1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
use crate::simd::{
cmp::SimdPartialOrd,
num::SimdUint,
ptr::{SimdConstPtr, SimdMutPtr},
LaneCount, Mask, MaskElement, SupportedLaneCount, Swizzle,
};
/// A SIMD vector with the shape of `[T; N]` but the operations of `T`.
///
/// `Simd<T, N>` supports the operators (+, *, etc.) that `T` does in "elementwise" fashion.
/// These take the element at each index from the left-hand side and right-hand side,
/// perform the operation, then return the result in the same index in a vector of equal size.
/// However, `Simd` differs from normal iteration and normal arrays:
/// - `Simd<T, N>` executes `N` operations in a single step with no `break`s
/// - `Simd<T, N>` can have an alignment greater than `T`, for better mechanical sympathy
///
/// By always imposing these constraints on `Simd`, it is easier to compile elementwise operations
/// into machine instructions that can themselves be executed in parallel.
///
/// ```rust
/// # #![feature(portable_simd)]
/// # use core::simd::{Simd};
/// # use core::array;
/// let a: [i32; 4] = [-2, 0, 2, 4];
/// let b = [10, 9, 8, 7];
/// let sum = array::from_fn(|i| a[i] + b[i]);
/// let prod = array::from_fn(|i| a[i] * b[i]);
///
/// // `Simd<T, N>` implements `From<[T; N]>`
/// let (v, w) = (Simd::from(a), Simd::from(b));
/// // Which means arrays implement `Into<Simd<T, N>>`.
/// assert_eq!(v + w, sum.into());
/// assert_eq!(v * w, prod.into());
/// ```
///
///
/// `Simd` with integer elements treats operators as wrapping, as if `T` was [`Wrapping<T>`].
/// Thus, `Simd` does not implement `wrapping_add`, because that is the default behavior.
/// This means there is no warning on overflows, even in "debug" builds.
/// For most applications where `Simd` is appropriate, it is "not a bug" to wrap,
/// and even "debug builds" are unlikely to tolerate the loss of performance.
/// You may want to consider using explicitly checked arithmetic if such is required.
/// Division by zero on integers still causes a panic, so
/// you may want to consider using `f32` or `f64` if that is unacceptable.
///
/// [`Wrapping<T>`]: core::num::Wrapping
///
/// # Layout
/// `Simd<T, N>` has a layout similar to `[T; N]` (identical "shapes"), with a greater alignment.
/// `[T; N]` is aligned to `T`, but `Simd<T, N>` will have an alignment based on both `T` and `N`.
/// Thus it is sound to [`transmute`] `Simd<T, N>` to `[T; N]` and should optimize to "zero cost",
/// but the reverse transmutation may require a copy the compiler cannot simply elide.
///
/// # ABI "Features"
/// Due to Rust's safety guarantees, `Simd<T, N>` is currently passed and returned via memory,
/// not SIMD registers, except as an optimization. Using `#[inline]` on functions that accept
/// `Simd<T, N>` or return it is recommended, at the cost of code generation time, as
/// inlining SIMD-using functions can omit a large function prolog or epilog and thus
/// improve both speed and code size. The need for this may be corrected in the future.
///
/// Using `#[inline(always)]` still requires additional care.
///
/// # Safe SIMD with Unsafe Rust
///
/// Operations with `Simd` are typically safe, but there are many reasons to want to combine SIMD with `unsafe` code.
/// Care must be taken to respect differences between `Simd` and other types it may be transformed into or derived from.
/// In particular, the layout of `Simd<T, N>` may be similar to `[T; N]`, and may allow some transmutations,
/// but references to `[T; N]` are not interchangeable with those to `Simd<T, N>`.
/// Thus, when using `unsafe` Rust to read and write `Simd<T, N>` through [raw pointers], it is a good idea to first try with
/// [`read_unaligned`] and [`write_unaligned`]. This is because:
/// - [`read`] and [`write`] require full alignment (in this case, `Simd<T, N>`'s alignment)
/// - `Simd<T, N>` is often read from or written to [`[T]`](slice) and other types aligned to `T`
/// - combining these actions violates the `unsafe` contract and explodes the program into
/// a puff of **undefined behavior**
/// - the compiler can implicitly adjust layouts to make unaligned reads or writes fully aligned
/// if it sees the optimization
/// - most contemporary processors with "aligned" and "unaligned" read and write instructions
/// exhibit no performance difference if the "unaligned" variant is aligned at runtime
///
/// Less obligations mean unaligned reads and writes are less likely to make the program unsound,
/// and may be just as fast as stricter alternatives.
/// When trying to guarantee alignment, [`[T]::as_simd`][as_simd] is an option for
/// converting `[T]` to `[Simd<T, N>]`, and allows soundly operating on an aligned SIMD body,
/// but it may cost more time when handling the scalar head and tail.
/// If these are not enough, it is most ideal to design data structures to be already aligned
/// to `mem::align_of::<Simd<T, N>>()` before using `unsafe` Rust to read or write.
/// Other ways to compensate for these facts, like materializing `Simd` to or from an array first,
/// are handled by safe methods like [`Simd::from_array`] and [`Simd::from_slice`].
