Mirrors/rust
1
0
Fork 0
mirror of https://github.com/rust-lang/rust.git synced 2026-05-15 20:45:45 +03:00
Code Issues Packages Projects Releases Wiki Activity

Make avx512fp16 functions const

Browse Source
This commit is contained in:
sayantn
2025-09-23 19:11:35 +05:30
parent 4506af45dd
commit 92098dd82d
2 changed files with 671 additions and 445 deletions
Download Patch File Download Diff File Expand all files Collapse all files
library/stdarch/crates/core_arch/src/x86/avx512fp16.rs
+664 -441
View File
@@ -9,7 +9,8 @@
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_set_ph(
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_set_ph(
e7: f16,
e6: f16,
e5: f16,
@@ -28,7 +29,8 @@ pub fn _mm_set_ph(
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_set_ph(
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_set_ph(
e15: f16,
e14: f16,
e13: f16,
@@ -57,7 +59,8 @@ pub fn _mm256_set_ph(
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_set_ph(
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_set_ph(
e31: f16,
e30: f16,
e29: f16,
@@ -104,7 +107,8 @@ pub fn _mm512_set_ph(
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_set_sh(a: f16) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_set_sh(a: f16) -> __m128h {
__m128h([a, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0])
}
@@ -114,7 +118,8 @@ pub fn _mm_set_sh(a: f16) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_set1_ph(a: f16) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_set1_ph(a: f16) -> __m128h {
unsafe { transmute(f16x8::splat(a)) }
}
@@ -124,7 +129,8 @@ pub fn _mm_set1_ph(a: f16) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_set1_ph(a: f16) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_set1_ph(a: f16) -> __m256h {
unsafe { transmute(f16x16::splat(a)) }
}
@@ -134,7 +140,8 @@ pub fn _mm256_set1_ph(a: f16) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_set1_ph(a: f16) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_set1_ph(a: f16) -> __m512h {
unsafe { transmute(f16x32::splat(a)) }
}
@@ -144,7 +151,8 @@ pub fn _mm512_set1_ph(a: f16) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_setr_ph(
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_setr_ph(
e0: f16,
e1: f16,
e2: f16,
@@ -163,7 +171,8 @@ pub fn _mm_setr_ph(
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_setr_ph(
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_setr_ph(
e0: f16,
e1: f16,
e2: f16,
@@ -192,7 +201,8 @@ pub fn _mm256_setr_ph(
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_setr_ph(
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_setr_ph(
e0: f16,
e1: f16,
e2: f16,
@@ -238,7 +248,8 @@ pub fn _mm512_setr_ph(
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_setzero_ph() -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_setzero_ph() -> __m128h {
unsafe { transmute(f16x8::ZERO) }
}
@@ -248,7 +259,8 @@ pub fn _mm_setzero_ph() -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_setzero_ph() -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_setzero_ph() -> __m256h {
f16x16::ZERO.as_m256h()
}
@@ -258,7 +270,8 @@ pub fn _mm256_setzero_ph() -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_setzero_ph() -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_setzero_ph() -> __m512h {
f16x32::ZERO.as_m512h()
}
@@ -271,7 +284,8 @@ pub fn _mm512_setzero_ph() -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_undefined_ph() -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_undefined_ph() -> __m128h {
f16x8::ZERO.as_m128h()
}
@@ -284,7 +298,8 @@ pub fn _mm_undefined_ph() -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_undefined_ph() -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_undefined_ph() -> __m256h {
f16x16::ZERO.as_m256h()
}
@@ -297,7 +312,8 @@ pub fn _mm256_undefined_ph() -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_undefined_ph() -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_undefined_ph() -> __m512h {
f16x32::ZERO.as_m512h()
}
@@ -308,7 +324,8 @@ pub fn _mm512_undefined_ph() -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_castpd_ph(a: __m128d) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_castpd_ph(a: __m128d) -> __m128h {
unsafe { transmute(a) }
}
@@ -319,7 +336,8 @@ pub fn _mm_castpd_ph(a: __m128d) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_castpd_ph(a: __m256d) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_castpd_ph(a: __m256d) -> __m256h {
unsafe { transmute(a) }
}
@@ -330,7 +348,8 @@ pub fn _mm256_castpd_ph(a: __m256d) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castpd_ph(a: __m512d) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castpd_ph(a: __m512d) -> __m512h {
unsafe { transmute(a) }
}
@@ -341,7 +360,8 @@ pub fn _mm512_castpd_ph(a: __m512d) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_castph_pd(a: __m128h) -> __m128d {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_castph_pd(a: __m128h) -> __m128d {
unsafe { transmute(a) }
}
@@ -352,7 +372,8 @@ pub fn _mm_castph_pd(a: __m128h) -> __m128d {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_castph_pd(a: __m256h) -> __m256d {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_castph_pd(a: __m256h) -> __m256d {
unsafe { transmute(a) }
}
@@ -363,7 +384,8 @@ pub fn _mm256_castph_pd(a: __m256h) -> __m256d {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castph_pd(a: __m512h) -> __m512d {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castph_pd(a: __m512h) -> __m512d {
unsafe { transmute(a) }
}
@@ -374,7 +396,8 @@ pub fn _mm512_castph_pd(a: __m512h) -> __m512d {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_castps_ph(a: __m128) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_castps_ph(a: __m128) -> __m128h {
unsafe { transmute(a) }
}
@@ -385,7 +408,8 @@ pub fn _mm_castps_ph(a: __m128) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_castps_ph(a: __m256) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_castps_ph(a: __m256) -> __m256h {
unsafe { transmute(a) }
}
@@ -396,7 +420,8 @@ pub fn _mm256_castps_ph(a: __m256) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castps_ph(a: __m512) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castps_ph(a: __m512) -> __m512h {
unsafe { transmute(a) }
}
@@ -407,7 +432,8 @@ pub fn _mm512_castps_ph(a: __m512) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_castph_ps(a: __m128h) -> __m128 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_castph_ps(a: __m128h) -> __m128 {
unsafe { transmute(a) }
}
@@ -418,7 +444,8 @@ pub fn _mm_castph_ps(a: __m128h) -> __m128 {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_castph_ps(a: __m256h) -> __m256 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_castph_ps(a: __m256h) -> __m256 {
unsafe { transmute(a) }
}
@@ -429,7 +456,8 @@ pub fn _mm256_castph_ps(a: __m256h) -> __m256 {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castph_ps(a: __m512h) -> __m512 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castph_ps(a: __m512h) -> __m512 {
unsafe { transmute(a) }
}
@@ -440,7 +468,8 @@ pub fn _mm512_castph_ps(a: __m512h) -> __m512 {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_castsi128_ph(a: __m128i) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_castsi128_ph(a: __m128i) -> __m128h {
unsafe { transmute(a) }
}
@@ -451,7 +480,8 @@ pub fn _mm_castsi128_ph(a: __m128i) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_castsi256_ph(a: __m256i) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_castsi256_ph(a: __m256i) -> __m256h {
unsafe { transmute(a) }
}
@@ -462,7 +492,8 @@ pub fn _mm256_castsi256_ph(a: __m256i) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castsi512_ph(a: __m512i) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castsi512_ph(a: __m512i) -> __m512h {
unsafe { transmute(a) }
}
@@ -473,7 +504,8 @@ pub fn _mm512_castsi512_ph(a: __m512i) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_castph_si128(a: __m128h) -> __m128i {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_castph_si128(a: __m128h) -> __m128i {
unsafe { transmute(a) }
}
@@ -484,7 +516,8 @@ pub fn _mm_castph_si128(a: __m128h) -> __m128i {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_castph_si256(a: __m256h) -> __m256i {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_castph_si256(a: __m256h) -> __m256i {
unsafe { transmute(a) }
}
@@ -495,7 +528,8 @@ pub fn _mm256_castph_si256(a: __m256h) -> __m256i {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castph_si512(a: __m512h) -> __m512i {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castph_si512(a: __m512h) -> __m512i {
unsafe { transmute(a) }
}
@@ -506,7 +540,8 @@ pub fn _mm512_castph_si512(a: __m512h) -> __m512i {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_castph256_ph128(a: __m256h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_castph256_ph128(a: __m256h) -> __m128h {
unsafe { simd_shuffle!(a, a, [0, 1, 2, 3, 4, 5, 6, 7]) }
}
@@ -517,7 +552,8 @@ pub fn _mm256_castph256_ph128(a: __m256h) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castph512_ph128(a: __m512h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castph512_ph128(a: __m512h) -> __m128h {
unsafe { simd_shuffle!(a, a, [0, 1, 2, 3, 4, 5, 6, 7]) }
}
@@ -528,7 +564,8 @@ pub fn _mm512_castph512_ph128(a: __m512h) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castph512_ph256(a: __m512h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castph512_ph256(a: __m512h) -> __m256h {
unsafe { simd_shuffle!(a, a, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]) }
}
@@ -540,7 +577,8 @@ pub fn _mm512_castph512_ph256(a: __m512h) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_castph128_ph256(a: __m128h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_castph128_ph256(a: __m128h) -> __m256h {
unsafe {
simd_shuffle!(
a,
@@ -558,7 +596,8 @@ pub fn _mm256_castph128_ph256(a: __m128h) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castph128_ph512(a: __m128h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castph128_ph512(a: __m128h) -> __m512h {
unsafe {
simd_shuffle!(
a,
@@ -579,7 +618,8 @@ pub fn _mm512_castph128_ph512(a: __m128h) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_castph256_ph512(a: __m256h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_castph256_ph512(a: __m256h) -> __m512h {
unsafe {
simd_shuffle!(
a,
@@ -600,7 +640,8 @@ pub fn _mm512_castph256_ph512(a: __m256h) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_zextph128_ph256(a: __m128h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_zextph128_ph256(a: __m128h) -> __m256h {
unsafe {
simd_shuffle!(
a,
@@ -618,7 +659,8 @@ pub fn _mm256_zextph128_ph256(a: __m128h) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_zextph256_ph512(a: __m256h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_zextph256_ph512(a: __m256h) -> __m512h {
unsafe {
simd_shuffle!(
a,
@@ -639,7 +681,8 @@ pub fn _mm512_zextph256_ph512(a: __m256h) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_zextph128_ph512(a: __m128h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_zextph128_ph512(a: __m128h) -> __m512h {
unsafe {
simd_shuffle!(
a,
@@ -1083,7 +1126,8 @@ pub fn _mm_ucomineq_sh(a: __m128h, b: __m128h) -> i32 {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm_load_ph(mem_addr: *const f16) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm_load_ph(mem_addr: *const f16) -> __m128h {
*mem_addr.cast()
}
@@ -1094,7 +1138,8 @@ pub unsafe fn _mm_load_ph(mem_addr: *const f16) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm256_load_ph(mem_addr: *const f16) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm256_load_ph(mem_addr: *const f16) -> __m256h {
*mem_addr.cast()
}
@@ -1105,7 +1150,8 @@ pub unsafe fn _mm256_load_ph(mem_addr: *const f16) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm512_load_ph(mem_addr: *const f16) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm512_load_ph(mem_addr: *const f16) -> __m512h {
*mem_addr.cast()
}
@@ -1116,7 +1162,8 @@ pub unsafe fn _mm512_load_ph(mem_addr: *const f16) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm_load_sh(mem_addr: *const f16) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm_load_sh(mem_addr: *const f16) -> __m128h {
_mm_set_sh(*mem_addr)
}
@@ -1165,7 +1212,8 @@ pub unsafe fn _mm_maskz_load_sh(k: __mmask8, mem_addr: *const f16) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm_loadu_ph(mem_addr: *const f16) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm_loadu_ph(mem_addr: *const f16) -> __m128h {
