packed_simd

packed_simd copied to clipboard

It looks like the speedup multipliers for `par_simd`, `ispc` and `ispc+tasks` are incorrect given the execution time. `par_simd` should be 7.37x `ispc` should be 2.55x `ispc+tasks` should be 5.09x. And...

pbrubaker

platform_intrinsics feature has been removed

4

......\packed_simd-0.3.9\src\lib.rs:218:5 | 218 | platform_intrinsics, | ^^^^^^^^^^^^^^^^^^^ feature has been removed | = note: SIMD intrinsics use the regular intrinsics ABI now --> ```rust #![feature(portable_simd)] use std::simd::f32x4; fn main() {...

lixiaoyong12

The crate is broken due to removal of stdsimd feature.

1

The removal of the stdsimd feature in nightly breaks the crate. Relevant commit: https://github.com/rust-lang/rust/commit/ea37e8091fe87ae0a7e204c034e7d55061e56790

R-Goc

The URL for this repository is still listed as https://rust-lang.github.io/packed_simd/packed_simd_2/ Presumably this should be changed to something current.

atouchet

Unrecognized platform-specific intrinsic function: `simd_shuffle64`

5

The packed_simd_2 crate fails to compile on the latest nightly with errors about `simd_shuffle64` ``` error: unrecognized platform-specific intrinsic function: `simd_shuffle64` --> /home/raphael/.cargo/registry/src/index.crates.io-6f17d22bba15001f/packed_simd_2-0.3.8/src/codegen/llvm.rs:15:5 ```

tustvold

Fix Android CI

1

This is modeled on https://github.com/rust-lang/backtrace-rs/commit/caf2f0b7593ade9e521fdc73f0b26bbb3b437493

workingjubilee

Difference between packed_simd and portable-simd, and which one should I use?

1

Hi thanks for the Rust language and the two libraries! I wonder which one should I use now? They seems to have similar features and one even links to the...

fzyzcjy

`Simd<A>` has the same layout and ABI as `A`

4

This will allow code to depend on `Simd`'s layout. My specific use case is to emit impls of `FromBytes` and `AsBytes` for [zerocopy](https://crates.io/crates/zerocopy) so that other crates can safely transmute...

joshlf

Hi thanks for the lib! My usage needs to index into an array based on a value. Shortly speaking, it is doing something like: ``` for x in 0..width {...

fzyzcjy

Move to new CI (GitHub Actions?)

1

QEMU sometimes throws apparently spurious errors, and in any case adds a slowdown. It would be nice to take advantage of the fact that Travis offers the following arch tags:...

workingjubilee