diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-17 12:02:58 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-17 12:02:58 +0000 |
commit | 698f8c2f01ea549d77d7dc3338a12e04c11057b9 (patch) | |
tree | 173a775858bd501c378080a10dca74132f05bc50 /library/stdarch/crates/core_arch/src/x86_64/avx2.rs | |
parent | Initial commit. (diff) | |
download | rustc-698f8c2f01ea549d77d7dc3338a12e04c11057b9.tar.xz rustc-698f8c2f01ea549d77d7dc3338a12e04c11057b9.zip |
Adding upstream version 1.64.0+dfsg1.upstream/1.64.0+dfsg1
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'library/stdarch/crates/core_arch/src/x86_64/avx2.rs')
-rw-r--r-- | library/stdarch/crates/core_arch/src/x86_64/avx2.rs | 47 |
1 files changed, 47 insertions, 0 deletions
diff --git a/library/stdarch/crates/core_arch/src/x86_64/avx2.rs b/library/stdarch/crates/core_arch/src/x86_64/avx2.rs new file mode 100644 index 000000000..14447a137 --- /dev/null +++ b/library/stdarch/crates/core_arch/src/x86_64/avx2.rs @@ -0,0 +1,47 @@ +//! Advanced Vector Extensions 2 (AVX) +//! +//! AVX2 expands most AVX commands to 256-bit wide vector registers and +//! adds [FMA](https://en.wikipedia.org/wiki/Fused_multiply-accumulate). +//! +//! The references are: +//! +//! - [Intel 64 and IA-32 Architectures Software Developer's Manual Volume 2: +//! Instruction Set Reference, A-Z][intel64_ref]. +//! - [AMD64 Architecture Programmer's Manual, Volume 3: General-Purpose and +//! System Instructions][amd64_ref]. +//! +//! Wikipedia's [AVX][wiki_avx] and [FMA][wiki_fma] pages provide a quick +//! overview of the instructions available. +//! +//! [intel64_ref]: http://www.intel.de/content/dam/www/public/us/en/documents/manuals/64-ia-32-architectures-software-developer-instruction-set-reference-manual-325383.pdf +//! [amd64_ref]: http://support.amd.com/TechDocs/24594.pdf +//! [wiki_avx]: https://en.wikipedia.org/wiki/Advanced_Vector_Extensions +//! [wiki_fma]: https://en.wikipedia.org/wiki/Fused_multiply-accumulate + +use crate::core_arch::{simd_llvm::*, x86::*}; + +/// Extracts a 64-bit integer from `a`, selected with `INDEX`. +/// +/// [Intel's documentation](https://software.intel.com/sites/landingpage/IntrinsicsGuide/#text=_mm256_extract_epi64) +#[inline] +#[target_feature(enable = "avx2")] +#[rustc_legacy_const_generics(1)] +// This intrinsic has no corresponding instruction. +#[stable(feature = "simd_x86", since = "1.27.0")] +pub unsafe fn _mm256_extract_epi64<const INDEX: i32>(a: __m256i) -> i64 { + static_assert_imm2!(INDEX); + simd_extract(a.as_i64x4(), INDEX as u32) +} + +#[cfg(test)] +mod tests { + use crate::core_arch::arch::x86_64::*; + use stdarch_test::simd_test; + + #[simd_test(enable = "avx2")] + unsafe fn test_mm256_extract_epi64() { + let a = _mm256_setr_epi64x(0, 1, 2, 3); + let r = _mm256_extract_epi64::<3>(a); + assert_eq!(r, 3); + } +} |