GitHub - AdamNiederer/faster: SIMD for humans
source link: https://github.com/AdamNiederer/faster
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
faster
SIMD for Humans
Easy, powerful, portable, absurdly fast numerical calculations. Includes static dispatch with inlining based on your platform and vector types, zero-allocation iteration, vectorized loading/storing, and support for uneven collections.
It looks something like this:
use faster::*;
let lots_of_3s = (&[-123.456f32; 128][..]).simd_iter()
.simd_map(f32s(0.0), |v| {
f32s(9.0) * v.abs().sqrt().rsqrt().ceil().sqrt() - f32s(4.0) - f32s(2.0)
})
.scalar_collect();
Which is analogous to this scalar code:
let lots_of_3s = (&[-123.456f32; 128][..]).iter()
.map(|v| {
9.0 * v.abs().sqrt().sqrt().recip().ceil().sqrt() - 4.0 - 2.0
})
.collect::<Vec<f32>>();
The vector size is entirely determined by the machine you’re compiling for - it attempts to use the largest vector size supported by your machine, and works on any platform or architecture (see below for details).
Compare this to traditional explicit SIMD:
use std::mem::transmute;
use stdsimd::{f32x4, f32x8};
let lots_of_3s = &mut [-123.456f32; 128][..];
if cfg!(all(not(target_feature = "avx"), target_feature = "sse")) {
for ch in init.chunks_mut(4) {
let v = f32x4::load(ch, 0);
let scalar_abs_mask = unsafe { transmute::<u32, f32>(0x7fffffff) };
let abs_mask = f32x4::splat(scalar_abs_mask);
// There isn't actually an absolute value intrinsic for floats - you
// have to look at the IEEE 754 spec and do some bit flipping
v = unsafe { _mm_and_ps(v, abs_mask) };
v = unsafe { _mm_sqrt_ps(v) };
v = unsafe { _mm_rsqrt_ps(v) };
v = unsafe { _mm_ceil_ps(v) };
v = unsafe { _mm_sqrt_ps(v) };
v = unsafe { _mm_mul_ps(v, 9.0) };
v = unsafe { _mm_sub_ps(v, 4.0) };
v = unsafe { _mm_sub_ps(v, 2.0) };
f32x4::store(ch, 0);
}
} else if cfg!(all(not(target_feature = "avx512"), target_feature = "avx")) {
for ch in init.chunks_mut(8) {
let v = f32x8::load(ch, 0);
let scalar_abs_mask = unsafe { transmute::<u32, f32>(0x7fffffff) };
let abs_mask = f32x8::splat(scalar_abs_mask);
v = unsafe { _mm256_and_ps(v, abs_mask) };
v = unsafe { _mm256_sqrt_ps(v) };
v = unsafe { _mm256_rsqrt_ps(v) };
v = unsafe { _mm256_ceil_ps(v) };
v = unsafe { _mm256_sqrt_ps(v) };
v = unsafe { _mm256_mul_ps(v, 9.0) };
v = unsafe { _mm256_sub_ps(v, 4.0) };
v = unsafe { _mm256_sub_ps(v, 2.0) };
f32x8::store(ch, 0);
}
}
Even with all of that boilerplate, this still only supports x86-64 machines with SSE or AVX - and you have to look up each intrinsic to ensure it’s usable for your compilation target.
Upcoming Features
A rewrite of the iterator API is upcoming, as well as internal changes to better match the direction Rust is taking with explicit SIMD.
Compatibility
Faster currently supports any architecture with floating point support, although hardware acceleration is only enabled on machines with x86’s vector extensions.
Performance
Here are some extremely unscientific benchmarks which, at least, prove that this isn’t any worse than scalar iterators. Even on ancient CPUs, a lot of performance can be extracted out of SIMD.