///
/// [`transmute`]: core::mem::transmute
/// [raw pointers]: pointer
/// [`read_unaligned`]: pointer::read_unaligned
/// [`write_unaligned`]: pointer::write_unaligned
/// [`read`]: pointer::read
/// [`write`]: pointer::write
/// [as_simd]: slice::as_simd
//
// NOTE: Accessing the inner array directly in any way (e.g. by using the `.0` field syntax) or
// directly constructing an instance of the type (i.e. `let vector = Simd(array)`) should be
// avoided, as it will likely become illegal on `#[repr(simd)]` structs in the future. It also
// causes rustc to emit illegal LLVM IR in some cases.
#[repr(simd, packed)]
pub struct Simd<T, const N: usize>([T; N])
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement;
impl<T, const N: usize> Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
/// Number of elements in this vector.
pub const LEN: usize = N;
/// Returns the number of elements in this SIMD vector.
///
/// # Examples
///
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::u32x4;
/// let v = u32x4::splat(0);
/// assert_eq!(v.len(), 4);
/// ```
#[inline]
#[allow(clippy::len_without_is_empty)]
pub const fn len(&self) -> usize {
Self::LEN
}
/// Constructs a new SIMD vector with all elements set to the given value.
///
/// # Examples
///
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::u32x4;
/// let v = u32x4::splat(8);
/// assert_eq!(v.as_array(), &[8, 8, 8, 8]);
/// ```
#[inline]
#[rustc_const_unstable(feature = "portable_simd", issue = "86656")]
pub const fn splat(value: T) -> Self {
const fn splat_const<T, const N: usize>(value: T) -> Simd<T, N>
where
T: SimdElement,
LaneCount<N>: SupportedLaneCount,
{
Simd::from_array([value; N])
}
fn splat_rt<T, const N: usize>(value: T) -> Simd<T, N>
where
T: SimdElement,
LaneCount<N>: SupportedLaneCount,
{
// This is preferred over `[value; N]`, since it's explicitly a splat:
// https://github.com/rust-lang/rust/issues/97804
struct Splat;
impl<const N: usize> Swizzle<N> for Splat {
const INDEX: [usize; N] = [0; N];
}
Splat::swizzle::<T, 1>(Simd::<T, 1>::from([value]))
}
core::intrinsics::const_eval_select((value,), splat_const, splat_rt)
}
/// Returns an array reference containing the entire SIMD vector.
///
/// # Examples
///
/// ```
/// # #![feature(portable_simd)]
/// # use core::simd::{Simd, u64x4};
/// let v: u64x4 = Simd::from_array([0, 1, 2, 3]);
/// assert_eq!(v.as_array(), &[0, 1, 2, 3]);
/// ```
#[inline]
pub const fn as_array(&self) -> &[T; N] {
// SAFETY: `Simd<T, N>` is just an overaligned `[T; N]` with
// potential padding at the end, so pointer casting to a
// `&[T; N]` is safe.
//
// NOTE: This deliberately doesn't just use `&self.0`, see the comment
// on the struct definition for details.
unsafe { &*(self as *const Self as *const [T; N]) }
}
/// Returns a mutable array reference containing the entire SIMD vector.
#[inline]
pub fn as_mut_array(&mut self) -> &mut [T; N] {
// SAFETY: `Simd<T, N>` is just an overaligned `[T; N]` with
// potential padding at the end, so pointer casting to a
// `&mut [T; N]` is safe.
//
// NOTE: This deliberately doesn't just use `&mut self.0`, see the comment
// on the struct definition for details.
unsafe { &mut *(self as *mut Self as *mut [T; N]) }
}
/// Load a vector from an array of `T`.
///
/// This function is necessary since `repr(simd)` has padding for non-power-of-2 vectors (at the time of writing).
/// With padding, `read_unaligned` will read past the end of an array of N elements.
///
/// # Safety
/// Reading `ptr` must be safe, as if by `<*const [T; N]>::read`.
#[inline]
const unsafe fn load(ptr: *const [T; N]) -> Self {
// There are potentially simpler ways to write this function, but this should result in
// LLVM `load <N x T>`
let mut tmp = core::mem::MaybeUninit::<Self>::uninit();
// SAFETY: `Simd<T, N>` always contains `N` elements of type `T`. It may have padding
// which does not need to be initialized. The safety of reading `ptr` is ensured by the
// caller.
unsafe {
core::ptr::copy_nonoverlapping(ptr, tmp.as_mut_ptr().cast(), 1);
tmp.assume_init()
}
}
/// Store a vector to an array of `T`.
///
/// See `load` as to why this function is necessary.
///
/// # Safety
/// Writing to `ptr` must be safe, as if by `<*mut [T; N]>::write`.
#[inline]
const unsafe fn store(self, ptr: *mut [T; N]) {
// There are potentially simpler ways to write this function, but this should result in
// LLVM `store <N x T>`
// Creating a temporary helps LLVM turn the memcpy into a store.
let tmp = self;
// SAFETY: `Simd<T, N>` always contains `N` elements of type `T`. The safety of writing
// `ptr` is ensured by the caller.
unsafe { core::ptr::copy_nonoverlapping(tmp.as_array(), ptr, 1) }
}
/// Converts an array to a SIMD vector.
