approach the AVX-512 substrate uses.

Neon = 7

ARM NEON TBL nibble-table multiply, 16 bytes per vqtbl1q_u8. The aarch64 byte-shuffle rung: the same nibble-table technique as SSSE3, so bit-identical to it (and to scalar). NEON is baseline on every aarch64 CPU, so this rung is always available on Apple Silicon / Neoverse.

Implementations

impl GfBackend

pub fn available(self) -> bool

Whether this backend can execute on the current host.

pub fn name(self) -> &'static str

Short identifier for diagnostics / bench output.

Trait Implementations

impl Clone for GfBackend

fn clone(&self) -> GfBackend

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for GfBackend

impl Debug for GfBackend

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Eq for GfBackend

impl PartialEq for GfBackend

fn eq(&self, other: &GfBackend) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for GfBackend