diff options
Diffstat (limited to 'third_party/rust/encoding_rs/src/simd_funcs.rs')
| -rw-r--r-- | third_party/rust/encoding_rs/src/simd_funcs.rs | 453 |
1 files changed, 453 insertions, 0 deletions
diff --git a/third_party/rust/encoding_rs/src/simd_funcs.rs b/third_party/rust/encoding_rs/src/simd_funcs.rs new file mode 100644 index 0000000000..96feeab5a6 --- /dev/null +++ b/third_party/rust/encoding_rs/src/simd_funcs.rs @@ -0,0 +1,453 @@ +// Copyright Mozilla Foundation. See the COPYRIGHT +// file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// https://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or https://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +use packed_simd::u16x8; +use packed_simd::u8x16; +use packed_simd::IntoBits; + +// TODO: Migrate unaligned access to stdlib code if/when the RFC +// https://github.com/rust-lang/rfcs/pull/1725 is implemented. + +#[inline(always)] +pub unsafe fn load16_unaligned(ptr: *const u8) -> u8x16 { + let mut simd = ::core::mem::uninitialized(); + ::core::ptr::copy_nonoverlapping(ptr, &mut simd as *mut u8x16 as *mut u8, 16); + simd +} + +#[allow(dead_code)] +#[inline(always)] +pub unsafe fn load16_aligned(ptr: *const u8) -> u8x16 { + *(ptr as *const u8x16) +} + +#[inline(always)] +pub unsafe fn store16_unaligned(ptr: *mut u8, s: u8x16) { + ::core::ptr::copy_nonoverlapping(&s as *const u8x16 as *const u8, ptr, 16); +} + +#[allow(dead_code)] +#[inline(always)] +pub unsafe fn store16_aligned(ptr: *mut u8, s: u8x16) { + *(ptr as *mut u8x16) = s; +} + +#[inline(always)] +pub unsafe fn load8_unaligned(ptr: *const u16) -> u16x8 { + let mut simd = ::core::mem::uninitialized(); + ::core::ptr::copy_nonoverlapping(ptr as *const u8, &mut simd as *mut u16x8 as *mut u8, 16); + simd +} + +#[allow(dead_code)] +#[inline(always)] +pub unsafe fn load8_aligned(ptr: *const u16) -> u16x8 { + *(ptr as *const u16x8) +} + +#[inline(always)] +pub unsafe fn store8_unaligned(ptr: *mut u16, s: u16x8) { + ::core::ptr::copy_nonoverlapping(&s as *const u16x8 as *const u8, ptr as *mut u8, 16); +} + +#[allow(dead_code)] +#[inline(always)] +pub unsafe fn store8_aligned(ptr: *mut u16, s: u16x8) { + *(ptr as *mut u16x8) = s; +} + +cfg_if! { + if #[cfg(all(target_feature = "sse2", target_arch = "x86_64"))] { + use core::arch::x86_64::__m128i; + use core::arch::x86_64::_mm_movemask_epi8; + use core::arch::x86_64::_mm_packus_epi16; + } else if #[cfg(all(target_feature = "sse2", target_arch = "x86"))] { + use core::arch::x86::__m128i; + use core::arch::x86::_mm_movemask_epi8; + use core::arch::x86::_mm_packus_epi16; + } else if #[cfg(target_arch = "aarch64")]{ + use core::arch::aarch64::vmaxvq_u8; + use core::arch::aarch64::vmaxvq_u16; + } else { + + } +} + +// #[inline(always)] +// fn simd_byte_swap_u8(s: u8x16) -> u8x16 { +// unsafe { +// shuffle!(s, s, [1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14]) +// } +// } + +// #[inline(always)] +// pub fn simd_byte_swap(s: u16x8) -> u16x8 { +// to_u16_lanes(simd_byte_swap_u8(to_u8_lanes(s))) +// } + +#[inline(always)] +pub fn simd_byte_swap(s: u16x8) -> u16x8 { + let left = s << 8; + let right = s >> 8; + left | right +} + +#[inline(always)] +pub fn to_u16_lanes(s: u8x16) -> u16x8 { + s.into_bits() +} + +cfg_if! { + if #[cfg(target_feature = "sse2")] { + + // Expose low-level mask instead of higher-level conclusion, + // because the non-ASCII case would perform less well otherwise. + #[inline(always)] + pub fn mask_ascii(s: u8x16) -> i32 { + unsafe { + _mm_movemask_epi8(s.into_bits()) + } + } + + } else { + + } +} + +cfg_if! { + if #[cfg(target_feature = "sse2")] { + #[inline(always)] + pub fn simd_is_ascii(s: u8x16) -> bool { + unsafe { + _mm_movemask_epi8(s.into_bits()) == 0 + } + } + } else if #[cfg(target_arch = "aarch64")]{ + #[inline(always)] + pub fn simd_is_ascii(s: u8x16) -> bool { + unsafe { + vmaxvq_u8(s.into_bits()) < 0x80 + } + } + } else { + #[inline(always)] + pub fn simd_is_ascii(s: u8x16) -> bool { + // This optimizes better on ARM than + // the lt formulation. + let highest_ascii = u8x16::splat(0x7F); + !s.gt(highest_ascii).any() + } + } +} + +cfg_if! { + if #[cfg(target_feature = "sse2")] { + #[inline(always)] + pub fn simd_is_str_latin1(s: u8x16) -> bool { + if simd_is_ascii(s) { + return true; + } + let above_str_latin1 = u8x16::splat(0xC4); + s.lt(above_str_latin1).all() + } + } else if #[cfg(target_arch = "aarch64")]{ + #[inline(always)] + pub fn simd_is_str_latin1(s: u8x16) -> bool { + unsafe { + vmaxvq_u8(s.into_bits()) < 0xC4 + } + } + } else { + #[inline(always)] + pub fn simd_is_str_latin1(s: u8x16) -> bool { + let above_str_latin1 = u8x16::splat(0xC4); + s.lt(above_str_latin1).all() + } + } +} + +cfg_if! { + if #[cfg(target_arch = "aarch64")]{ + #[inline(always)] + pub fn simd_is_basic_latin(s: u16x8) -> bool { + unsafe { + vmaxvq_u16(s.into_bits()) < 0x80 + } + } + + #[inline(always)] + pub fn simd_is_latin1(s: u16x8) -> bool { + unsafe { + vmaxvq_u16(s.into_bits()) < 0x100 + } + } + } else { + #[inline(always)] + pub fn simd_is_basic_latin(s: u16x8) -> bool { + let above_ascii = u16x8::splat(0x80); + s.lt(above_ascii).all() + } + + #[inline(always)] + pub fn simd_is_latin1(s: u16x8) -> bool { + // For some reason, on SSE2 this formulation + // seems faster in this case while the above + // function is better the other way round... + let highest_latin1 = u16x8::splat(0xFF); + !s.gt(highest_latin1).any() + } + } +} + +#[inline(always)] +pub fn contains_surrogates(s: u16x8) -> bool { + let mask = u16x8::splat(0xF800); + let surrogate_bits = u16x8::splat(0xD800); + (s & mask).eq(surrogate_bits).any() +} + +cfg_if! { + if #[cfg(target_arch = "aarch64")]{ + macro_rules! aarch64_return_false_if_below_hebrew { + ($s:ident) => ({ + unsafe { + if vmaxvq_u16($s.into_bits()) < 0x0590 { + return false; + } + } + }) + } + + macro_rules! non_aarch64_return_false_if_all { + ($s:ident) => () + } + } else { + macro_rules! aarch64_return_false_if_below_hebrew { + ($s:ident) => () + } + + macro_rules! non_aarch64_return_false_if_all { + ($s:ident) => ({ + if $s.all() { + return false; + } + }) + } + } +} + +macro_rules! in_range16x8 { + ($s:ident, $start:expr, $end:expr) => {{ + // SIMD sub is wrapping + ($s - u16x8::splat($start)).lt(u16x8::splat($end - $start)) + }}; +} + +#[inline(always)] +pub fn is_u16x8_bidi(s: u16x8) -> bool { + // We try to first quickly refute the RTLness of the vector. If that + // fails, we do the real RTL check, so in that case we end up wasting + // the work for the up-front quick checks. Even the quick-check is + // two-fold in order to return `false` ASAP if everything is below + // Hebrew. + + aarch64_return_false_if_below_hebrew!(s); + + let below_hebrew = s.lt(u16x8::splat(0x0590)); + + non_aarch64_return_false_if_all!(below_hebrew); + + if (below_hebrew | in_range16x8!(s, 0x0900, 0x200F) | in_range16x8!(s, 0x2068, 0xD802)).all() { + return false; + } + + // Quick refutation failed. Let's do the full check. + + (in_range16x8!(s, 0x0590, 0x0900) + | in_range16x8!(s, 0xFB1D, 0xFE00) + | in_range16x8!(s, 0xFE70, 0xFEFF) + | in_range16x8!(s, 0xD802, 0xD804) + | in_range16x8!(s, 0xD83A, 0xD83C) + | s.eq(u16x8::splat(0x200F)) + | s.eq(u16x8::splat(0x202B)) + | s.eq(u16x8::splat(0x202E)) + | s.eq(u16x8::splat(0x2067))) + .any() +} + +#[inline(always)] +pub fn simd_unpack(s: u8x16) -> (u16x8, u16x8) { + unsafe { + let first: u8x16 = shuffle!( + s, + u8x16::splat(0), + [0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23] + ); + let second: u8x16 = shuffle!