Commit Graph

252 Commits

Author SHA1 Message Date
Caleb Zulawski
138b9cf4bf Use intrinsic for min/max 2022-01-13 17:59:55 -05:00
Caleb Zulawski
65cb2c90a0 Fix mask alias 2022-01-09 13:12:22 -05:00
Jubilee
09fa72ae84 Merge portable-simd#219 - ./std-float
impl std::simd::StdFloat

 This introduces an extension trait to allow use of floating point methods
 that need runtime support. It is *excessively* documented because its mere
 existence is quite vexing, as the entire thing constitutes a leakage of
 implementation details into user observable space. Eventually the entire
 thing will ideally be folded into core and restructured to match the rest
 of the library, whatever that structure might look like at the time. This
 is preferred in lieu of the "lang item" path because any energy the lang
 items require (and it will be significant, by Simulacrum's estimation) is
 better spent on implementing our libmvec.
2021-12-31 15:50:20 -08:00
Jubilee Young
af26e3b9fd Tear down and rewrite support for float testing 2021-12-31 14:28:14 -08:00
Jubilee
4bbef261b6
Merge portable-simd#210 - ./wrap-shifts
Refactor ops.rs with wrapping shifts

This approaches reducing macro nesting in a slightly different way. Instead of just flattening details, make one macro apply another. This allows specifying all details up-front in the first macro invocation, making it easier to audit and refactor in the future.

This refactor also has some functional changes. Only one is a true behavior change, however:
- The visible one is that SIMD shifts are now wrapping, not panicking on overflow
- `core::simd` now has a lot more instances of `#[must_use]`, which merely lints
- div/rem now perform a SIMD check but remain as before, which should improve performance but be invisible
2021-12-30 01:22:01 -08:00
Jubilee Young
a42420583b Use Mask::any in div check 2021-12-23 23:17:13 -08:00
Jubilee Young
bc326a2bbc Refactor ops.rs with a recursive macro
This approaches reducing macro nesting in a slightly different way.
Instead of just flattening details, make one macro apply another.
This allows specifying all details up-front in the first macro
invocation, making it easier to audit and refactor in the future.
2021-12-22 15:37:05 -08:00
Jubilee Young
533f0fc81a Use relative intrinsics paths for bitmasks 2021-12-10 22:11:27 -08:00
Jubilee Young
5dcd397f47 Finish refactoring ints in ops.rs
This should perform a SIMD check for whether or not we can div/rem,
so that we can panic several times faster!
2021-12-09 13:23:45 -08:00
Jubilee Young
049e8ca7f7 Refactor float arith with #[must_use] 2021-12-08 18:09:46 -08:00
Jubilee Young
8aef340b8b Refactor bitops with #[must_use] 2021-12-08 18:09:32 -08:00
Jubilee Young
b6d0eec3de Wrap bitshifts in ops.rs
For all other operators, we use wrapping logic where applicable.
This is another case it applies. Per rust-lang/rust#91237, we may
wish to specify this as the natural behavior of `simd_{shl,shr}`.
2021-12-08 18:08:18 -08:00
Caleb Zulawski
d9f82f9c4d Remove Select trait 2021-12-04 05:54:15 +00:00
Jubilee
a8385522ad
Merge portable-simd#195 - portable-simd:trait-ops
Generic `core::ops` for `Simd<T, _>`

In order to maintain type soundness, we need to be sure we only implement an operation for `Simd<T, _> where T: SimdElement`... and also valid for that operation in general. While we could do this purely parametrically, it is more sound to implement the operators directly for the base scalar type arguments and then use type parameters to extend the operators to the "higher order" operations.

This implements that strategy and cleans up `simd::ops` into a few submodules:
- assign.rs: `core::ops::*Assign`
- deref.rs:  `core::ops` impls which "deref" borrowed versions of the arguments
- unary.rs: encloses the logic for unary operators on `Simd`, as unary ops are much simpler

This is possible since everything need not be nested in a single maze of macros anymore. The result simplifies the logic and allows reasoning about what operators are valid based on the expressed trait bounds, and also reduces the size of the trait implementation output in rustdoc, for a huge win of 4 MB off the size of `struct.Simd.html`! This addresses a common user complaint, as the original was over 5.5 MB and capable of crashing browsers!

