Crate simd_aligned

Expand description

NOTE - Do not use this crate for now. It has been reactivated to make FFSVM compile again, but needs some architectural work.

In One Sentence

You want to use std::simd but realized there is no simple, safe and fast way to align your f32x8 (and friends) in memory and treat them as regular f32 slices for easy loading and manipulation; simd_aligned to the rescue.

Highlights

built on top of std::simd for easy data handling
supports everything from u8x2 to f64x8
think in flat slices (&[f32]), but get performance of properly aligned SIMD vectors (&[f32x16])
defines u8s, …, f36s as “best guess” for current platform (WIP)
provides N-dimensional VectorD and NxM-dimensional MatrixD.

Note: Right now this is an experimental crate. Features might be added or removed depending on how std::simd evolves. At the end of the day it’s just about being able to load and manipulate data without much fuzz.

Examples

Produces a vector that can hold 10 elements of type f64. Might internally allocate 5 elements of type f64x2, or 3 of type f64x4, depending on the platform. All elements are guaranteed to be properly aligned for fast access.

#![feature(portable_simd)]
use std::simd::*;
use simd_aligned::*;

// Create vectors of `10` f64 elements with value `0.0`.
let mut v1 = VectorD::<f64s>::with(0.0, 10);
let mut v2 = VectorD::<f64s>::with(0.0, 10);

// Get "flat", mutable view of the vector, and set individual elements:
let v1_m = v1.flat_mut();
let v2_m = v2.flat_mut();

// Set some elements on v1
v1_m[0] = 0.0;
v1_m[4] = 4.0;
v1_m[8] = 8.0;

// Set some others on v2
v2_m[1] = 0.0;
v2_m[5] = 5.0;
v2_m[9] = 9.0;

let mut sum = f64s::splat(0.0);

// Eventually, do something with the actual SIMD types. Does
// `std::simd` vector math, e.g., f64x8 + f64x8 in one operation:
sum = v1[0] + v2[0];

Benchmarks

There is no performance penalty for using simd_aligned, while retaining all the simplicity of handling flat arrays.

test vectors::packed       ... bench:          77 ns/iter (+/- 4)
test vectors::scalar       ... bench:       1,177 ns/iter (+/- 464)
test vectors::simd_aligned ... bench:          71 ns/iter (+/- 5)

FAQ

How does it relate to faster and `std::simd`?

simd_aligned builds on top of std::simd. At aims to provide common, SIMD-aligned data structure that support simple and safe scalar access patterns.
faster (as of today) is really good if you already have exiting flat slices in your code and want operate them “full SIMD ahead”. However, in particular when dealing with multiple slices at the same time (e.g., kernel computations) the performance impact of unaligned arrays can become a bit more noticeable (e.g., in the case of ffsvm up to 10% - 20%).

Re-exports

pub use crate::arch::current::*;

Modules

arch
Contains vector definitions with a fixed bit width.
traits
Unified views on SIMD types.

Macros

simd_swizzleExperimental
Constructs a new SIMD vector by copying elements from selected lanes in other vectors.

Structs

LaneCountExperimental
Specifies the number of lanes in a SIMD vector as a type.
MaskExperimental
A SIMD vector mask for LANES elements of width specified by Element.
MatrixD
A dynamic (heap allocated) matrix with one axis aligned for fast and safe SIMD access that also provides a flat view on its data.
MatrixFlat
Produced by MatrixD::flat, this allow for flat matrix access.
MatrixFlatMut
Provided by MatrixD::flat_mut, this allow for flat, mutable matrix access.
SimdExperimental
A SIMD vector of LANES elements of type T. Simd<T, N> has the same shape as [T; N], but operates like T.
VectorD
A dynamic (heap allocated) vector aligned for fast and safe SIMD access that also provides a flat view on its data.

Enums

WhichExperimental
Specifies a lane index into one of two SIMD vectors.

Traits

MaskElementExperimental
Marker trait for types that may be used as SIMD mask elements.
SimdElementExperimental
Marker trait for types that may be used as SIMD vector elements.
SimdExt
SimdFloatExperimental
Operations on SIMD vectors of floats.
SimdIntExperimental
Operations on SIMD vectors of signed integers.
SimdOrdExperimental
Parallel Ord.
SimdPartialEqExperimental
Parallel PartialEq.
SimdPartialOrdExperimental
Parallel PartialOrd.
SimdUintExperimental
Operations on SIMD vectors of unsigned integers.
StdFloatExperimental
This trait provides a possibly-temporary implementation of float functions that may, in the absence of hardware support, canonicalize to calling an operating system’s math.h dynamically-loaded library (also known as a shared object). As these conditionally require runtime support, they should only appear in binaries built assuming OS support: std.
SupportedLaneCountExperimental
Statically guarantees that a lane count is marked as supported.
SwizzleExperimental
Create a vector from the elements of another vector.
Swizzle2Experimental
Create a vector from the elements of two other vectors.
ToBitMaskExperimental
Converts masks to and from integer bitmasks.

Functions

packed_as_flat
Converts an slice of SIMD vectors into a flat slice of elements.
packed_as_flat_mut
Converts a mutable slice of SIMD vectors into a flat slice of elements.