#[inline]
pub const fn from_array(array: [T; N]) -> Self {
// SAFETY: `&array` is safe to read.
//
// FIXME: We currently use a pointer load instead of `transmute_copy` because `repr(simd)`
// results in padding for non-power-of-2 vectors (so vectors are larger than arrays).
//
// NOTE: This deliberately doesn't just use `Self(array)`, see the comment
// on the struct definition for details.
unsafe { Self::load(&array) }
}
/// Converts a SIMD vector to an array.
#[inline]
pub const fn to_array(self) -> [T; N] {
let mut tmp = core::mem::MaybeUninit::uninit();
// SAFETY: writing to `tmp` is safe and initializes it.
//
// FIXME: We currently use a pointer store instead of `transmute_copy` because `repr(simd)`
// results in padding for non-power-of-2 vectors (so vectors are larger than arrays).
//
// NOTE: This deliberately doesn't just use `self.0`, see the comment
// on the struct definition for details.
unsafe {
self.store(tmp.as_mut_ptr());
tmp.assume_init()
}
}
/// Converts a slice to a SIMD vector containing `slice[..N]`.
///
/// # Panics
///
/// Panics if the slice's length is less than the vector's `Simd::N`.
/// Use `load_or_default` for an alternative that does not panic.
///
/// # Example
///
/// ```
/// # #![feature(portable_simd)]
/// # use core::simd::u32x4;
/// let source = vec![1, 2, 3, 4, 5, 6];
/// let v = u32x4::from_slice(&source);
/// assert_eq!(v.as_array(), &[1, 2, 3, 4]);
/// ```
#[must_use]
#[inline]
#[track_caller]
pub const fn from_slice(slice: &[T]) -> Self {
assert!(
slice.len() >= Self::LEN,
"slice length must be at least the number of elements"
);
// SAFETY: We just checked that the slice contains
// at least `N` elements.
unsafe { Self::load(slice.as_ptr().cast()) }
}
/// Writes a SIMD vector to the first `N` elements of a slice.
///
/// # Panics
///
/// Panics if the slice's length is less than the vector's `Simd::N`.
///
/// # Example
///
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::u32x4;
/// let mut dest = vec![0; 6];
/// let v = u32x4::from_array([1, 2, 3, 4]);
/// v.copy_to_slice(&mut dest);
/// assert_eq!(&dest, &[1, 2, 3, 4, 0, 0]);
/// ```
#[inline]
#[track_caller]
pub fn copy_to_slice(self, slice: &mut [T]) {
assert!(
slice.len() >= Self::LEN,
"slice length must be at least the number of elements"
);
// SAFETY: We just checked that the slice contains
// at least `N` elements.
unsafe { self.store(slice.as_mut_ptr().cast()) }
}
/// Reads contiguous elements from `slice`. Elements are read so long as they're in-bounds for
/// the `slice`. Otherwise, the default value for the element type is returned.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::Simd;
/// let vec: Vec<i32> = vec![10, 11];
///
/// let result = Simd::<i32, 4>::load_or_default(&vec);
/// assert_eq!(result, Simd::from_array([10, 11, 0, 0]));
/// ```
#[must_use]
#[inline]
pub fn load_or_default(slice: &[T]) -> Self
where
T: Default,
{
Self::load_or(slice, Default::default())
}
/// Reads contiguous elements from `slice`. Elements are read so long as they're in-bounds for
/// the `slice`. Otherwise, the corresponding value from `or` is passed through.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::Simd;
/// let vec: Vec<i32> = vec![10, 11];
/// let or = Simd::from_array([-5, -4, -3, -2]);
///
/// let result = Simd::load_or(&vec, or);
/// assert_eq!(result, Simd::from_array([10, 11, -3, -2]));
/// ```
#[must_use]
#[inline]
pub fn load_or(slice: &[T], or: Self) -> Self {
Self::load_select(slice, Mask::splat(true), or)
}
/// Reads contiguous elements from `slice`. Each element is read from memory if its
/// corresponding element in `enable` is `true`.
///
/// When the element is disabled or out of bounds for the slice, that memory location
/// is not accessed and the corresponding value from `or` is passed through.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::{Simd, Mask};
/// let vec: Vec<i32> = vec![10, 11, 12, 13, 14, 15, 16, 17, 18];
/// let enable = Mask::from_array([true, true, false, true]);
/// let or = Simd::from_array([-5, -4, -3, -2]);
///
/// let result = Simd::load_select(&vec, enable, or);
/// assert_eq!(result, Simd::from_array([10, 11, -3, 13]));
/// ```
#[must_use]
#[inline]
pub fn load_select_or_default(slice: &[T], enable: Mask<<T as SimdElement>::Mask, N>) -> Self
where
T: Default,
{
Self::load_select(slice, enable, Default::default())
}
/// Reads contiguous elements from `slice`. Each element is read from memory if its
/// corresponding element in `enable` is `true`.