ptr::read_unaligned(mem_addr.cast())
}
@@ -1176,7 +1224,8 @@ pub unsafe fn _mm_loadu_ph(mem_addr: *const f16) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm256_loadu_ph(mem_addr: *const f16) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm256_loadu_ph(mem_addr: *const f16) -> __m256h {
ptr::read_unaligned(mem_addr.cast())
}
@@ -1187,7 +1236,8 @@ pub unsafe fn _mm256_loadu_ph(mem_addr: *const f16) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm512_loadu_ph(mem_addr: *const f16) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm512_loadu_ph(mem_addr: *const f16) -> __m512h {
ptr::read_unaligned(mem_addr.cast())
}
@@ -1199,7 +1249,8 @@ pub unsafe fn _mm512_loadu_ph(mem_addr: *const f16) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_move_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_move_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let mut mov: f16 = simd_extract!(src, 0);
if (k & 1) != 0 {
@@ -1217,7 +1268,8 @@ pub fn _mm_mask_move_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_move_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_move_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let mut mov: f16 = 0.;
if (k & 1) != 0 {
@@ -1234,7 +1286,8 @@ pub fn _mm_maskz_move_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_move_sh(a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_move_sh(a: __m128h, b: __m128h) -> __m128h {
unsafe {
let mov: f16 = simd_extract!(b, 0);
simd_insert!(a, 0, mov)
@@ -1248,7 +1301,8 @@ pub fn _mm_move_sh(a: __m128h, b: __m128h) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm_store_ph(mem_addr: *mut f16, a: __m128h) {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm_store_ph(mem_addr: *mut f16, a: __m128h) {
*mem_addr.cast() = a;
}
@@ -1259,7 +1313,8 @@ pub unsafe fn _mm_store_ph(mem_addr: *mut f16, a: __m128h) {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm256_store_ph(mem_addr: *mut f16, a: __m256h) {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm256_store_ph(mem_addr: *mut f16, a: __m256h) {
*mem_addr.cast() = a;
}
@@ -1270,7 +1325,8 @@ pub unsafe fn _mm256_store_ph(mem_addr: *mut f16, a: __m256h) {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm512_store_ph(mem_addr: *mut f16, a: __m512h) {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm512_store_ph(mem_addr: *mut f16, a: __m512h) {
*mem_addr.cast() = a;
}
@@ -1280,7 +1336,8 @@ pub unsafe fn _mm512_store_ph(mem_addr: *mut f16, a: __m512h) {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm_store_sh(mem_addr: *mut f16, a: __m128h) {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm_store_sh(mem_addr: *mut f16, a: __m128h) {
*mem_addr = simd_extract!(a, 0);
}
@@ -1307,7 +1364,8 @@ pub unsafe fn _mm_mask_store_sh(mem_addr: *mut f16, k: __mmask8, a: __m128h) {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm_storeu_ph(mem_addr: *mut f16, a: __m128h) {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm_storeu_ph(mem_addr: *mut f16, a: __m128h) {
ptr::write_unaligned(mem_addr.cast(), a);
}
@@ -1318,7 +1376,8 @@ pub unsafe fn _mm_storeu_ph(mem_addr: *mut f16, a: __m128h) {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm256_storeu_ph(mem_addr: *mut f16, a: __m256h) {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm256_storeu_ph(mem_addr: *mut f16, a: __m256h) {
ptr::write_unaligned(mem_addr.cast(), a);
}
@@ -1329,7 +1388,8 @@ pub unsafe fn _mm256_storeu_ph(mem_addr: *mut f16, a: __m256h) {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub unsafe fn _mm512_storeu_ph(mem_addr: *mut f16, a: __m512h) {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const unsafe fn _mm512_storeu_ph(mem_addr: *mut f16, a: __m512h) {
ptr::write_unaligned(mem_addr.cast(), a);
}
@@ -1340,7 +1400,8 @@ pub unsafe fn _mm512_storeu_ph(mem_addr: *mut f16, a: __m512h) {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vaddph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_add_ph(a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_add_ph(a: __m128h, b: __m128h) -> __m128h {
unsafe { simd_add(a, b) }
}
@@ -1352,7 +1413,8 @@ pub fn _mm_add_ph(a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vaddph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_add_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_add_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let r = _mm_add_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -1367,7 +1429,8 @@ pub fn _mm_mask_add_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vaddph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_add_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_add_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let r = _mm_add_ph(a, b);
simd_select_bitmask(k, r, _mm_setzero_ph())
@@ -1381,7 +1444,8 @@ pub fn _mm_maskz_add_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vaddph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_add_ph(a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_add_ph(a: __m256h, b: __m256h) -> __m256h {
unsafe { simd_add(a, b) }
}
@@ -1393,7 +1457,8 @@ pub fn _mm256_add_ph(a: __m256h, b: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vaddph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_add_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_add_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
unsafe {
let r = _mm256_add_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -1408,7 +1473,8 @@ pub fn _mm256_mask_add_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vaddph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_add_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_add_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
unsafe {
let r = _mm256_add_ph(a, b);
simd_select_bitmask(k, r, _mm256_setzero_ph())
@@ -1422,7 +1488,8 @@ pub fn _mm256_maskz_add_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vaddph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_add_ph(a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_add_ph(a: __m512h, b: __m512h) -> __m512h {
unsafe { simd_add(a, b) }
}
@@ -1434,7 +1501,8 @@ pub fn _mm512_add_ph(a: __m512h, b: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vaddph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_add_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_add_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
unsafe {
let r = _mm512_add_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -1449,7 +1517,8 @@ pub fn _mm512_mask_add_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) ->
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vaddph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_add_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_add_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
unsafe {
let r = _mm512_add_ph(a, b);
simd_select_bitmask(k, r, _mm512_setzero_ph())
@@ -1614,7 +1683,8 @@ pub fn _mm_maskz_add_round_sh<const ROUNDING: i32>(k: __mmask8, a: __m128h, b: _
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vaddsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_add_sh(a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_add_sh(a: __m128h, b: __m128h) -> __m128h {
unsafe { simd_insert!(a, 0, _mm_cvtsh_h(a) + _mm_cvtsh_h(b)) }
}
@@ -1627,7 +1697,8 @@ pub fn _mm_add_sh(a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vaddsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_add_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_add_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let extractsrc: f16 = simd_extract!(src, 0);
let mut add: f16 = extractsrc;
@@ -1649,7 +1720,8 @@ pub fn _mm_mask_add_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vaddsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_add_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_add_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let mut add: f16 = 0.;
if (k & 0b00000001) != 0 {
@@ -1668,7 +1740,8 @@ pub fn _mm_maskz_add_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vsubph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_sub_ph(a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_sub_ph(a: __m128h, b: __m128h) -> __m128h {
unsafe { simd_sub(a, b) }
}
@@ -1680,7 +1753,8 @@ pub fn _mm_sub_ph(a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vsubph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_sub_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_sub_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let r = _mm_sub_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -1695,7 +1769,8 @@ pub fn _mm_mask_sub_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vsubph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_sub_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_sub_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let r = _mm_sub_ph(a, b);
simd_select_bitmask(k, r, _mm_setzero_ph())
@@ -1709,7 +1784,8 @@ pub fn _mm_maskz_sub_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vsubph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_sub_ph(a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_sub_ph(a: __m256h, b: __m256h) -> __m256h {
unsafe { simd_sub(a, b) }
}
@@ -1721,7 +1797,8 @@ pub fn _mm256_sub_ph(a: __m256h, b: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vsubph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_sub_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_sub_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
unsafe {
let r = _mm256_sub_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -1736,7 +1813,8 @@ pub fn _mm256_mask_sub_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vsubph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_sub_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_sub_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
unsafe {
let r = _mm256_sub_ph(a, b);
simd_select_bitmask(k, r, _mm256_setzero_ph())
@@ -1750,7 +1828,8 @@ pub fn _mm256_maskz_sub_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vsubph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_sub_ph(a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_sub_ph(a: __m512h, b: __m512h) -> __m512h {
unsafe { simd_sub(a, b) }
}
@@ -1762,7 +1841,8 @@ pub fn _mm512_sub_ph(a: __m512h, b: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vsubph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_sub_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_sub_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
unsafe {
let r = _mm512_sub_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -1777,7 +1857,8 @@ pub fn _mm512_mask_sub_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) ->
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vsubph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_sub_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_sub_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
unsafe {
let r = _mm512_sub_ph(a, b);
simd_select_bitmask(k, r, _mm512_setzero_ph())
@@ -1943,7 +2024,8 @@ pub fn _mm_maskz_sub_round_sh<const ROUNDING: i32>(k: __mmask8, a: __m128h, b: _
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vsubsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_sub_sh(a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_sub_sh(a: __m128h, b: __m128h) -> __m128h {
unsafe { simd_insert!(a, 0, _mm_cvtsh_h(a) - _mm_cvtsh_h(b)) }
}
@@ -1956,7 +2038,8 @@ pub fn _mm_sub_sh(a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vsubsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_sub_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_sub_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let extractsrc: f16 = simd_extract!(src, 0);
let mut add: f16 = extractsrc;
@@ -1978,7 +2061,8 @@ pub fn _mm_mask_sub_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vsubsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_sub_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_sub_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let mut add: f16 = 0.;
if (k & 0b00000001) != 0 {
@@ -1997,7 +2081,8 @@ pub fn _mm_maskz_sub_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vmulph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mul_ph(a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mul_ph(a: __m128h, b: __m128h) -> __m128h {
unsafe { simd_mul(a, b) }
}
