Crate fastlanes 

Modules

scalar

Portable scalar transpose using a 64-bit gather and the classic 8x8 bit-matrix transpose. Used as the fallback when no SIMD implementation is available.

x86

x86-64 transpose implementations: BMI2 (PEXT/PDEP) and AVX-512 VBMI.

Macros

iterate

pack

seq_t

unpack

Constants

FL_ORDER

Traits

BitPacking

BitPack into a compile-time known bit-width.

BitPackingCompare

Delta

FastLanes

FastLanesComparable

FoR

RLE

Transpose

Functions

transpose

Return the corresponding index in a transposed FastLanes vector.

transpose_bits

Transpose 1024 bits into FastLanes layout, dispatching to the best implementation.

untranspose_bits

Untranspose a T-width comparison mask (1024 bits) from FastLanes layout into logical row order, dispatching to the best implementation. For T = u64 this is the canonical FastLanes bit untranspose; narrower T undo the per-lane packing produced by unpack_cmp for that width.