///
/// When the element is disabled or out of bounds for the slice, that memory location
/// is not accessed and the corresponding value from `or` is passed through.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::{Simd, Mask};
/// let vec: Vec<i32> = vec![10, 11, 12, 13, 14, 15, 16, 17, 18];
/// let enable = Mask::from_array([true, true, false, true]);
/// let or = Simd::from_array([-5, -4, -3, -2]);
///
/// let result = Simd::load_select(&vec, enable, or);
/// assert_eq!(result, Simd::from_array([10, 11, -3, 13]));
/// ```
#[must_use]
#[inline]
pub fn load_select(
slice: &[T],
mut enable: Mask<<T as SimdElement>::Mask, N>,
or: Self,
) -> Self {
enable &= mask_up_to(slice.len());
// SAFETY: We performed the bounds check by updating the mask. &[T] is properly aligned to
// the element.
unsafe { Self::load_select_ptr(slice.as_ptr(), enable, or) }
}
/// Reads contiguous elements from `slice`. Each element is read from memory if its
/// corresponding element in `enable` is `true`.
///
/// When the element is disabled, that memory location is not accessed and the corresponding
/// value from `or` is passed through.
///
/// # Safety
/// Enabled loads must not exceed the length of `slice`.
#[must_use]
#[inline]
pub unsafe fn load_select_unchecked(
slice: &[T],
enable: Mask<<T as SimdElement>::Mask, N>,
or: Self,
) -> Self {
let ptr = slice.as_ptr();
// SAFETY: The safety of reading elements from `slice` is ensured by the caller.
unsafe { Self::load_select_ptr(ptr, enable, or) }
}
/// Reads contiguous elements starting at `ptr`. Each element is read from memory if its
/// corresponding element in `enable` is `true`.
///
/// When the element is disabled, that memory location is not accessed and the corresponding
/// value from `or` is passed through.
///
/// # Safety
/// Enabled `ptr` elements must be safe to read as if by `std::ptr::read`.
#[must_use]
#[inline]
pub unsafe fn load_select_ptr(
ptr: *const T,
enable: Mask<<T as SimdElement>::Mask, N>,
or: Self,
) -> Self {
// SAFETY: The safety of reading elements through `ptr` is ensured by the caller.
unsafe { core::intrinsics::simd::simd_masked_load(enable.to_int(), ptr, or) }
}
/// Reads from potentially discontiguous indices in `slice` to construct a SIMD vector.
/// If an index is out-of-bounds, the element is instead selected from the `or` vector.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # use core::simd::Simd;
/// let vec: Vec<i32> = vec![10, 11, 12, 13, 14, 15, 16, 17, 18];
/// let idxs = Simd::from_array([9, 3, 0, 5]); // Note the index that is out-of-bounds
/// let alt = Simd::from_array([-5, -4, -3, -2]);
///
/// let result = Simd::gather_or(&vec, idxs, alt);
/// assert_eq!(result, Simd::from_array([-5, 13, 10, 15]));
/// ```
#[must_use]
#[inline]
pub fn gather_or(slice: &[T], idxs: Simd<usize, N>, or: Self) -> Self {
Self::gather_select(slice, Mask::splat(true), idxs, or)
}
/// Reads from indices in `slice` to construct a SIMD vector.
/// If an index is out-of-bounds, the element is set to the default given by `T: Default`.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # use core::simd::Simd;
/// let vec: Vec<i32> = vec![10, 11, 12, 13, 14, 15, 16, 17, 18];
/// let idxs = Simd::from_array([9, 3, 0, 5]); // Note the index that is out-of-bounds
///
/// let result = Simd::gather_or_default(&vec, idxs);
/// assert_eq!(result, Simd::from_array([0, 13, 10, 15]));
/// ```
#[must_use]
#[inline]
pub fn gather_or_default(slice: &[T], idxs: Simd<usize, N>) -> Self
where
T: Default,
{
Self::gather_or(slice, idxs, Self::splat(T::default()))
}
/// Reads from indices in `slice` to construct a SIMD vector.
/// The mask `enable`s all `true` indices and disables all `false` indices.
/// If an index is disabled or is out-of-bounds, the element is selected from the `or` vector.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # use core::simd::{Simd, Mask};
/// let vec: Vec<i32> = vec![10, 11, 12, 13, 14, 15, 16, 17, 18];
/// let idxs = Simd::from_array([9, 3, 0, 5]); // Includes an out-of-bounds index
/// let alt = Simd::from_array([-5, -4, -3, -2]);
/// let enable = Mask::from_array([true, true, true, false]); // Includes a masked element
///
/// let result = Simd::gather_select(&vec, enable, idxs, alt);
/// assert_eq!(result, Simd::from_array([-5, 13, 10, -2]));
/// ```
#[must_use]
#[inline]
pub fn gather_select(
slice: &[T],
enable: Mask<isize, N>,
idxs: Simd<usize, N>,
or: Self,
) -> Self {
let enable: Mask<isize, N> = enable & idxs.simd_lt(Simd::splat(slice.len()));
// Safety: We have masked-off out-of-bounds indices.
unsafe { Self::gather_select_unchecked(slice, enable, idxs, or) }
}
/// Reads from indices in `slice` to construct a SIMD vector.
/// The mask `enable`s all `true` indices and disables all `false` indices.
/// If an index is disabled, the element is selected from the `or` vector.