@@ -2009,7 +2094,8 @@ pub fn _mm_mul_ph(a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vmulph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_mul_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_mul_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let r = _mm_mul_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -2024,7 +2110,8 @@ pub fn _mm_mask_mul_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vmulph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_mul_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_mul_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let r = _mm_mul_ph(a, b);
simd_select_bitmask(k, r, _mm_setzero_ph())
@@ -2038,7 +2125,8 @@ pub fn _mm_maskz_mul_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vmulph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mul_ph(a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mul_ph(a: __m256h, b: __m256h) -> __m256h {
unsafe { simd_mul(a, b) }
}
@@ -2050,7 +2138,8 @@ pub fn _mm256_mul_ph(a: __m256h, b: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vmulph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_mul_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_mul_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
unsafe {
let r = _mm256_mul_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -2065,7 +2154,8 @@ pub fn _mm256_mask_mul_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vmulph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_mul_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_mul_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
unsafe {
let r = _mm256_mul_ph(a, b);
simd_select_bitmask(k, r, _mm256_setzero_ph())
@@ -2079,7 +2169,8 @@ pub fn _mm256_maskz_mul_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vmulph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mul_ph(a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mul_ph(a: __m512h, b: __m512h) -> __m512h {
unsafe { simd_mul(a, b) }
}
@@ -2091,7 +2182,8 @@ pub fn _mm512_mul_ph(a: __m512h, b: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vmulph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_mul_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_mul_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
unsafe {
let r = _mm512_mul_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -2106,7 +2198,8 @@ pub fn _mm512_mask_mul_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) ->
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vmulph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_mul_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_mul_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
unsafe {
let r = _mm512_mul_ph(a, b);
simd_select_bitmask(k, r, _mm512_setzero_ph())
@@ -2272,7 +2365,8 @@ pub fn _mm_maskz_mul_round_sh<const ROUNDING: i32>(k: __mmask8, a: __m128h, b: _
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vmulsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mul_sh(a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mul_sh(a: __m128h, b: __m128h) -> __m128h {
unsafe { simd_insert!(a, 0, _mm_cvtsh_h(a) * _mm_cvtsh_h(b)) }
}
@@ -2285,7 +2379,8 @@ pub fn _mm_mul_sh(a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vmulsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_mul_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_mul_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let extractsrc: f16 = simd_extract!(src, 0);
let mut add: f16 = extractsrc;
@@ -2307,7 +2402,8 @@ pub fn _mm_mask_mul_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vmulsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_mul_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_mul_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let mut add: f16 = 0.;
if (k & 0b00000001) != 0 {
@@ -2326,7 +2422,8 @@ pub fn _mm_maskz_mul_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vdivph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_div_ph(a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_div_ph(a: __m128h, b: __m128h) -> __m128h {
unsafe { simd_div(a, b) }
}
@@ -2338,7 +2435,8 @@ pub fn _mm_div_ph(a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vdivph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_div_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_div_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let r = _mm_div_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -2353,7 +2451,8 @@ pub fn _mm_mask_div_ph(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vdivph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_div_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_div_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let r = _mm_div_ph(a, b);
simd_select_bitmask(k, r, _mm_setzero_ph())
@@ -2367,7 +2466,8 @@ pub fn _mm_maskz_div_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vdivph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_div_ph(a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_div_ph(a: __m256h, b: __m256h) -> __m256h {
unsafe { simd_div(a, b) }
}
@@ -2379,7 +2479,8 @@ pub fn _mm256_div_ph(a: __m256h, b: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vdivph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_div_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_div_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
unsafe {
let r = _mm256_div_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -2394,7 +2495,8 @@ pub fn _mm256_mask_div_ph(src: __m256h, k: __mmask16, a: __m256h, b: __m256h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vdivph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_div_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_div_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
unsafe {
let r = _mm256_div_ph(a, b);
simd_select_bitmask(k, r, _mm256_setzero_ph())
@@ -2408,7 +2510,8 @@ pub fn _mm256_maskz_div_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vdivph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_div_ph(a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_div_ph(a: __m512h, b: __m512h) -> __m512h {
unsafe { simd_div(a, b) }
}
@@ -2420,7 +2523,8 @@ pub fn _mm512_div_ph(a: __m512h, b: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vdivph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_div_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_div_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
unsafe {
let r = _mm512_div_ph(a, b);
simd_select_bitmask(k, r, src)
@@ -2435,7 +2539,8 @@ pub fn _mm512_mask_div_ph(src: __m512h, k: __mmask32, a: __m512h, b: __m512h) ->
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vdivph))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_div_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_div_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
unsafe {
let r = _mm512_div_ph(a, b);
simd_select_bitmask(k, r, _mm512_setzero_ph())
@@ -2601,7 +2706,8 @@ pub fn _mm_maskz_div_round_sh<const ROUNDING: i32>(k: __mmask8, a: __m128h, b: _
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vdivsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_div_sh(a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_div_sh(a: __m128h, b: __m128h) -> __m128h {
unsafe { simd_insert!(a, 0, _mm_cvtsh_h(a) / _mm_cvtsh_h(b)) }
}
@@ -2614,7 +2720,8 @@ pub fn _mm_div_sh(a: __m128h, b: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vdivsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_div_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_div_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let extractsrc: f16 = simd_extract!(src, 0);
let mut add: f16 = extractsrc;
@@ -2636,7 +2743,8 @@ pub fn _mm_mask_div_sh(src: __m128h, k: __mmask8, a: __m128h, b: __m128h) -> __m
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vdivsh))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_div_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_div_sh(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe {
let mut add: f16 = 0.;
if (k & 0b00000001) != 0 {
@@ -3967,7 +4075,8 @@ pub fn _mm_maskz_fcmul_round_sch<const ROUNDING: i32>(
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_abs_ph(v2: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_abs_ph(v2: __m128h) -> __m128h {
unsafe { transmute(_mm_and_si128(transmute(v2), _mm_set1_epi16(i16::MAX))) }
}
@@ -3978,7 +4087,8 @@ pub fn _mm_abs_ph(v2: __m128h) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_abs_ph(v2: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_abs_ph(v2: __m256h) -> __m256h {
unsafe { transmute(_mm256_and_si256(transmute(v2), _mm256_set1_epi16(i16::MAX))) }
}
@@ -3989,7 +4099,8 @@ pub fn _mm256_abs_ph(v2: __m256h) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_abs_ph(v2: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_abs_ph(v2: __m512h) -> __m512h {
unsafe { transmute(_mm512_and_si512(transmute(v2), _mm512_set1_epi16(i16::MAX))) }
}
@@ -4002,7 +4113,8 @@ pub fn _mm512_abs_ph(v2: __m512h) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_conj_pch(a: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_conj_pch(a: __m128h) -> __m128h {
unsafe { transmute(_mm_xor_si128(transmute(a), _mm_set1_epi32(i32::MIN))) }
}
@@ -4015,7 +4127,8 @@ pub fn _mm_conj_pch(a: __m128h) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_conj_pch(src: __m128h, k: __mmask8, a: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_conj_pch(src: __m128h, k: __mmask8, a: __m128h) -> __m128h {
unsafe {
let r: __m128 = transmute(_mm_conj_pch(a));
transmute(simd_select_bitmask(k, r, transmute(src)))
@@ -4031,7 +4144,8 @@ pub fn _mm_mask_conj_pch(src: __m128h, k: __mmask8, a: __m128h) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_conj_pch(k: __mmask8, a: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_conj_pch(k: __mmask8, a: __m128h) -> __m128h {
_mm_mask_conj_pch(_mm_setzero_ph(), k, a)
}
@@ -4043,7 +4157,8 @@ pub fn _mm_maskz_conj_pch(k: __mmask8, a: __m128h) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_conj_pch(a: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_conj_pch(a: __m256h) -> __m256h {
unsafe { transmute(_mm256_xor_si256(transmute(a), _mm256_set1_epi32(i32::MIN))) }
}
@@ -4056,7 +4171,8 @@ pub fn _mm256_conj_pch(a: __m256h) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_conj_pch(src: __m256h, k: __mmask8, a: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_conj_pch(src: __m256h, k: __mmask8, a: __m256h) -> __m256h {
unsafe {
let r: __m256 = transmute(_mm256_conj_pch(a));
transmute(simd_select_bitmask(k, r, transmute(src)))
@@ -4072,7 +4188,8 @@ pub fn _mm256_mask_conj_pch(src: __m256h, k: __mmask8, a: __m256h) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_conj_pch(k: __mmask8, a: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_conj_pch(k: __mmask8, a: __m256h) -> __m256h {
_mm256_mask_conj_pch(_mm256_setzero_ph(), k, a)
}
@@ -4084,7 +4201,8 @@ pub fn _mm256_maskz_conj_pch(k: __mmask8, a: __m256h) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_conj_pch(a: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_conj_pch(a: __m512h) -> __m512h {
unsafe { transmute(_mm512_xor_si512(transmute(a), _mm512_set1_epi32(i32::MIN))) }
}
@@ -4097,7 +4215,8 @@ pub fn _mm512_conj_pch(a: __m512h) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_conj_pch(src: __m512h, k: __mmask16, a: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_conj_pch(src: __m512h, k: __mmask16, a: __m512h) -> __m512h {
unsafe {
let r: __m512 = transmute(_mm512_conj_pch(a));
transmute(simd_select_bitmask(k, r, transmute(src)))
@@ -4113,7 +4232,8 @@ pub fn _mm512_mask_conj_pch(src: __m512h, k: __mmask16, a: __m512h) -> __m512h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_conj_pch(k: __mmask16, a: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_conj_pch(k: __mmask16, a: __m512h) -> __m512h {
_mm512_mask_conj_pch(_mm512_setzero_ph(), k, a)
}
@@ -5198,7 +5318,8 @@ pub fn _mm_maskz_fcmadd_round_sch<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fmadd_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fmadd_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_fma(a, b, c) }
}
@@ -5211,7 +5332,8 @@ pub fn _mm_fmadd_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fmadd_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fmadd_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmadd_ph(a, b, c), a) }
}