$ RUSTFLAGS="-C target-cpu=ivybridge" cargo bench # host is ivybridge; target has AVX
test tests::base100_enc_scalar ... bench: 1,307 ns/iter (+/- 45)
test tests::base100_enc_simd ... bench: 332 ns/iter (+/- 10)
test tests::determinant2_scalar ... bench: 486 ns/iter (+/- 8)
test tests::determinant2_simd ... bench: 215 ns/iter (+/- 3)
test tests::determinant3_scalar ... bench: 389 ns/iter (+/- 6)
test tests::determinant3_simd ... bench: 209 ns/iter (+/- 3)
test tests::map_fill_simd ... bench: 835 ns/iter (+/- 12)
test tests::map_scalar ... bench: 6,963 ns/iter (+/- 117)
test tests::map_simd ... bench: 879 ns/iter (+/- 18)
test tests::map_uneven_simd ... bench: 884 ns/iter (+/- 10)
test tests::nop_scalar ... bench: 49 ns/iter (+/- 0)
test tests::nop_simd ... bench: 34 ns/iter (+/- 0)
test tests::reduce_scalar ... bench: 6,905 ns/iter (+/- 107)
test tests::reduce_simd ... bench: 839 ns/iter (+/- 13)
test tests::reduce_uneven_simd ... bench: 838 ns/iter (+/- 11)
test tests::zip_nop_scalar ... bench: 824 ns/iter (+/- 18)
test tests::zip_nop_simd ... bench: 231 ns/iter (+/- 5)
test tests::zip_scalar ... bench: 901 ns/iter (+/- 29)
test tests::zip_simd ... bench: 1,128 ns/iter (+/- 12)
RUSTFLAGS="-C target-cpu=x86-64" cargo bench # host is ivybridge; target has SSE2
test tests::base100_enc_scalar ... bench: 760 ns/iter (+/- 11)
test tests::base100_enc_simd ... bench: 492 ns/iter (+/- 2)
test tests::determinant2_scalar ... bench: 477 ns/iter (+/- 3)
test tests::determinant2_simd ... bench: 277 ns/iter (+/- 1)
test tests::determinant3_scalar ... bench: 380 ns/iter (+/- 3)
test tests::determinant3_simd ... bench: 285 ns/iter (+/- 2)
test tests::map_fill_simd ... bench: 1,797 ns/iter (+/- 8)
test tests::map_scalar ... bench: 7,237 ns/iter (+/- 51)
test tests::map_simd ... bench: 1,879 ns/iter (+/- 12)
test tests::map_uneven_simd ... bench: 1,878 ns/iter (+/- 9)
test tests::nop_scalar ... bench: 47 ns/iter (+/- 0)
test tests::nop_simd ... bench: 34 ns/iter (+/- 0)
test tests::reduce_scalar ... bench: 7,021 ns/iter (+/- 39)
test tests::reduce_simd ... bench: 1,801 ns/iter (+/- 8)
test tests::reduce_uneven_simd ... bench: 1,734 ns/iter (+/- 9)
test tests::zip_nop_scalar ... bench: 803 ns/iter (+/- 9)
test tests::zip_nop_simd ... bench: 257 ns/iter (+/- 1)
test tests::zip_scalar ... bench: 988 ns/iter (+/- 6)
test tests::zip_simd ... bench: 629 ns/iter (+/- 5)
$ RUSTFLAGS="-C target-cpu=pentium" cargo bench # host is ivybridge; this only runs the polyfills!
test tests::bench_determinant2_scalar ... bench: 427 ns/iter (+/- 2)
test tests::bench_determinant2_simd ... bench: 402 ns/iter (+/- 1)
test tests::bench_determinant3_scalar ... bench: 354 ns/iter (+/- 1)
test tests::bench_determinant3_simd ... bench: 593 ns/iter (+/- 1)
test tests::bench_map_scalar ... bench: 7,195 ns/iter (+/- 28)
test tests::bench_map_simd ... bench: 6,271 ns/iter (+/- 22)
test tests::bench_map_uneven_simd ... bench: 6,288 ns/iter (+/- 22)
test tests::bench_nop_scalar ... bench: 38 ns/iter (+/- 0)
test tests::bench_nop_simd ... bench: 69 ns/iter (+/- 0)
test tests::bench_reduce_scalar ... bench: 7,004 ns/iter (+/- 17)
test tests::bench_reduce_simd ... bench: 6,063 ns/iter (+/- 17)
test tests::bench_reduce_uneven_simd ... bench: 6,107 ns/iter (+/- 11)
test tests::bench_zip_nop_scalar ... bench: 623 ns/iter (+/- 2)
test tests::bench_zip_nop_simd ... bench: 289 ns/iter (+/- 1)
test tests::bench_zip_scalar ... bench: 972 ns/iter (+/- 3)
test tests::bench_zip_simd ... bench: 621 ns/iter (+/- 3)
Recommend
About Joyk
Aggregate valuable and interesting links.
Joyk means Joy of geeK