///
/// # Safety
///
/// Calling this function with an `enable`d out-of-bounds index is *[undefined behavior]*
/// even if the resulting value is not used.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::{Simd, cmp::SimdPartialOrd, Mask};
/// let vec: Vec<i32> = vec![10, 11, 12, 13, 14, 15, 16, 17, 18];
/// let idxs = Simd::from_array([9, 3, 0, 5]); // Includes an out-of-bounds index
/// let alt = Simd::from_array([-5, -4, -3, -2]);
/// let enable = Mask::from_array([true, true, true, false]); // Includes a masked element
/// // If this mask was used to gather, it would be unsound. Let's fix that.
/// let enable = enable & idxs.simd_lt(Simd::splat(vec.len()));
///
/// // The out-of-bounds index has been masked, so it's safe to gather now.
/// let result = unsafe { Simd::gather_select_unchecked(&vec, enable, idxs, alt) };
/// assert_eq!(result, Simd::from_array([-5, 13, 10, -2]));
/// ```
/// [undefined behavior]: https://doc.rust-lang.org/reference/behavior-considered-undefined.html
#[must_use]
#[inline]
#[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces
pub unsafe fn gather_select_unchecked(
slice: &[T],
enable: Mask<isize, N>,
idxs: Simd<usize, N>,
or: Self,
) -> Self {
let base_ptr = Simd::<*const T, N>::splat(slice.as_ptr());
// Ferris forgive me, I have done pointer arithmetic here.
let ptrs = base_ptr.wrapping_add(idxs);
// Safety: The caller is responsible for determining the indices are okay to read
unsafe { Self::gather_select_ptr(ptrs, enable, or) }
}
/// Read elementwise from pointers into a SIMD vector.
///
/// # Safety
///
/// Each read must satisfy the same conditions as [`core::ptr::read`].
///
/// # Example
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::prelude::*;
/// let values = [6, 2, 4, 9];
/// let offsets = Simd::from_array([1, 0, 0, 3]);
/// let source = Simd::splat(values.as_ptr()).wrapping_add(offsets);
/// let gathered = unsafe { Simd::gather_ptr(source) };
/// assert_eq!(gathered, Simd::from_array([2, 6, 6, 9]));
/// ```
#[must_use]
#[inline]
#[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces
pub unsafe fn gather_ptr(source: Simd<*const T, N>) -> Self
where
T: Default,
{
// TODO: add an intrinsic that doesn't use a passthru vector, and remove the T: Default bound
// Safety: The caller is responsible for upholding all invariants
unsafe { Self::gather_select_ptr(source, Mask::splat(true), Self::default()) }
}
/// Conditionally read elementwise from pointers into a SIMD vector.
/// The mask `enable`s all `true` pointers and disables all `false` pointers.
/// If a pointer is disabled, the element is selected from the `or` vector,
/// and no read is performed.
///
/// # Safety
///
/// Enabled elements must satisfy the same conditions as [`core::ptr::read`].
///
/// # Example
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::prelude::*;
/// let values = [6, 2, 4, 9];
/// let enable = Mask::from_array([true, true, false, true]);
/// let offsets = Simd::from_array([1, 0, 0, 3]);
/// let source = Simd::splat(values.as_ptr()).wrapping_add(offsets);
/// let gathered = unsafe { Simd::gather_select_ptr(source, enable, Simd::splat(0)) };
/// assert_eq!(gathered, Simd::from_array([2, 6, 0, 9]));
/// ```
#[must_use]
#[inline]
#[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces
pub unsafe fn gather_select_ptr(
source: Simd<*const T, N>,
enable: Mask<isize, N>,
or: Self,
) -> Self {
// Safety: The caller is responsible for upholding all invariants
unsafe { core::intrinsics::simd::simd_gather(or, source, enable.to_int()) }
}
/// Conditionally write contiguous elements to `slice`. The `enable` mask controls
/// which elements are written, as long as they're in-bounds of the `slice`.
/// If the element is disabled or out of bounds, no memory access to that location
/// is made.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::{Simd, Mask};
/// let mut arr = [0i32; 4];
/// let write = Simd::from_array([-5, -4, -3, -2]);
/// let enable = Mask::from_array([false, true, true, true]);
///
/// write.store_select(&mut arr[..3], enable);
/// assert_eq!(arr, [0, -4, -3, 0]);
/// ```
#[inline]
pub fn store_select(self, slice: &mut [T], mut enable: Mask<<T as SimdElement>::Mask, N>) {
enable &= mask_up_to(slice.len());
// SAFETY: We performed the bounds check by updating the mask. &[T] is properly aligned to
// the element.
unsafe { self.store_select_ptr(slice.as_mut_ptr(), enable) }
}
/// Conditionally write contiguous elements to `slice`. The `enable` mask controls
/// which elements are written.
///
/// # Safety
///
/// Every enabled element must be in bounds for the `slice`.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::{Simd, Mask};
/// let mut arr = [0i32; 4];
/// let write = Simd::from_array([-5, -4, -3, -2]);
/// let enable = Mask::from_array([false, true, true, true]);
///
/// unsafe { write.store_select_unchecked(&mut arr, enable) };
/// assert_eq!(arr, [0, -4, -3, -2]);
/// ```
#[inline]
pub unsafe fn store_select_unchecked(
self,
slice: &mut [T],
enable: Mask<<T as SimdElement>::Mask, N>,
) {
let ptr = slice.as_mut_ptr();
// SAFETY: The safety of writing elements in `slice` is ensured by the caller.
unsafe { self.store_select_ptr(ptr, enable) }
}
/// Conditionally write contiguous elements starting from `ptr`.