@@ -5224,7 +5346,8 @@ pub fn _mm_mask_fmadd_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fmadd_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fmadd_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmadd_ph(a, b, c), c) }
}
@@ -5237,7 +5360,8 @@ pub fn _mm_mask3_fmadd_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fmadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fmadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmadd_ph(a, b, c), _mm_setzero_ph()) }
}
@@ -5249,7 +5373,8 @@ pub fn _mm_maskz_fmadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_fmadd_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_fmadd_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_fma(a, b, c) }
}
@@ -5262,7 +5387,8 @@ pub fn _mm256_fmadd_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_fmadd_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_fmadd_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmadd_ph(a, b, c), a) }
}
@@ -5275,7 +5401,8 @@ pub fn _mm256_mask_fmadd_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask3_fmadd_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask3_fmadd_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmadd_ph(a, b, c), c) }
}
@@ -5288,7 +5415,8 @@ pub fn _mm256_mask3_fmadd_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_fmadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_fmadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmadd_ph(a, b, c), _mm256_setzero_ph()) }
}
@@ -5300,7 +5428,8 @@ pub fn _mm256_maskz_fmadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_fmadd_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_fmadd_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_fma(a, b, c) }
}
@@ -5313,7 +5442,8 @@ pub fn _mm512_fmadd_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_fmadd_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_fmadd_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmadd_ph(a, b, c), a) }
}
@@ -5326,7 +5456,8 @@ pub fn _mm512_mask_fmadd_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) ->
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask3_fmadd_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask3_fmadd_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmadd_ph(a, b, c), c) }
}
@@ -5339,7 +5470,8 @@ pub fn _mm512_mask3_fmadd_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_fmadd_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_fmadd_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmadd_ph(a, b, c), _mm512_setzero_ph()) }
}
@@ -5470,7 +5602,8 @@ pub fn _mm512_maskz_fmadd_round_ph<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fmadd_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fmadd_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let extracta: f16 = simd_extract!(a, 0);
let extractb: f16 = simd_extract!(b, 0);
@@ -5490,7 +5623,8 @@ pub fn _mm_fmadd_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fmadd_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fmadd_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let mut fmadd: f16 = simd_extract!(a, 0);
if k & 1 != 0 {
@@ -5512,7 +5646,8 @@ pub fn _mm_mask_fmadd_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fmadd_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fmadd_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe {
let mut fmadd: f16 = simd_extract!(c, 0);
if k & 1 != 0 {
@@ -5534,7 +5669,8 @@ pub fn _mm_mask3_fmadd_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fmadd_sh(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fmadd_sh(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let mut fmadd: f16 = 0.0;
if k & 1 != 0 {
@@ -5697,7 +5833,8 @@ pub fn _mm_maskz_fmadd_round_sh<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fmsub_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fmsub_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_fma(a, b, simd_neg(c)) }
}
@@ -5710,7 +5847,8 @@ pub fn _mm_fmsub_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fmsub_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fmsub_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmsub_ph(a, b, c), a) }
}
@@ -5723,7 +5861,8 @@ pub fn _mm_mask_fmsub_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fmsub_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fmsub_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmsub_ph(a, b, c), c) }
}
@@ -5736,7 +5875,8 @@ pub fn _mm_mask3_fmsub_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fmsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fmsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmsub_ph(a, b, c), _mm_setzero_ph()) }
}
@@ -5748,7 +5888,8 @@ pub fn _mm_maskz_fmsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_fmsub_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_fmsub_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_fma(a, b, simd_neg(c)) }
}
@@ -5761,7 +5902,8 @@ pub fn _mm256_fmsub_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_fmsub_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_fmsub_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmsub_ph(a, b, c), a) }
}
@@ -5774,7 +5916,8 @@ pub fn _mm256_mask_fmsub_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask3_fmsub_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask3_fmsub_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmsub_ph(a, b, c), c) }
}
@@ -5787,7 +5930,8 @@ pub fn _mm256_mask3_fmsub_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_fmsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_fmsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmsub_ph(a, b, c), _mm256_setzero_ph()) }
}
@@ -5799,7 +5943,8 @@ pub fn _mm256_maskz_fmsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_fmsub_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_fmsub_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_fma(a, b, simd_neg(c)) }
}
@@ -5812,7 +5957,8 @@ pub fn _mm512_fmsub_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_fmsub_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_fmsub_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmsub_ph(a, b, c), a) }
}
@@ -5825,7 +5971,8 @@ pub fn _mm512_mask_fmsub_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) ->
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask3_fmsub_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask3_fmsub_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmsub_ph(a, b, c), c) }
}
@@ -5838,7 +5985,8 @@ pub fn _mm512_mask3_fmsub_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_fmsub_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_fmsub_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmsub_ph(a, b, c), _mm512_setzero_ph()) }
}
@@ -5969,7 +6117,8 @@ pub fn _mm512_maskz_fmsub_round_ph<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fmsub_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fmsub_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let extracta: f16 = simd_extract!(a, 0);
let extractb: f16 = simd_extract!(b, 0);
@@ -5989,7 +6138,8 @@ pub fn _mm_fmsub_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fmsub_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fmsub_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let mut fmsub: f16 = simd_extract!(a, 0);
if k & 1 != 0 {
@@ -6011,7 +6161,8 @@ pub fn _mm_mask_fmsub_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fmsub_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fmsub_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe {
let mut fmsub: f16 = simd_extract!(c, 0);
if k & 1 != 0 {
@@ -6033,7 +6184,8 @@ pub fn _mm_mask3_fmsub_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fmsub_sh(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fmsub_sh(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let mut fmsub: f16 = 0.0;
if k & 1 != 0 {
@@ -6187,7 +6339,8 @@ pub fn _mm_maskz_fmsub_round_sh<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fnmadd_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fnmadd_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_fma(simd_neg(a), b, c) }
}
@@ -6200,7 +6353,8 @@ pub fn _mm_fnmadd_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fnmadd_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fnmadd_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fnmadd_ph(a, b, c), a) }
}
@@ -6213,7 +6367,8 @@ pub fn _mm_mask_fnmadd_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fnmadd_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fnmadd_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fnmadd_ph(a, b, c), c) }
}
@@ -6226,7 +6381,8 @@ pub fn _mm_mask3_fnmadd_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> _
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fnmadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fnmadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fnmadd_ph(a, b, c), _mm_setzero_ph()) }
}
@@ -6238,7 +6394,8 @@ pub fn _mm_maskz_fnmadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> _
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_fnmadd_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_fnmadd_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_fma(simd_neg(a), b, c) }
}
@@ -6251,7 +6408,8 @@ pub fn _mm256_fnmadd_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_fnmadd_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_fnmadd_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fnmadd_ph(a, b, c), a) }
}
@@ -6264,7 +6422,8 @@ pub fn _mm256_mask_fnmadd_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask3_fnmadd_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask3_fnmadd_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fnmadd_ph(a, b, c), c) }
}
@@ -6277,7 +6436,8 @@ pub fn _mm256_mask3_fnmadd_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16)
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_fnmadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_fnmadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fnmadd_ph(a, b, c), _mm256_setzero_ph()) }
}
@@ -6289,7 +6449,8 @@ pub fn _mm256_maskz_fnmadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h)
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_fnmadd_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_fnmadd_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_fma(simd_neg(a), b, c) }
}
@@ -6302,7 +6463,8 @@ pub fn _mm512_fnmadd_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_fnmadd_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_fnmadd_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fnmadd_ph(a, b, c), a) }
}
@@ -6315,7 +6477,8 @@ pub fn _mm512_mask_fnmadd_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask3_fnmadd_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask3_fnmadd_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fnmadd_ph(a, b, c), c) }
}
@@ -6328,7 +6491,8 @@ pub fn _mm512_mask3_fnmadd_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32)
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_fnmadd_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_fnmadd_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fnmadd_ph(a, b, c), _mm512_setzero_ph()) }
}
@@ -6459,7 +6623,8 @@ pub fn _mm512_maskz_fnmadd_round_ph<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fnmadd_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fnmadd_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let extracta: f16 = simd_extract!(a, 0);
let extractb: f16 = simd_extract!(b, 0);
@@ -6479,7 +6644,8 @@ pub fn _mm_fnmadd_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fnmadd_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fnmadd_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let mut fnmadd: f16 = simd_extract!(a, 0);
if k & 1 != 0 {
@@ -6501,7 +6667,8 @@ pub fn _mm_mask_fnmadd_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fnmadd_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fnmadd_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe {
let mut fnmadd: f16 = simd_extract!(c, 0);
if k & 1 != 0 {
@@ -6523,7 +6690,8 @@ pub fn _mm_mask3_fnmadd_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> _
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fnmadd_sh(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fnmadd_sh(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let mut fnmadd: f16 = 0.0;
if k & 1 != 0 {