Type Definitions

f32x2Experimental
A 64-bit SIMD vector with two elements of type f32.
f32x4Experimental
A 128-bit SIMD vector with four elements of type f32.
f32x8Experimental
A 256-bit SIMD vector with eight elements of type f32.
f32x16Experimental
A 512-bit SIMD vector with 16 elements of type f32.
f64x2Experimental
A 128-bit SIMD vector with two elements of type f64.
f64x4Experimental
A 256-bit SIMD vector with four elements of type f64.
f64x8Experimental
A 512-bit SIMD vector with eight elements of type f64.
i8x4Experimental
A 32-bit SIMD vector with four elements of type i8.
i8x8Experimental
A 64-bit SIMD vector with eight elements of type i8.
i8x16Experimental
A 128-bit SIMD vector with 16 elements of type i8.
i8x32Experimental
A 256-bit SIMD vector with 32 elements of type i8.
i8x64Experimental
A 512-bit SIMD vector with 64 elements of type i8.
i16x2Experimental
A 32-bit SIMD vector with two elements of type i16.
i16x4Experimental
A 64-bit SIMD vector with four elements of type i16.
i16x8Experimental
A 128-bit SIMD vector with eight elements of type i16.
i16x16Experimental
A 256-bit SIMD vector with 16 elements of type i16.
i16x32Experimental
A 512-bit SIMD vector with 32 elements of type i16.
i32x2Experimental
A 64-bit SIMD vector with two elements of type i32.
i32x4Experimental
A 128-bit SIMD vector with four elements of type i32.
i32x8Experimental
A 256-bit SIMD vector with eight elements of type i32.
i32x16Experimental
A 512-bit SIMD vector with 16 elements of type i32.
i64x2Experimental
A 128-bit SIMD vector with two elements of type i64.
i64x4Experimental
A 256-bit SIMD vector with four elements of type i64.
i64x8Experimental
A 512-bit SIMD vector with eight elements of type i64.
isizex2Experimental
A SIMD vector with two elements of type isize.
isizex4Experimental
A SIMD vector with four elements of type isize.
isizex8Experimental
A SIMD vector with eight elements of type isize.
mask8x8Experimental
A mask for SIMD vectors with eight elements of 8 bits.
mask8x16Experimental
A mask for SIMD vectors with 16 elements of 8 bits.
mask8x32Experimental
A mask for SIMD vectors with 32 elements of 8 bits.
mask8x64Experimental
A mask for SIMD vectors with 64 elements of 8 bits.
mask16x4Experimental
A mask for SIMD vectors with four elements of 16 bits.
mask16x8Experimental
A mask for SIMD vectors with eight elements of 16 bits.
mask16x16Experimental
A mask for SIMD vectors with 16 elements of 16 bits.
mask16x32Experimental
A mask for SIMD vectors with 32 elements of 16 bits.
mask32x2Experimental
A mask for SIMD vectors with two elements of 32 bits.
mask32x4Experimental
A mask for SIMD vectors with four elements of 32 bits.
mask32x8Experimental
A mask for SIMD vectors with eight elements of 32 bits.
mask32x16Experimental
A mask for SIMD vectors with 16 elements of 32 bits.
mask64x2Experimental
A mask for SIMD vectors with two elements of 64 bits.
mask64x4Experimental
A mask for SIMD vectors with four elements of 64 bits.
mask64x8Experimental
A mask for SIMD vectors with eight elements of 64 bits.
masksizex2Experimental
A mask for SIMD vectors with two elements of pointer width.
masksizex4Experimental
A mask for SIMD vectors with four elements of pointer width.
masksizex8Experimental
A mask for SIMD vectors with eight elements of pointer width.
u8x4Experimental
A 32-bit SIMD vector with four elements of type u8.
u8x8Experimental
A 64-bit SIMD vector with eight elements of type u8.
u8x16Experimental
A 128-bit SIMD vector with 16 elements of type u8.
u8x32Experimental
A 256-bit SIMD vector with 32 elements of type u8.
u8x64Experimental
A 512-bit SIMD vector with 64 elements of type u8.
u16x2Experimental
A 32-bit SIMD vector with two elements of type u16.
u16x4Experimental
A 64-bit SIMD vector with four elements of type u16.
u16x8Experimental
A 128-bit SIMD vector with eight elements of type u16.
u16x16Experimental
A 256-bit SIMD vector with 16 elements of type u16.
u16x32Experimental
A 512-bit SIMD vector with 32 elements of type u16.
u32x2Experimental
A 64-bit SIMD vector with two elements of type u32.
u32x4Experimental
A 128-bit SIMD vector with four elements of type u32.
u32x8Experimental
A 256-bit SIMD vector with eight elements of type u32.
u32x16Experimental
A 512-bit SIMD vector with 16 elements of type u32.
u64x2Experimental
A 128-bit SIMD vector with two elements of type u64.
u64x4Experimental
A 256-bit SIMD vector with four elements of type u64.
u64x8Experimental
A 512-bit SIMD vector with eight elements of type u64.
usizex2Experimental
A SIMD vector with two elements of type usize.
usizex4Experimental
A SIMD vector with four elements of type usize.
usizex8Experimental
A SIMD vector with eight elements of type usize.