/// The `enable` mask controls which elements are written.
/// When disabled, the memory location corresponding to that element is not accessed.
///
/// # Safety
///
/// Memory addresses for element are calculated [`pointer::wrapping_offset`] and
/// each enabled element must satisfy the same conditions as [`core::ptr::write`].
#[inline]
pub unsafe fn store_select_ptr(self, ptr: *mut T, enable: Mask<<T as SimdElement>::Mask, N>) {
// SAFETY: The safety of writing elements through `ptr` is ensured by the caller.
unsafe { core::intrinsics::simd::simd_masked_store(enable.to_int(), ptr, self) }
}
/// Writes the values in a SIMD vector to potentially discontiguous indices in `slice`.
/// If an index is out-of-bounds, the write is suppressed without panicking.
/// If two elements in the scattered vector would write to the same index
/// only the last element is guaranteed to actually be written.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # use core::simd::Simd;
/// let mut vec: Vec<i32> = vec![10, 11, 12, 13, 14, 15, 16, 17, 18];
/// let idxs = Simd::from_array([9, 3, 0, 0]); // Note the duplicate index.
/// let vals = Simd::from_array([-27, 82, -41, 124]);
///
/// vals.scatter(&mut vec, idxs); // two logical writes means the last wins.
/// assert_eq!(vec, vec![124, 11, 12, 82, 14, 15, 16, 17, 18]);
/// ```
#[inline]
pub fn scatter(self, slice: &mut [T], idxs: Simd<usize, N>) {
self.scatter_select(slice, Mask::splat(true), idxs)
}
/// Writes values from a SIMD vector to multiple potentially discontiguous indices in `slice`.
/// The mask `enable`s all `true` indices and disables all `false` indices.
/// If an enabled index is out-of-bounds, the write is suppressed without panicking.
/// If two enabled elements in the scattered vector would write to the same index,
/// only the last element is guaranteed to actually be written.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::{Simd, Mask};
/// let mut vec: Vec<i32> = vec![10, 11, 12, 13, 14, 15, 16, 17, 18];
/// let idxs = Simd::from_array([9, 3, 0, 0]); // Includes an out-of-bounds index
/// let vals = Simd::from_array([-27, 82, -41, 124]);
/// let enable = Mask::from_array([true, true, true, false]); // Includes a masked element
///
/// vals.scatter_select(&mut vec, enable, idxs); // The last write is masked, thus omitted.
/// assert_eq!(vec, vec![-41, 11, 12, 82, 14, 15, 16, 17, 18]);
/// ```
#[inline]
pub fn scatter_select(self, slice: &mut [T], enable: Mask<isize, N>, idxs: Simd<usize, N>) {
let enable: Mask<isize, N> = enable & idxs.simd_lt(Simd::splat(slice.len()));
// Safety: We have masked-off out-of-bounds indices.
unsafe { self.scatter_select_unchecked(slice, enable, idxs) }
}
/// Writes values from a SIMD vector to multiple potentially discontiguous indices in `slice`.
/// The mask `enable`s all `true` indices and disables all `false` indices.
/// If two enabled elements in the scattered vector would write to the same index,
/// only the last element is guaranteed to actually be written.
///
/// # Safety
///
/// Calling this function with an enabled out-of-bounds index is *[undefined behavior]*,
/// and may lead to memory corruption.
///
/// # Examples
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::{Simd, cmp::SimdPartialOrd, Mask};
/// let mut vec: Vec<i32> = vec![10, 11, 12, 13, 14, 15, 16, 17, 18];
/// let idxs = Simd::from_array([9, 3, 0, 0]);
/// let vals = Simd::from_array([-27, 82, -41, 124]);
/// let enable = Mask::from_array([true, true, true, false]); // Masks the final index
/// // If this mask was used to scatter, it would be unsound. Let's fix that.
/// let enable = enable & idxs.simd_lt(Simd::splat(vec.len()));
///
/// // We have masked the OOB index, so it's safe to scatter now.
/// unsafe { vals.scatter_select_unchecked(&mut vec, enable, idxs); }
/// // The second write to index 0 was masked, thus omitted.
/// assert_eq!(vec, vec![-41, 11, 12, 82, 14, 15, 16, 17, 18]);
/// ```
/// [undefined behavior]: https://doc.rust-lang.org/reference/behavior-considered-undefined.html
#[inline]
#[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces
pub unsafe fn scatter_select_unchecked(
self,
slice: &mut [T],
enable: Mask<isize, N>,
idxs: Simd<usize, N>,
) {
// Safety: This block works with *mut T derived from &mut 'a [T],
// which means it is delicate in Rust's borrowing model, circa 2021:
// &mut 'a [T] asserts uniqueness, so deriving &'a [T] invalidates live *mut Ts!
// Even though this block is largely safe methods, it must be exactly this way
// to prevent invalidating the raw ptrs while they're live.