@@ -6685,7 +6853,8 @@ pub fn _mm_maskz_fnmadd_round_sh<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fnmsub_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fnmsub_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_fma(simd_neg(a), b, simd_neg(c)) }
}
@@ -6698,7 +6867,8 @@ pub fn _mm_fnmsub_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fnmsub_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fnmsub_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fnmsub_ph(a, b, c), a) }
}
@@ -6711,7 +6881,8 @@ pub fn _mm_mask_fnmsub_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fnmsub_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fnmsub_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fnmsub_ph(a, b, c), c) }
}
@@ -6724,7 +6895,8 @@ pub fn _mm_mask3_fnmsub_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> _
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fnmsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fnmsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fnmsub_ph(a, b, c), _mm_setzero_ph()) }
}
@@ -6736,7 +6908,8 @@ pub fn _mm_maskz_fnmsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> _
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_fnmsub_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_fnmsub_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_fma(simd_neg(a), b, simd_neg(c)) }
}
@@ -6749,7 +6922,8 @@ pub fn _mm256_fnmsub_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_fnmsub_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_fnmsub_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fnmsub_ph(a, b, c), a) }
}
@@ -6762,7 +6936,8 @@ pub fn _mm256_mask_fnmsub_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask3_fnmsub_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask3_fnmsub_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fnmsub_ph(a, b, c), c) }
}
@@ -6775,7 +6950,8 @@ pub fn _mm256_mask3_fnmsub_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16)
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_fnmsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_fnmsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fnmsub_ph(a, b, c), _mm256_setzero_ph()) }
}
@@ -6787,7 +6963,8 @@ pub fn _mm256_maskz_fnmsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h)
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_fnmsub_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_fnmsub_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_fma(simd_neg(a), b, simd_neg(c)) }
}
@@ -6800,7 +6977,8 @@ pub fn _mm512_fnmsub_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_fnmsub_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_fnmsub_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fnmsub_ph(a, b, c), a) }
}
@@ -6813,7 +6991,8 @@ pub fn _mm512_mask_fnmsub_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask3_fnmsub_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask3_fnmsub_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fnmsub_ph(a, b, c), c) }
}
@@ -6826,7 +7005,8 @@ pub fn _mm512_mask3_fnmsub_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32)
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_fnmsub_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_fnmsub_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fnmsub_ph(a, b, c), _mm512_setzero_ph()) }
}
@@ -6957,7 +7137,8 @@ pub fn _mm512_maskz_fnmsub_round_ph<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fnmsub_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fnmsub_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let extracta: f16 = simd_extract!(a, 0);
let extractb: f16 = simd_extract!(b, 0);
@@ -6977,7 +7158,8 @@ pub fn _mm_fnmsub_sh(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fnmsub_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fnmsub_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let mut fnmsub: f16 = simd_extract!(a, 0);
if k & 1 != 0 {
@@ -6999,7 +7181,8 @@ pub fn _mm_mask_fnmsub_sh(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fnmsub_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fnmsub_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe {
let mut fnmsub: f16 = simd_extract!(c, 0);
if k & 1 != 0 {
@@ -7021,7 +7204,8 @@ pub fn _mm_mask3_fnmsub_sh(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> _
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfnmsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fnmsub_sh(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fnmsub_sh(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let mut fnmsub: f16 = 0.0;
if k & 1 != 0 {
@@ -7183,7 +7367,8 @@ pub fn _mm_maskz_fnmsub_round_sh<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fmaddsub_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fmaddsub_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe {
let add = simd_fma(a, b, c);
let sub = simd_fma(a, b, simd_neg(c));
@@ -7200,7 +7385,8 @@ pub fn _mm_fmaddsub_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fmaddsub_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fmaddsub_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmaddsub_ph(a, b, c), a) }
}
@@ -7213,7 +7399,8 @@ pub fn _mm_mask_fmaddsub_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fmaddsub_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fmaddsub_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmaddsub_ph(a, b, c), c) }
}
@@ -7226,7 +7413,8 @@ pub fn _mm_mask3_fmaddsub_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fmaddsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fmaddsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmaddsub_ph(a, b, c), _mm_setzero_ph()) }
}
@@ -7238,7 +7426,8 @@ pub fn _mm_maskz_fmaddsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_fmaddsub_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_fmaddsub_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe {
let add = simd_fma(a, b, c);
let sub = simd_fma(a, b, simd_neg(c));
@@ -7259,7 +7448,8 @@ pub fn _mm256_fmaddsub_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_fmaddsub_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_fmaddsub_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmaddsub_ph(a, b, c), a) }
}
@@ -7272,7 +7462,8 @@ pub fn _mm256_mask_fmaddsub_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h)
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask3_fmaddsub_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask3_fmaddsub_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmaddsub_ph(a, b, c), c) }
}
@@ -7285,7 +7476,8 @@ pub fn _mm256_mask3_fmaddsub_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_fmaddsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_fmaddsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmaddsub_ph(a, b, c), _mm256_setzero_ph()) }
}
@@ -7297,7 +7489,8 @@ pub fn _mm256_maskz_fmaddsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_fmaddsub_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_fmaddsub_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe {
let add = simd_fma(a, b, c);
let sub = simd_fma(a, b, simd_neg(c));
@@ -7321,7 +7514,8 @@ pub fn _mm512_fmaddsub_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_fmaddsub_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_fmaddsub_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmaddsub_ph(a, b, c), a) }
}
@@ -7334,7 +7528,8 @@ pub fn _mm512_mask_fmaddsub_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h)
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask3_fmaddsub_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask3_fmaddsub_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmaddsub_ph(a, b, c), c) }
}
@@ -7347,7 +7542,8 @@ pub fn _mm512_mask3_fmaddsub_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmaddsub))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_fmaddsub_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_fmaddsub_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmaddsub_ph(a, b, c), _mm512_setzero_ph()) }
}
@@ -7481,7 +7677,8 @@ pub fn _mm512_maskz_fmaddsub_round_ph<const ROUNDING: i32>(
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_fmsubadd_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_fmsubadd_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
_mm_fmaddsub_ph(a, b, unsafe { simd_neg(c) })
}
@@ -7494,7 +7691,8 @@ pub fn _mm_fmsubadd_ph(a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_fmsubadd_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_fmsubadd_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmsubadd_ph(a, b, c), a) }
}
@@ -7507,7 +7705,8 @@ pub fn _mm_mask_fmsubadd_ph(a: __m128h, k: __mmask8, b: __m128h, c: __m128h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask3_fmsubadd_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask3_fmsubadd_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmsubadd_ph(a, b, c), c) }
}
@@ -7520,7 +7719,8 @@ pub fn _mm_mask3_fmsubadd_ph(a: __m128h, b: __m128h, c: __m128h, k: __mmask8) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_maskz_fmsubadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_maskz_fmsubadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, _mm_fmsubadd_ph(a, b, c), _mm_setzero_ph()) }
}
@@ -7532,7 +7732,8 @@ pub fn _mm_maskz_fmsubadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) ->
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_fmsubadd_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_fmsubadd_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
_mm256_fmaddsub_ph(a, b, unsafe { simd_neg(c) })
}
@@ -7545,7 +7746,8 @@ pub fn _mm256_fmsubadd_ph(a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_fmsubadd_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_fmsubadd_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmsubadd_ph(a, b, c), a) }
}
@@ -7558,7 +7760,8 @@ pub fn _mm256_mask_fmsubadd_ph(a: __m256h, k: __mmask16, b: __m256h, c: __m256h)
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask3_fmsubadd_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask3_fmsubadd_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmsubadd_ph(a, b, c), c) }
}
@@ -7571,7 +7774,8 @@ pub fn _mm256_mask3_fmsubadd_ph(a: __m256h, b: __m256h, c: __m256h, k: __mmask16
#[target_feature(enable = "avx512fp16,avx512vl")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_maskz_fmsubadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_maskz_fmsubadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, _mm256_fmsubadd_ph(a, b, c), _mm256_setzero_ph()) }
}
@@ -7583,7 +7787,8 @@ pub fn _mm256_maskz_fmsubadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_fmsubadd_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_fmsubadd_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
_mm512_fmaddsub_ph(a, b, unsafe { simd_neg(c) })
}
@@ -7596,7 +7801,8 @@ pub fn _mm512_fmsubadd_ph(a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_fmsubadd_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_fmsubadd_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmsubadd_ph(a, b, c), a) }
}
@@ -7609,7 +7815,8 @@ pub fn _mm512_mask_fmsubadd_ph(a: __m512h, k: __mmask32, b: __m512h, c: __m512h)
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask3_fmsubadd_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask3_fmsubadd_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmsubadd_ph(a, b, c), c) }
}
@@ -7622,7 +7829,8 @@ pub fn _mm512_mask3_fmsubadd_ph(a: __m512h, b: __m512h, c: __m512h, k: __mmask32
#[target_feature(enable = "avx512fp16")]
#[cfg_attr(test, assert_instr(vfmsubadd))]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_maskz_fmsubadd_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_maskz_fmsubadd_ph(k: __mmask32, a: __m512h, b: __m512h, c: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, _mm512_fmsubadd_ph(a, b, c), _mm512_setzero_ph()) }
}
@@ -11117,7 +11325,8 @@ pub fn _mm_maskz_reduce_round_sh<const IMM8: i32, const SAE: i32>(
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_reduce_add_ph(a: __m128h) -> f16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_reduce_add_ph(a: __m128h) -> f16 {
unsafe {
let b = simd_shuffle!(a, a, [4, 5, 6, 7, 0, 1, 2, 3]);
let a = _mm_add_ph(a, b);
@@ -11134,7 +11343,8 @@ pub fn _mm_reduce_add_ph(a: __m128h) -> f16 {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_reduce_add_ph(a: __m256h) -> f16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_reduce_add_ph(a: __m256h) -> f16 {
unsafe {