( + s, + u8x16::splat(0), + [8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31] + ); + (first.into_bits(), second.into_bits()) + } +} + +cfg_if! { + if #[cfg(target_feature = "sse2")] { + #[inline(always)] + pub fn simd_pack(a: u16x8, b: u16x8) -> u8x16 { + unsafe { + _mm_packus_epi16(a.into_bits(), b.into_bits()).into_bits() + } + } + } else { + #[inline(always)] + pub fn simd_pack(a: u16x8, b: u16x8) -> u8x16 { + unsafe { + let first: u8x16 = a.into_bits(); + let second: u8x16 = b.into_bits(); + shuffle!( + first, + second, + [0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30] + ) + } + } + } +} + +#[cfg(test)] +mod tests { + use super::*; + use alloc::vec::Vec; + + #[test] + fn test_unpack() { + let ascii: [u8; 16] = [ + 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let basic_latin: [u16; 16] = [ + 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let simd = unsafe { load16_unaligned(ascii.as_ptr()) }; + let mut vec = Vec::with_capacity(16); + vec.resize(16, 0u16); + let (first, second) = simd_unpack(simd); + let ptr = vec.as_mut_ptr(); + unsafe { + store8_unaligned(ptr, first); + store8_unaligned(ptr.add(8), second); + } + assert_eq!(&vec[..], &basic_latin[..]); + } + + #[test] + fn test_simd_is_basic_latin_success() { + let ascii: [u8; 16] = [ + 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let basic_latin: [u16; 16] = [ + 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let first = unsafe { load8_unaligned(basic_latin.as_ptr()) }; + let second = unsafe { load8_unaligned(basic_latin.as_ptr().add(8)) }; + let mut vec = Vec::with_capacity(16); + vec.resize(16, 0u8); + let ptr = vec.as_mut_ptr(); + assert!(simd_is_basic_latin(first | second)); + unsafe { + store16_unaligned(ptr, simd_pack(first, second)); + } + assert_eq!(&vec[..], &ascii[..]); + } + + #[test] + fn test_simd_is_basic_latin_c0() { + let input: [u16; 16] = [ + 0x61, 0x62, 0x63, 0x81, 0x65, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let first = unsafe { load8_unaligned(input.as_ptr()) }; + let second = unsafe { load8_unaligned(input.as_ptr().add(8)) }; + assert!(!simd_is_basic_latin(first | second)); + } + + #[test] + fn test_simd_is_basic_latin_0fff() { + let input: [u16; 16] = [ + 0x61, 0x62, 0x63, 0x0FFF, 0x65, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let first = unsafe { load8_unaligned(input.as_ptr()) }; + let second = unsafe { load8_unaligned(input.as_ptr().add(8)) }; + assert!(!simd_is_basic_latin(first | second)); + } + + #[test] + fn test_simd_is_basic_latin_ffff() { + let input: [u16; 16] = [ + 0x61, 0x62, 0x63, 0xFFFF, 0x65, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let first = unsafe { load8_unaligned(input.as_ptr()) }; + let second = unsafe { load8_unaligned(input.as_ptr().add(8)) }; + assert!(!simd_is_basic_latin(first | second)); + } + + #[test] + fn test_simd_is_ascii_success() { + let ascii: [u8; 16] = [ + 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let simd = unsafe { load16_unaligned(ascii.as_ptr()) }; + assert!(simd_is_ascii(simd)); + } + + #[test] + fn test_simd_is_ascii_failure() { + let input: [u8; 16] = [ + 0x61, 0x62, 0x63, 0x64, 0x81, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let simd = unsafe { load16_unaligned(input.as_ptr()) }; + assert!(!simd_is_ascii(simd)); + } + + #[cfg(target_feature = "sse2")] + #[test] + fn test_check_ascii() { + let input: [u8; 16] = [ + 0x61, 0x62, 0x63, 0x64, 0x81, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let simd = unsafe { load16_unaligned(input.as_ptr()) }; + let mask = mask_ascii(simd); + assert_ne!(mask, 0); + assert_eq!(mask.trailing_zeros(), 4); + } + + #[test] + fn test_alu() { + let input: [u8; 16] = [ + 0x61, 0x62, 0x63, 0x64, 0x81, 0x66, 0x67, 0x68, 0x69, 0x70, 0x71, 0x72, 0x73, 0x74, + 0x75, 0x76, + ]; + let mut alu = 0u64; + unsafe { + ::core::ptr::copy_nonoverlapping(input.as_ptr(), &mut alu as *mut u64 as *mut u8, 8); + } + let masked = alu & 0x8080808080808080; + assert_eq!(masked.trailing_zeros(), 39); + } +} |