// Thus, entering this block requires all values to use being already ready:
// 0. idxs we want to write to, which are used to construct the mask.
// 1. enable, which depends on an initial &'a [T] and the idxs.
// 2. actual values to scatter (self).
// 3. &mut [T] which will become our base ptr.
unsafe {
// Now Entering ☢️ *mut T Zone
let base_ptr = Simd::<*mut T, N>::splat(slice.as_mut_ptr());
// Ferris forgive me, I have done pointer arithmetic here.
let ptrs = base_ptr.wrapping_add(idxs);
// The ptrs have been bounds-masked to prevent memory-unsafe writes insha'allah
self.scatter_select_ptr(ptrs, enable);
// Cleared ☢️ *mut T Zone
}
}
/// Write pointers elementwise into a SIMD vector.
///
/// # Safety
///
/// Each write must satisfy the same conditions as [`core::ptr::write`].
///
/// # Example
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::{Simd, ptr::SimdMutPtr};
/// let mut values = [0; 4];
/// let offset = Simd::from_array([3, 2, 1, 0]);
/// let ptrs = Simd::splat(values.as_mut_ptr()).wrapping_add(offset);
/// unsafe { Simd::from_array([6, 3, 5, 7]).scatter_ptr(ptrs); }
/// assert_eq!(values, [7, 5, 3, 6]);
/// ```
#[inline]
#[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces
pub unsafe fn scatter_ptr(self, dest: Simd<*mut T, N>) {
// Safety: The caller is responsible for upholding all invariants
unsafe { self.scatter_select_ptr(dest, Mask::splat(true)) }
}
/// Conditionally write pointers elementwise into a SIMD vector.
/// The mask `enable`s all `true` pointers and disables all `false` pointers.
/// If a pointer is disabled, the write to its pointee is skipped.
///
/// # Safety
///
/// Enabled pointers must satisfy the same conditions as [`core::ptr::write`].
///
/// # Example
/// ```
/// # #![feature(portable_simd)]
/// # #[cfg(feature = "as_crate")] use core_simd::simd;
/// # #[cfg(not(feature = "as_crate"))] use core::simd;
/// # use simd::{Mask, Simd, ptr::SimdMutPtr};
/// let mut values = [0; 4];
/// let offset = Simd::from_array([3, 2, 1, 0]);
/// let ptrs = Simd::splat(values.as_mut_ptr()).wrapping_add(offset);
/// let enable = Mask::from_array([true, true, false, false]);
/// unsafe { Simd::from_array([6, 3, 5, 7]).scatter_select_ptr(ptrs, enable); }
/// assert_eq!(values, [0, 0, 3, 6]);
/// ```
#[inline]
#[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces
pub unsafe fn scatter_select_ptr(self, dest: Simd<*mut T, N>, enable: Mask<isize, N>) {
// Safety: The caller is responsible for upholding all invariants
unsafe { core::intrinsics::simd::simd_scatter(self, dest, enable.to_int()) }
}
}
impl<T, const N: usize> Copy for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
}
impl<T, const N: usize> Clone for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
#[inline]
fn clone(&self) -> Self {
*self
}
}
impl<T, const N: usize> Default for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement + Default,
{
#[inline]
fn default() -> Self {
Self::splat(T::default())
}
}
impl<T, const N: usize> PartialEq for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement + PartialEq,
{
#[inline]
fn eq(&self, other: &Self) -> bool {
// Safety: All SIMD vectors are SimdPartialEq, and the comparison produces a valid mask.
let mask = unsafe {
let tfvec: Simd<<T as SimdElement>::Mask, N> =
core::intrinsics::simd::simd_eq(*self, *other);
Mask::from_int_unchecked(tfvec)
};
// Two vectors are equal if all elements are equal when compared elementwise
mask.all()
}
#[allow(clippy::partialeq_ne_impl)]
#[inline]
fn ne(&self, other: &Self) -> bool {
// Safety: All SIMD vectors are SimdPartialEq, and the comparison produces a valid mask.
let mask = unsafe {
let tfvec: Simd<<T as SimdElement>::Mask, N> =
core::intrinsics::simd::simd_ne(*self, *other);
Mask::from_int_unchecked(tfvec)
};
// Two vectors are non-equal if any elements are non-equal when compared elementwise
mask.any()
}
}
/// Lexicographic order. For the SIMD elementwise minimum and maximum, use simd_min and simd_max instead.
impl<T, const N: usize> PartialOrd for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement + PartialOrd,
{
#[inline]
fn partial_cmp(&self, other: &Self) -> Option<core::cmp::Ordering> {
// TODO use SIMD equality
self.to_array().partial_cmp(other.as_ref())
}
}
impl<T, const N: usize> Eq for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement + Eq,
{
}
/// Lexicographic order. For the SIMD elementwise minimum and maximum, use simd_min and simd_max instead.