let p = simd_shuffle!(a, a, [0, 1, 2, 3, 4, 5, 6, 7]);
let q = simd_shuffle!(a, a, [8, 9, 10, 11, 12, 13, 14, 15]);
@@ -11149,7 +11359,8 @@ pub fn _mm256_reduce_add_ph(a: __m256h) -> f16 {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_reduce_add_ph(a: __m512h) -> f16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_reduce_add_ph(a: __m512h) -> f16 {
unsafe {
let p = simd_shuffle!(a, a, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]);
let q = simd_shuffle!(
@@ -11170,7 +11381,8 @@ pub fn _mm512_reduce_add_ph(a: __m512h) -> f16 {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_reduce_mul_ph(a: __m128h) -> f16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_reduce_mul_ph(a: __m128h) -> f16 {
unsafe {
let b = simd_shuffle!(a, a, [4, 5, 6, 7, 0, 1, 2, 3]);
let a = _mm_mul_ph(a, b);
@@ -11187,7 +11399,8 @@ pub fn _mm_reduce_mul_ph(a: __m128h) -> f16 {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_reduce_mul_ph(a: __m256h) -> f16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_reduce_mul_ph(a: __m256h) -> f16 {
unsafe {
let p = simd_shuffle!(a, a, [0, 1, 2, 3, 4, 5, 6, 7]);
let q = simd_shuffle!(a, a, [8, 9, 10, 11, 12, 13, 14, 15]);
@@ -11202,7 +11415,8 @@ pub fn _mm256_reduce_mul_ph(a: __m256h) -> f16 {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_reduce_mul_ph(a: __m512h) -> f16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_reduce_mul_ph(a: __m512h) -> f16 {
unsafe {
let p = simd_shuffle!(a, a, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]);
let q = simd_shuffle!(
@@ -11566,7 +11780,8 @@ pub fn _mm_mask_fpclass_sh_mask<const IMM8: i32>(k1: __mmask8, a: __m128h) -> __
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_mask_blend_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_mask_blend_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
unsafe { simd_select_bitmask(k, b, a) }
}
@@ -11577,7 +11792,8 @@ pub fn _mm_mask_blend_ph(k: __mmask8, a: __m128h, b: __m128h) -> __m128h {
#[inline]
#[target_feature(enable = "avx512fp16,avx512vl")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_mask_blend_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_mask_blend_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
unsafe { simd_select_bitmask(k, b, a) }
}
@@ -11588,7 +11804,8 @@ pub fn _mm256_mask_blend_ph(k: __mmask16, a: __m256h, b: __m256h) -> __m256h {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_mask_blend_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_mask_blend_ph(k: __mmask32, a: __m512h, b: __m512h) -> __m512h {
unsafe { simd_select_bitmask(k, b, a) }
}
@@ -16304,7 +16521,8 @@ pub fn _mm_maskz_cvt_roundsh_sd<const SAE: i32>(k: __mmask8, a: __m128d, b: __m1
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_cvtsh_h(a: __m128h) -> f16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_cvtsh_h(a: __m128h) -> f16 {
unsafe { simd_extract!(a, 0) }
}
@@ -16314,7 +16532,8 @@ pub fn _mm_cvtsh_h(a: __m128h) -> f16 {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm256_cvtsh_h(a: __m256h) -> f16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_cvtsh_h(a: __m256h) -> f16 {
unsafe { simd_extract!(a, 0) }
}
@@ -16324,7 +16543,8 @@ pub fn _mm256_cvtsh_h(a: __m256h) -> f16 {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm512_cvtsh_h(a: __m512h) -> f16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm512_cvtsh_h(a: __m512h) -> f16 {
unsafe { simd_extract!(a, 0) }
}
@@ -16334,7 +16554,8 @@ pub fn _mm512_cvtsh_h(a: __m512h) -> f16 {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_cvtsi128_si16(a: __m128i) -> i16 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_cvtsi128_si16(a: __m128i) -> i16 {
unsafe { simd_extract!(a.as_i16x8(), 0) }
}
@@ -16344,7 +16565,8 @@ pub fn _mm_cvtsi128_si16(a: __m128i) -> i16 {
#[inline]
#[target_feature(enable = "avx512fp16")]
#[unstable(feature = "stdarch_x86_avx512_f16", issue = "127213")]
pub fn _mm_cvtsi16_si128(a: i16) -> __m128i {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_cvtsi16_si128(a: i16) -> __m128i {
unsafe { transmute(simd_insert!(i16x8::ZERO, 0, a)) }
}
@@ -16708,6 +16930,7 @@ fn vreducesh(a: __m128h, b: __m128h, src: __m128h, k: __mmask8, imm8: i32, sae:
#[cfg(test)]
mod tests {
use crate::core_arch::assert_eq_const as assert_eq;
use crate::core_arch::x86::*;
use crate::mem::transmute;
use crate::ptr::{addr_of, addr_of_mut};
@@ -16737,14 +16960,14 @@ const fn _mm512_set1_pch(re: f16, im: f16) -> __m512h {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_set_ph() {
const unsafe fn test_mm_set_ph() {
let r = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let e = _mm_setr_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
assert_eq_m128h(r, e);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_set_ph() {
const unsafe fn test_mm256_set_ph() {
let r = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -16755,7 +16978,7 @@ unsafe fn test_mm256_set_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_set_ph() {
const unsafe fn test_mm512_set_ph() {
let r = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -16770,21 +16993,21 @@ unsafe fn test_mm512_set_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_set_sh() {
const unsafe fn test_mm_set_sh() {
let r = _mm_set_sh(1.0);
let e = _mm_set_ph(0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0);
assert_eq_m128h(r, e);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_set1_ph() {
const unsafe fn test_mm_set1_ph() {
let r = _mm_set1_ph(1.0);
let e = _mm_set_ph(1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0);
assert_eq_m128h(r, e);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_set1_ph() {
const unsafe fn test_mm256_set1_ph() {
let r = _mm256_set1_ph(1.0);
let e = _mm256_set_ph(
1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,
@@ -16793,7 +17016,7 @@ unsafe fn test_mm256_set1_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_set1_ph() {
const unsafe fn test_mm512_set1_ph() {
let r = _mm512_set1_ph(1.0);
let e = _mm512_set_ph(
1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,
@@ -16803,14 +17026,14 @@ unsafe fn test_mm512_set1_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_setr_ph() {
const unsafe fn test_mm_setr_ph() {
let r = _mm_setr_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let e = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
assert_eq_m128h(r, e);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_setr_ph() {
const unsafe fn test_mm256_setr_ph() {
let r = _mm256_setr_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -16821,7 +17044,7 @@ unsafe fn test_mm256_setr_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_setr_ph() {
const unsafe fn test_mm512_setr_ph() {
let r = _mm512_setr_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -16836,28 +17059,28 @@ unsafe fn test_mm512_setr_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_setzero_ph() {
const unsafe fn test_mm_setzero_ph() {
let r = _mm_setzero_ph();
let e = _mm_set1_ph(0.0);
assert_eq_m128h(r, e);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_setzero_ph() {
const unsafe fn test_mm256_setzero_ph() {
let r = _mm256_setzero_ph();
let e = _mm256_set1_ph(0.0);
assert_eq_m256h(r, e);
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_setzero_ph() {
const unsafe fn test_mm512_setzero_ph() {
let r = _mm512_setzero_ph();
let e = _mm512_set1_ph(0.0);
assert_eq_m512h(r, e);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_castsi128_ph() {
const unsafe fn test_mm_castsi128_ph() {
let a = _mm_set1_epi16(0x3c00);
let r = _mm_castsi128_ph(a);
let e = _mm_set1_ph(1.0);
@@ -16865,7 +17088,7 @@ unsafe fn test_mm_castsi128_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_castsi256_ph() {
const unsafe fn test_mm256_castsi256_ph() {
let a = _mm256_set1_epi16(0x3c00);
let r = _mm256_castsi256_ph(a);
let e = _mm256_set1_ph(1.0);
@@ -16873,7 +17096,7 @@ unsafe fn test_mm256_castsi256_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_castsi512_ph() {
const unsafe fn test_mm512_castsi512_ph() {
let a = _mm512_set1_epi16(0x3c00);
let r = _mm512_castsi512_ph(a);
let e = _mm512_set1_ph(1.0);
@@ -16881,7 +17104,7 @@ unsafe fn test_mm512_castsi512_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm_castph_si128() {
const unsafe fn test_mm_castph_si128() {
let a = _mm_set1_ph(1.0);
let r = _mm_castph_si128(a);
let e = _mm_set1_epi16(0x3c00);
@@ -16889,7 +17112,7 @@ unsafe fn test_mm_castph_si128() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm256_castph_si256() {
const unsafe fn test_mm256_castph_si256() {
let a = _mm256_set1_ph(1.0);
let r = _mm256_castph_si256(a);
let e = _mm256_set1_epi16(0x3c00);
@@ -16897,7 +17120,7 @@ unsafe fn test_mm256_castph_si256() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_castph_si512() {
const unsafe fn test_mm512_castph_si512() {
let a = _mm512_set1_ph(1.0);
let r = _mm512_castph_si512(a);
let e = _mm512_set1_epi16(0x3c00);
@@ -16905,7 +17128,7 @@ unsafe fn test_mm512_castph_si512() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_castps_ph() {
const unsafe fn test_mm_castps_ph() {
let a = _mm_castsi128_ps(_mm_set1_epi16(0x3c00));
let r = _mm_castps_ph(a);
let e = _mm_set1_ph(1.0);
@@ -16913,7 +17136,7 @@ unsafe fn test_mm_castps_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_castps_ph() {
const unsafe fn test_mm256_castps_ph() {
let a = _mm256_castsi256_ps(_mm256_set1_epi16(0x3c00));
let r = _mm256_castps_ph(a);
let e = _mm256_set1_ph(1.0);
@@ -16921,7 +17144,7 @@ unsafe fn test_mm256_castps_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_castps_ph() {
const unsafe fn test_mm512_castps_ph() {
let a = _mm512_castsi512_ps(_mm512_set1_epi16(0x3c00));
let r = _mm512_castps_ph(a);
let e = _mm512_set1_ph(1.0);
@@ -16929,7 +17152,7 @@ unsafe fn test_mm512_castps_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm_castph_ps() {
const unsafe fn test_mm_castph_ps() {
let a = _mm_castsi128_ph(_mm_set1_epi32(0x3f800000));
let r = _mm_castph_ps(a);
let e = _mm_set1_ps(1.0);
@@ -16937,7 +17160,7 @@ unsafe fn test_mm_castph_ps() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm256_castph_ps() {
const unsafe fn test_mm256_castph_ps() {
let a = _mm256_castsi256_ph(_mm256_set1_epi32(0x3f800000));
let r = _mm256_castph_ps(a);
let e = _mm256_set1_ps(1.0);
@@ -16945,7 +17168,7 @@ unsafe fn test_mm256_castph_ps() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_castph_ps() {
const unsafe fn test_mm512_castph_ps() {
let a = _mm512_castsi512_ph(_mm512_set1_epi32(0x3f800000));
let r = _mm512_castph_ps(a);
let e = _mm512_set1_ps(1.0);
@@ -16953,7 +17176,7 @@ unsafe fn test_mm512_castph_ps() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_castpd_ph() {
const unsafe fn test_mm_castpd_ph() {
let a = _mm_castsi128_pd(_mm_set1_epi16(0x3c00));
let r = _mm_castpd_ph(a);
let e = _mm_set1_ph(1.0);
@@ -16961,7 +17184,7 @@ unsafe fn test_mm_castpd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_castpd_ph() {
const unsafe fn test_mm256_castpd_ph() {
let a = _mm256_castsi256_pd(_mm256_set1_epi16(0x3c00));
let r = _mm256_castpd_ph(a);
let e = _mm256_set1_ph(1.0);
@@ -16969,7 +17192,7 @@ unsafe fn test_mm256_castpd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_castpd_ph() {
const unsafe fn test_mm512_castpd_ph() {
let a = _mm512_castsi512_pd(_mm512_set1_epi16(0x3c00));
let r = _mm512_castpd_ph(a);
let e = _mm512_set1_ph(1.0);
@@ -16977,7 +17200,7 @@ unsafe fn test_mm512_castpd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm_castph_pd() {
const unsafe fn test_mm_castph_pd() {
let a = _mm_castsi128_ph(_mm_set1_epi64x(0x3ff0000000000000));
let r = _mm_castph_pd(a);
let e = _mm_set1_pd(1.0);
@@ -16985,7 +17208,7 @@ unsafe fn test_mm_castph_pd() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm256_castph_pd() {
const unsafe fn test_mm256_castph_pd() {
let a = _mm256_castsi256_ph(_mm256_set1_epi64x(0x3ff0000000000000));
let r = _mm256_castph_pd(a);
let e = _mm256_set1_pd(1.0);
@@ -16993,7 +17216,7 @@ unsafe fn test_mm256_castph_pd() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_castph_pd() {
const unsafe fn test_mm512_castph_pd() {
let a = _mm512_castsi512_ph(_mm512_set1_epi64(0x3ff0000000000000));
let r = _mm512_castph_pd(a);
let e = _mm512_set1_pd(1.0);
@@ -17001,7 +17224,7 @@ unsafe fn test_mm512_castph_pd() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_castph256_ph128() {
const unsafe fn test_mm256_castph256_ph128() {
