LaihoE/SIMD-itertools

Faster implementations of standard library operations like find, filter, position etc.

Rust

186

70 commits

updated Feb 8, 2025

See the code

README

SIMD-accelerated iterators

crates.io

Unmatched flexibility 🤯

let needles = [42, 52, 94];
arr.iter().any_simd(|x| needles.contains(x) || x > 156);
  • Works by letting LLVM do the vectorization (may change in the future).
  • Functions are made easy to paste into Godbolt for inspection.

Currently the following are implemented:

find filter position all any argmin/argmax

Tradeoffs

Every piece of software makes tradeoffs. The goal of this library it to provide the majority of performance gains gained from going scalar -> vectorized, while staying user-friendly. If you are looking to shave off the last few cycles this might not be what you are looking for.

⚠️ Warning ⚠️:

The library makes one extra assumption over the stdlib: The closure may be executed any number of times:

arr.iter().simd_position(|x| {
    println!("hello world");
    *x == 42
})

May print a different number of times compared to the standard library. This shouldn't be an issue under normal use-cases but something to keep in mind.

Why is this not part of the standard library

It's tricky. Hopefully one day.

algorithms
iterators
itertools
rust
simd

Contributors

LaihoE

57 commits

smu160

12 commits

PaulDotSH

1 commits

LaihoE/SIMD-itertools

Faster implementations of standard library operations like find, filter, position etc.

Rust

186

70 commits

updated Feb 8, 2025

See the code

README

SIMD-accelerated iterators

crates.io

Unmatched flexibility 🤯

let needles = [42, 52, 94];
arr.iter().any_simd(|x| needles.contains(x) || x > 156);
  • Works by letting LLVM do the vectorization (may change in the future).
  • Functions are made easy to paste into Godbolt for inspection.

Currently the following are implemented:

find filter position all any argmin/argmax

Tradeoffs

Every piece of software makes tradeoffs. The goal of this library it to provide the majority of performance gains gained from going scalar -> vectorized, while staying user-friendly. If you are looking to shave off the last few cycles this might not be what you are looking for.

⚠️ Warning ⚠️:

The library makes one extra assumption over the stdlib: The closure may be executed any number of times:

arr.iter().simd_position(|x| {
    println!("hello world");
    *x == 42
})

May print a different number of times compared to the standard library. This shouldn't be an issue under normal use-cases but something to keep in mind.

Why is this not part of the standard library

It's tricky. Hopefully one day.

algorithms
iterators
itertools
rust
simd

Contributors

LaihoE

57 commits

smu160

12 commits

PaulDotSH

1 commits

Languages

Rust

100.0%