impl<T, const N: usize> Ord for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement + Ord,
{
#[inline]
fn cmp(&self, other: &Self) -> core::cmp::Ordering {
// TODO use SIMD equality
self.to_array().cmp(other.as_ref())
}
}
impl<T, const N: usize> core::hash::Hash for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement + core::hash::Hash,
{
#[inline]
fn hash<H>(&self, state: &mut H)
where
H: core::hash::Hasher,
{
self.as_array().hash(state)
}
}
// array references
impl<T, const N: usize> AsRef<[T; N]> for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
#[inline]
fn as_ref(&self) -> &[T; N] {
self.as_array()
}
}
impl<T, const N: usize> AsMut<[T; N]> for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
#[inline]
fn as_mut(&mut self) -> &mut [T; N] {
self.as_mut_array()
}
}
// slice references
impl<T, const N: usize> AsRef<[T]> for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
#[inline]
fn as_ref(&self) -> &[T] {
self.as_array()
}
}
impl<T, const N: usize> AsMut<[T]> for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
#[inline]
fn as_mut(&mut self) -> &mut [T] {
self.as_mut_array()
}
}
// vector/array conversion
impl<T, const N: usize> From<[T; N]> for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
#[inline]
fn from(array: [T; N]) -> Self {
Self::from_array(array)
}
}
impl<T, const N: usize> From<Simd<T, N>> for [T; N]
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
#[inline]
fn from(vector: Simd<T, N>) -> Self {
vector.to_array()
}
}
impl<T, const N: usize> TryFrom<&[T]> for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
type Error = core::array::TryFromSliceError;
#[inline]
fn try_from(slice: &[T]) -> Result<Self, core::array::TryFromSliceError> {
Ok(Self::from_array(slice.try_into()?))
}
}
impl<T, const N: usize> TryFrom<&mut [T]> for Simd<T, N>
where
LaneCount<N>: SupportedLaneCount,
T: SimdElement,
{
type Error = core::array::TryFromSliceError;
#[inline]
fn try_from(slice: &mut [T]) -> Result<Self, core::array::TryFromSliceError> {
Ok(Self::from_array(slice.try_into()?))
}
}
mod sealed {
pub trait Sealed {}
}
use sealed::Sealed;
/// Marker trait for types that may be used as SIMD vector elements.
///
/// # Safety
/// This trait, when implemented, asserts the compiler can monomorphize
/// `#[repr(simd)]` structs with the marked type as an element.
/// Strictly, it is valid to impl if the vector will not be miscompiled.
/// Practically, it is user-unfriendly to impl it if the vector won't compile,
/// even when no soundness guarantees are broken by allowing the user to try.
pub unsafe trait SimdElement: Sealed + Copy {
/// The mask element type corresponding to this element type.
type Mask: MaskElement;
}
impl Sealed for u8 {}
// Safety: u8 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for u8 {
type Mask = i8;
}
impl Sealed for u16 {}
// Safety: u16 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for u16 {
type Mask = i16;
}
impl Sealed for u32 {}
// Safety: u32 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for u32 {
type Mask = i32;
}
impl Sealed for u64 {}
// Safety: u64 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for u64 {
type Mask = i64;
}
impl Sealed for usize {}
// Safety: usize is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for usize {
type Mask = isize;
}
impl Sealed for i8 {}
// Safety: i8 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for i8 {
type Mask = i8;
}
impl Sealed for i16 {}
// Safety: i16 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for i16 {
type Mask = i16;
}
impl Sealed for i32 {}
// Safety: i32 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for i32 {
type Mask = i32;
}
impl Sealed for i64 {}
// Safety: i64 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for i64 {
type Mask = i64;
}
impl Sealed for isize {}
// Safety: isize is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for isize {
type Mask = isize;
}
impl Sealed for f32 {}
// Safety: f32 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for f32 {
type Mask = i32;
}
impl Sealed for f64 {}
// Safety: f64 is a valid SIMD element type, and is supported by this API
unsafe impl SimdElement for f64 {
type Mask = i64;
}
impl<T> Sealed for *const T {}
// Safety: (thin) const pointers are valid SIMD element types, and are supported by this API
//
// Fat pointers may be supported in the future.
unsafe impl<T> SimdElement for *const T
where
T: core::ptr::Pointee<Metadata = ()>,
{
type Mask = isize;
}
impl<T> Sealed for *mut T {}
// Safety: (thin) mut pointers are valid SIMD element types, and are supported by this API
//
// Fat pointers may be supported in the future.
unsafe impl<T> SimdElement for *mut T
where
T: core::ptr::Pointee<Metadata = ()>,
{
type Mask = isize;
}
#[inline]
fn lane_indices<const N: usize>() -> Simd<usize, N>
where
LaneCount<N>: SupportedLaneCount,
{
#![allow(clippy::needless_range_loop)]
let mut index = [0; N];
for i in 0..N {
index[i] = i;
}
Simd::from_array(index)
}
#[inline]
fn mask_up_to<M, const N: usize>(len: usize) -> Mask<M, N>
where
LaneCount<N>: SupportedLaneCount,
M: MaskElement,
{
let index = lane_indices::<N>();
let max_value: u64 = M::max_unsigned();
macro_rules! case {
($ty:ty) => {
if N < <$ty>::MAX as usize && max_value as $ty as u64 == max_value {
return index.cast().simd_lt(Simd::splat(len.min(N) as $ty)).cast();
}
};
}
case!(u8);
case!(u16);
case!(u32);
case!(u64);
index.simd_lt(Simd::splat(len)).cast()
}