let a = _mm256_setr_ph(
1., 2., 3., 4., 5., 6., 7., 8., 9., 10., 11., 12., 13., 14., 15., 16.,
);
@@ -17011,7 +17234,7 @@ unsafe fn test_mm256_castph256_ph128() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm512_castph512_ph128() {
const unsafe fn test_mm512_castph512_ph128() {
let a = _mm512_setr_ph(
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.,
@@ -17022,7 +17245,7 @@ unsafe fn test_mm512_castph512_ph128() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm512_castph512_ph256() {
const unsafe fn test_mm512_castph512_ph256() {
let a = _mm512_setr_ph(
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.,
@@ -17035,21 +17258,21 @@ unsafe fn test_mm512_castph512_ph256() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_castph128_ph256() {
const unsafe fn test_mm256_castph128_ph256() {
let a = _mm_setr_ph(1., 2., 3., 4., 5., 6., 7., 8.);
let r = _mm256_castph128_ph256(a);
assert_eq_m128h(_mm256_castph256_ph128(r), a);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm512_castph128_ph512() {
const unsafe fn test_mm512_castph128_ph512() {
let a = _mm_setr_ph(1., 2., 3., 4., 5., 6., 7., 8.);
let r = _mm512_castph128_ph512(a);
assert_eq_m128h(_mm512_castph512_ph128(r), a);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm512_castph256_ph512() {
const unsafe fn test_mm512_castph256_ph512() {
let a = _mm256_setr_ph(
1., 2., 3., 4., 5., 6., 7., 8., 9., 10., 11., 12., 13., 14., 15., 16.,
);
@@ -17058,7 +17281,7 @@ unsafe fn test_mm512_castph256_ph512() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_zextph128_ph256() {
const unsafe fn test_mm256_zextph128_ph256() {
let a = _mm_setr_ph(1., 2., 3., 4., 5., 6., 7., 8.);
let r = _mm256_zextph128_ph256(a);
let e = _mm256_setr_ph(
@@ -17068,7 +17291,7 @@ unsafe fn test_mm256_zextph128_ph256() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_zextph128_ph512() {
const unsafe fn test_mm512_zextph128_ph512() {
let a = _mm_setr_ph(1., 2., 3., 4., 5., 6., 7., 8.);
let r = _mm512_zextph128_ph512(a);
let e = _mm512_setr_ph(
@@ -17079,7 +17302,7 @@ unsafe fn test_mm512_zextph128_ph512() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_zextph256_ph512() {
const unsafe fn test_mm512_zextph256_ph512() {
let a = _mm256_setr_ph(
1., 2., 3., 4., 5., 6., 7., 8., 9., 10., 11., 12., 13., 14., 15., 16.,
);
@@ -17346,14 +17569,14 @@ unsafe fn test_mm_ucomineq_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_load_ph() {
const unsafe fn test_mm_load_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_load_ph(addr_of!(a).cast());
assert_eq_m128h(a, b);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_load_ph() {
const unsafe fn test_mm256_load_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -17362,7 +17585,7 @@ unsafe fn test_mm256_load_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_load_ph() {
const unsafe fn test_mm512_load_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -17373,7 +17596,7 @@ unsafe fn test_mm512_load_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_load_sh() {
const unsafe fn test_mm_load_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_load_sh(addr_of!(a).cast());
assert_eq_m128h(a, b);
@@ -17399,7 +17622,7 @@ unsafe fn test_mm_maskz_load_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_loadu_ph() {
const unsafe fn test_mm_loadu_ph() {
let array = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0];
let r = _mm_loadu_ph(array.as_ptr());
let e = _mm_setr_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
@@ -17407,7 +17630,7 @@ unsafe fn test_mm_loadu_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_loadu_ph() {
const unsafe fn test_mm256_loadu_ph() {
let array = [
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
];
@@ -17419,7 +17642,7 @@ unsafe fn test_mm256_loadu_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_loadu_ph() {
const unsafe fn test_mm512_loadu_ph() {
let array = [
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -17435,7 +17658,7 @@ unsafe fn test_mm512_loadu_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_move_sh() {
const unsafe fn test_mm_move_sh() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_sh(9.0);
let r = _mm_move_sh(a, b);
@@ -17444,7 +17667,7 @@ unsafe fn test_mm_move_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_move_sh() {
const unsafe fn test_mm_mask_move_sh() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_sh(9.0);
let src = _mm_set_sh(10.0);
@@ -17454,7 +17677,7 @@ unsafe fn test_mm_mask_move_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_move_sh() {
const unsafe fn test_mm_maskz_move_sh() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_sh(9.0);
let r = _mm_maskz_move_sh(0, a, b);
@@ -17463,7 +17686,7 @@ unsafe fn test_mm_maskz_move_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_store_ph() {
const unsafe fn test_mm_store_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let mut b = _mm_setzero_ph();
_mm_store_ph(addr_of_mut!(b).cast(), a);
@@ -17471,7 +17694,7 @@ unsafe fn test_mm_store_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_store_ph() {
const unsafe fn test_mm256_store_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -17481,7 +17704,7 @@ unsafe fn test_mm256_store_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_store_ph() {
const unsafe fn test_mm512_store_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -17493,7 +17716,7 @@ unsafe fn test_mm512_store_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_store_sh() {
const unsafe fn test_mm_store_sh() {
let a = _mm_set_sh(1.0);
let mut b = _mm_setzero_ph();
_mm_store_sh(addr_of_mut!(b).cast(), a);
@@ -17511,7 +17734,7 @@ unsafe fn test_mm_mask_store_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_storeu_ph() {
const unsafe fn test_mm_storeu_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let mut array = [0.0; 8];
_mm_storeu_ph(array.as_mut_ptr(), a);
@@ -17519,7 +17742,7 @@ unsafe fn test_mm_storeu_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_storeu_ph() {
const unsafe fn test_mm256_storeu_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -17529,7 +17752,7 @@ unsafe fn test_mm256_storeu_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_storeu_ph() {
const unsafe fn test_mm512_storeu_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -17541,7 +17764,7 @@ unsafe fn test_mm512_storeu_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_add_ph() {
const unsafe fn test_mm_add_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
let r = _mm_add_ph(a, b);
@@ -17550,7 +17773,7 @@ unsafe fn test_mm_add_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_add_ph() {
const unsafe fn test_mm_mask_add_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
let src = _mm_set_ph(10., 11., 12., 13., 14., 15., 16., 17.);
@@ -17560,7 +17783,7 @@ unsafe fn test_mm_mask_add_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_add_ph() {
const unsafe fn test_mm_maskz_add_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
let r = _mm_maskz_add_ph(0b01010101, a, b);
@@ -17569,7 +17792,7 @@ unsafe fn test_mm_maskz_add_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_add_ph() {
const unsafe fn test_mm256_add_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -17582,7 +17805,7 @@ unsafe fn test_mm256_add_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_add_ph() {
const unsafe fn test_mm256_mask_add_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -17600,7 +17823,7 @@ unsafe fn test_mm256_mask_add_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_add_ph() {
const unsafe fn test_mm256_maskz_add_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -17615,7 +17838,7 @@ unsafe fn test_mm256_maskz_add_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_add_ph() {
const unsafe fn test_mm512_add_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -17632,7 +17855,7 @@ unsafe fn test_mm512_add_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_add_ph() {
const unsafe fn test_mm512_mask_add_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -17656,7 +17879,7 @@ unsafe fn test_mm512_mask_add_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_add_ph() {
const unsafe fn test_mm512_maskz_add_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -17786,7 +18009,7 @@ unsafe fn test_mm_maskz_add_round_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_add_sh() {
const unsafe fn test_mm_add_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let r = _mm_add_sh(a, b);
@@ -17795,7 +18018,7 @@ unsafe fn test_mm_add_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_add_sh() {
const unsafe fn test_mm_mask_add_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let src = _mm_set_sh(4.0);
@@ -17808,7 +18031,7 @@ unsafe fn test_mm_mask_add_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_add_sh() {
const unsafe fn test_mm_maskz_add_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let r = _mm_maskz_add_sh(0, a, b);
@@ -17820,7 +18043,7 @@ unsafe fn test_mm_maskz_add_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_sub_ph() {
const unsafe fn test_mm_sub_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
let r = _mm_sub_ph(a, b);
@@ -17829,7 +18052,7 @@ unsafe fn test_mm_sub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_sub_ph() {
const unsafe fn test_mm_mask_sub_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
let src = _mm_set_ph(10., 11., 12., 13., 14., 15., 16., 17.);
@@ -17839,7 +18062,7 @@ unsafe fn test_mm_mask_sub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_sub_ph() {
const unsafe fn test_mm_maskz_sub_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
let r = _mm_maskz_sub_ph(0b01010101, a, b);
@@ -17848,7 +18071,7 @@ unsafe fn test_mm_maskz_sub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_sub_ph() {
const unsafe fn test_mm256_sub_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -17864,7 +18087,7 @@ unsafe fn test_mm256_sub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_sub_ph() {
const unsafe fn test_mm256_mask_sub_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -17882,7 +18105,7 @@ unsafe fn test_mm256_mask_sub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_sub_ph() {
const unsafe fn test_mm256_maskz_sub_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -17897,7 +18120,7 @@ unsafe fn test_mm256_maskz_sub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_sub_ph() {
const unsafe fn test_mm512_sub_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -17918,7 +18141,7 @@ unsafe fn test_mm512_sub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_sub_ph() {
const unsafe fn test_mm512_mask_sub_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -17942,7 +18165,7 @@ unsafe fn test_mm512_mask_sub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_sub_ph() {
const unsafe fn test_mm512_maskz_sub_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -18076,7 +18299,7 @@ unsafe fn test_mm_maskz_sub_round_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_sub_sh() {
const unsafe fn test_mm_sub_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let r = _mm_sub_sh(a, b);
@@ -18085,7 +18308,7 @@ unsafe fn test_mm_sub_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_sub_sh() {
const unsafe fn test_mm_mask_sub_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let src = _mm_set_sh(4.0);
@@ -18098,7 +18321,7 @@ unsafe fn test_mm_mask_sub_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_sub_sh() {
const unsafe fn test_mm_maskz_sub_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let r = _mm_maskz_sub_sh(0, a, b);
@@ -18110,7 +18333,7 @@ unsafe fn test_mm_maskz_sub_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mul_ph() {
const unsafe fn test_mm_mul_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
let r = _mm_mul_ph(a, b);
@@ -18119,7 +18342,7 @@ unsafe fn test_mm_mul_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_mul_ph() {
const unsafe fn test_mm_mask_mul_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
let src = _mm_set_ph(10., 11., 12., 13., 14., 15., 16., 17.);
@@ -18129,7 +18352,7 @@ unsafe fn test_mm_mask_mul_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_mul_ph() {
const unsafe fn test_mm_maskz_mul_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(8.0, 7.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0);
let r = _mm_maskz_mul_ph(0b01010101, a, b);