This also carries a fix for a type-inference-related breakage, by removing the autosplatting (vector + scalar binop) impls, as unfortunately the presence of autosplatting was capable of busting type inference. We will likely need to see results from a Crater run before we can understand how to re-land autosplatting.
2021-12-02 17:41:30 -08:00
Jubilee Young
8003b04323 impl Op<&'_ RHS> for &'_ LHS 2021-12-01 15:45:01 -08:00
Jubilee Young
257fa7aa6d Drop splats for Simd<T, _>
Unfortunately, splatting impls currently break several crates.
Rust needs more time to review possible mitigations, so
drop the impls for the `impl Add<T> for Simd<T, _>` pattern, for now.
2021-12-01 15:40:53 -08:00
Jubilee Young
6094f22ceb impl unary.rs for Simd<{i,u}{8,16,32,64,size}, _>
In order to assure type soundness, these "base" impls
need to go directly on Simd<T, _> for every scalar type argument.
A bit of cleanup of ops.rs is still warranted.
2021-12-01 15:40:46 -08:00
Alexander Ronald Altman
b2dac7124b Uncomment AVX512 byte vector conversions
Resolves my comment in #197, at least for now; #187 is pending but since these are already here, just commented, it seemed to make sense to me to re-enable them anyway.
2021-11-26 13:48:52 -08:00
Jubilee Young
ae612100d2 Generically implement horizontal_{and,or,xor} 2021-11-24 08:07:41 -08:00
Jubilee Young
51ff925925 impl assign.rs<U> for Simd<T, _>
Instead of implementing {Op}Assign traits for individual scalar type args
to Simd<_, _>, use parametric impls that reassert the bounds of the binary op.
2021-11-23 18:11:48 -08:00
Jubilee Young
0a6992f5bf impl deref.rs<&Self> for Simd<T, _>
Instead of implementing each "deref" pattern for every single scalar,
we can use type parameters for Simd operating on &Self.
We can use a macro, but keep it cleaner and more explicit.
2021-11-23 18:11:48 -08:00
Jubilee Young
ced3a05526 Attempt to support to 64 lanes 2021-11-23 15:19:01 -08:00
Jubilee Young
f7b0358573 Sprinkle the crate with #[must_use] 2021-11-14 12:11:35 -08:00
Caleb Zulawski
36e198b97a Use new bitmask intrinsics with byte arrays 2021-11-13 13:22:06 -08:00
Jubilee Young
1ce1c645cf Rewrite Arm transmutes, reading std::arch closer 2021-11-12 16:56:14 -08:00
Jubilee Young
6ddf7ad8e1 Restrict Arm types to Arm v7+
This mostly mirrors the restrictions in std::arch.
It can be loosened slightly with later refactoring.
2021-11-11 11:40:18 -08:00
Jubilee Young
949f71c0dc Deny warnings in CI and fix 2021-11-10 14:40:32 -08:00
Jubilee
081240a663
Merge pull request #175 from rust-lang/feature/more-actions
Delete travis config, move tests to github actions.
2021-11-09 20:12:47 -08:00
Caleb Zulawski
349a61143c Delete travis config, move tests to github actions. 2021-11-10 02:18:41 +00:00
Caleb Zulawski
0ecf9871fc
Merge pull request #181 from rust-lang/rotate_lanes
rotate_{left,right} -> rotate_lanes_{left,right}
2021-11-08 23:01:05 -05:00
Proloy Mishra
d2e87281fc
add Simd::from_slice (#177)
* add `Simd::from_slice`

uses a zeroed initial array and loops so that it can be const.
unfortunately, parameterizing the assert with slice length
needs `#![feature(const_fn_fn_ptr_basics)]` to work.
2021-11-08 17:28:43 -08:00
Caleb Zulawski
4e00aa68c7 rotate_{left,right} -> rotate_lanes_{left,right} 2021-11-06 00:34:23 +00:00
Caleb Zulawski
772bf2090e Hide select impl in sealed trait 2021-10-22 00:10:44 -07:00
Caleb Zulawski
7c2d295a76 Hide mask impl details in sealed trait. 2021-10-22 00:10:44 -07:00
Jubilee Young
ab8eec7cba Fixup import pathing for core
This changes simd_swizzle! to a decl_macro to give it a path,
so it can be imported using a path and not the crate root.
It also adds various uses that were missed and adjusts paths.
2021-10-21 18:20:06 -07:00
Caleb Zulawski
5b4282edcd Improve docs 2021-10-11 13:18:59 -07:00
Caleb Zulawski
765bee6362 Update crates/core_simd/src/swizzle.rs
Co-authored-by: Jubilee <46493976+workingjubilee@users.noreply.github.com>
2021-10-11 13:18:59 -07:00
Caleb Zulawski
cd7ecba19f Remove adt_const_params feature 2021-10-11 13:18:59 -07:00
Caleb Zulawski
37797d9c0a simd_shuffle -> simd_swizzle 2021-10-11 13:18:59 -07:00
Caleb Zulawski
98e4fcae5a Fix macro in core
Co-authored-by: Jubilee <46493976+workingjubilee@users.noreply.github.com>
2021-10-11 13:18:59 -07:00
Caleb Zulawski
10168fb7c4 Add new swizzle API
Expand swizzle API and migrate existing functions. Add rotate_left, rotate_right.

Hide implementation details

Add simd_shuffle macro
2021-10-11 13:18:59 -07:00
Jubilee
a16b481a08 Simplify language for scatter/gather
Co-authored-by: Caleb Zulawski <caleb.zulawski@gmail.com>
2021-10-03 14:35:07 -07:00
Jubilee Young
9be26656d2 Rewrite gather/scatter docs
Headings with # Safety and # Examples are more "std style".
Use terms like "enable" and "disable", rather than "mask" jargon.
2021-10-03 14:33:56 -07:00
Jubilee
01e9816ace docs: fix typo gather -> scatter
Co-authored-by: Jacob Lifshay <programmerjake@gmail.com>
2021-10-03 11:04:58 -07:00
Jubilee Young
6d23662689 Add {gather,scatter}_select_unchecked
This unsafe variant allows the thinnest API, in case LLVM cannot
perform loop-invariant code motion on a hot loop when the safe
form is used.

An unchecked variant could be added to other forms, but doesn't
seem likely to improve anything, since it would just add heavier
codegen.
2021-10-03 11:04:58 -07:00
Jubilee
436ca7f7aa
Add lanes() and associated LANES const 2021-09-29 14:19:45 -07:00
Jubilee Young
b506e3e28e Renovate for Edition 2021
In a still-future edition, `unsafe_op_in_unsafe_fn` may error.
Let's get ahead of that.
2021-09-29 14:11:40 -07:00
Magnus Ulimoen
ec05dfbbf9 Add associated LANES const 2021-09-29 17:20:07 +00:00
Magnus Ulimoen
4fbccafc66 Add lanes() 2021-09-29 16:38:03 +00:00
Jubilee Young
afd7c5a5ee Make sure MaskElement is in bitmasks.rs 2021-09-27 15:27:34 -07:00