@@ -18138,7 +18361,7 @@ unsafe fn test_mm_maskz_mul_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mul_ph() {
const unsafe fn test_mm256_mul_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -18154,7 +18377,7 @@ unsafe fn test_mm256_mul_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_mul_ph() {
const unsafe fn test_mm256_mask_mul_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -18172,7 +18395,7 @@ unsafe fn test_mm256_mask_mul_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_mul_ph() {
const unsafe fn test_mm256_maskz_mul_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -18187,7 +18410,7 @@ unsafe fn test_mm256_maskz_mul_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mul_ph() {
const unsafe fn test_mm512_mul_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -18208,7 +18431,7 @@ unsafe fn test_mm512_mul_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_mul_ph() {
const unsafe fn test_mm512_mask_mul_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -18232,7 +18455,7 @@ unsafe fn test_mm512_mask_mul_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_mul_ph() {
const unsafe fn test_mm512_maskz_mul_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -18366,7 +18589,7 @@ unsafe fn test_mm_maskz_mul_round_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mul_sh() {
const unsafe fn test_mm_mul_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let r = _mm_mul_sh(a, b);
@@ -18375,7 +18598,7 @@ unsafe fn test_mm_mul_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_mul_sh() {
const unsafe fn test_mm_mask_mul_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let src = _mm_set_sh(4.0);
@@ -18388,7 +18611,7 @@ unsafe fn test_mm_mask_mul_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_mul_sh() {
const unsafe fn test_mm_maskz_mul_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let r = _mm_maskz_mul_sh(0, a, b);
@@ -18400,7 +18623,7 @@ unsafe fn test_mm_maskz_mul_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_div_ph() {
const unsafe fn test_mm_div_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let r = _mm_div_ph(a, b);
@@ -18409,7 +18632,7 @@ unsafe fn test_mm_div_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_div_ph() {
const unsafe fn test_mm_mask_div_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let src = _mm_set_ph(4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0);
@@ -18419,7 +18642,7 @@ unsafe fn test_mm_mask_div_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_div_ph() {
const unsafe fn test_mm_maskz_div_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let r = _mm_maskz_div_ph(0b01010101, a, b);
@@ -18428,7 +18651,7 @@ unsafe fn test_mm_maskz_div_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_div_ph() {
const unsafe fn test_mm256_div_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let r = _mm256_div_ph(a, b);
@@ -18437,7 +18660,7 @@ unsafe fn test_mm256_div_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_div_ph() {
const unsafe fn test_mm256_mask_div_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let src = _mm256_set_ph(
@@ -18452,7 +18675,7 @@ unsafe fn test_mm256_mask_div_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_div_ph() {
const unsafe fn test_mm256_maskz_div_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let r = _mm256_maskz_div_ph(0b0101010101010101, a, b);
@@ -18463,7 +18686,7 @@ unsafe fn test_mm256_maskz_div_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_div_ph() {
const unsafe fn test_mm512_div_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let r = _mm512_div_ph(a, b);
@@ -18472,7 +18695,7 @@ unsafe fn test_mm512_div_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_div_ph() {
const unsafe fn test_mm512_mask_div_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let src = _mm512_set_ph(
@@ -18489,7 +18712,7 @@ unsafe fn test_mm512_mask_div_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_div_ph() {
const unsafe fn test_mm512_maskz_div_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let r = _mm512_maskz_div_ph(0b01010101010101010101010101010101, a, b);
@@ -18588,7 +18811,7 @@ unsafe fn test_mm_maskz_div_round_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_div_sh() {
const unsafe fn test_mm_div_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let r = _mm_div_sh(a, b);
@@ -18597,7 +18820,7 @@ unsafe fn test_mm_div_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_div_sh() {
const unsafe fn test_mm_mask_div_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let src = _mm_set_sh(4.0);
@@ -18610,7 +18833,7 @@ unsafe fn test_mm_mask_div_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_div_sh() {
const unsafe fn test_mm_maskz_div_sh() {
let a = _mm_set_sh(1.0);
let b = _mm_set_sh(2.0);
let r = _mm_maskz_div_sh(0, a, b);
@@ -19454,7 +19677,7 @@ unsafe fn test_mm_maskz_fcmul_round_sch() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_abs_ph() {
const unsafe fn test_mm_abs_ph() {
let a = _mm_set_ph(-1.0, 0.0, 1.0, -2.0, 3.0, -4.0, 5.0, -6.0);
let r = _mm_abs_ph(a);
let e = _mm_set_ph(1.0, 0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0);
@@ -19462,7 +19685,7 @@ unsafe fn test_mm_abs_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_abs_ph() {
const unsafe fn test_mm256_abs_ph() {
let a = _mm256_set_ph(
-1.0, 0.0, 1.0, -2.0, 3.0, -4.0, 5.0, -6.0, 7.0, -8.0, 9.0, -10.0, 11.0, -12.0, 13.0,
-14.0,
@@ -19475,7 +19698,7 @@ unsafe fn test_mm256_abs_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_abs_ph() {
const unsafe fn test_mm512_abs_ph() {
let a = _mm512_set_ph(
-1.0, 0.0, 1.0, -2.0, 3.0, -4.0, 5.0, -6.0, 7.0, -8.0, 9.0, -10.0, 11.0, -12.0, 13.0,
-14.0, 15.0, -16.0, 17.0, -18.0, 19.0, -20.0, 21.0, -22.0, 23.0, -24.0, 25.0, -26.0,
@@ -19491,7 +19714,7 @@ unsafe fn test_mm512_abs_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_conj_pch() {
const unsafe fn test_mm_conj_pch() {
let a = _mm_set1_pch(0.0, 1.0);
let r = _mm_conj_pch(a);
let e = _mm_set1_pch(0.0, -1.0);
@@ -19499,7 +19722,7 @@ unsafe fn test_mm_conj_pch() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_conj_pch() {
const unsafe fn test_mm_mask_conj_pch() {
let a = _mm_set1_pch(0.0, 1.0);
let src = _mm_setr_ph(2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0);
let r = _mm_mask_conj_pch(src, 0b0101, a);
@@ -19508,7 +19731,7 @@ unsafe fn test_mm_mask_conj_pch() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_conj_pch() {
const unsafe fn test_mm_maskz_conj_pch() {
let a = _mm_set1_pch(0.0, 1.0);
let r = _mm_maskz_conj_pch(0b0101, a);
let e = _mm_setr_ph(0.0, -1.0, 0.0, 0.0, 0.0, -1.0, 0.0, 0.0);
@@ -19516,7 +19739,7 @@ unsafe fn test_mm_maskz_conj_pch() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_conj_pch() {
const unsafe fn test_mm256_conj_pch() {
let a = _mm256_set1_pch(0.0, 1.0);
let r = _mm256_conj_pch(a);
let e = _mm256_set1_pch(0.0, -1.0);
@@ -19524,7 +19747,7 @@ unsafe fn test_mm256_conj_pch() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_conj_pch() {
const unsafe fn test_mm256_mask_conj_pch() {
let a = _mm256_set1_pch(0.0, 1.0);
let src = _mm256_setr_ph(
2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 17.0,
@@ -19537,7 +19760,7 @@ unsafe fn test_mm256_mask_conj_pch() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_conj_pch() {
const unsafe fn test_mm256_maskz_conj_pch() {
let a = _mm256_set1_pch(0.0, 1.0);
let r = _mm256_maskz_conj_pch(0b01010101, a);
let e = _mm256_setr_ph(
@@ -19547,7 +19770,7 @@ unsafe fn test_mm256_maskz_conj_pch() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_conj_pch() {
const unsafe fn test_mm512_conj_pch() {
let a = _mm512_set1_pch(0.0, 1.0);
let r = _mm512_conj_pch(a);
let e = _mm512_set1_pch(0.0, -1.0);
@@ -19555,7 +19778,7 @@ unsafe fn test_mm512_conj_pch() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_conj_pch() {
const unsafe fn test_mm512_mask_conj_pch() {
let a = _mm512_set1_pch(0.0, 1.0);
let src = _mm512_setr_ph(
2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 17.0,
@@ -19572,7 +19795,7 @@ unsafe fn test_mm512_mask_conj_pch() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_conj_pch() {
const unsafe fn test_mm512_maskz_conj_pch() {
let a = _mm512_set1_pch(0.0, 1.0);
let r = _mm512_maskz_conj_pch(0b0101010101010101, a);
let e = _mm512_setr_ph(
@@ -20203,7 +20426,7 @@ unsafe fn test_mm_maskz_fcmadd_round_sch() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fmadd_ph() {
const unsafe fn test_mm_fmadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20213,7 +20436,7 @@ unsafe fn test_mm_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fmadd_ph() {
const unsafe fn test_mm_mask_fmadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20223,7 +20446,7 @@ unsafe fn test_mm_mask_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fmadd_ph() {
const unsafe fn test_mm_mask3_fmadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20233,7 +20456,7 @@ unsafe fn test_mm_mask3_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fmadd_ph() {
const unsafe fn test_mm_maskz_fmadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20243,7 +20466,7 @@ unsafe fn test_mm_maskz_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_fmadd_ph() {
const unsafe fn test_mm256_fmadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20253,7 +20476,7 @@ unsafe fn test_mm256_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_fmadd_ph() {
const unsafe fn test_mm256_mask_fmadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20265,7 +20488,7 @@ unsafe fn test_mm256_mask_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask3_fmadd_ph() {
const unsafe fn test_mm256_mask3_fmadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20277,7 +20500,7 @@ unsafe fn test_mm256_mask3_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_fmadd_ph() {
const unsafe fn test_mm256_maskz_fmadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20289,7 +20512,7 @@ unsafe fn test_mm256_maskz_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_fmadd_ph() {
const unsafe fn test_mm512_fmadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20299,7 +20522,7 @@ unsafe fn test_mm512_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_fmadd_ph() {
const unsafe fn test_mm512_mask_fmadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20312,7 +20535,7 @@ unsafe fn test_mm512_mask_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask3_fmadd_ph() {
const unsafe fn test_mm512_mask3_fmadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20325,7 +20548,7 @@ unsafe fn test_mm512_mask3_fmadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_fmadd_ph() {
const unsafe fn test_mm512_maskz_fmadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20402,7 +20625,7 @@ unsafe fn test_mm512_maskz_fmadd_round_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fmadd_sh() {
const unsafe fn test_mm_fmadd_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -20412,7 +20635,7 @@ unsafe fn test_mm_fmadd_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fmadd_sh() {
const unsafe fn test_mm_mask_fmadd_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -20425,7 +20648,7 @@ unsafe fn test_mm_mask_fmadd_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fmadd_sh() {
const unsafe fn test_mm_mask3_fmadd_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -20438,7 +20661,7 @@ unsafe fn test_mm_mask3_fmadd_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fmadd_sh() {
const unsafe fn test_mm_maskz_fmadd_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -20512,7 +20735,7 @@ unsafe fn test_mm_maskz_fmadd_round_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fmsub_ph() {
const unsafe fn test_mm_fmsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20522,7 +20745,7 @@ unsafe fn test_mm_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fmsub_ph() {
const unsafe fn test_mm_mask_fmsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20532,7 +20755,7 @@ unsafe fn test_mm_mask_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fmsub_ph() {
const unsafe fn test_mm_mask3_fmsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20542,7 +20765,7 @@ unsafe fn test_mm_mask3_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fmsub_ph() {
const unsafe fn test_mm_maskz_fmsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20552,7 +20775,7 @@ unsafe fn test_mm_maskz_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_fmsub_ph() {
const unsafe fn test_mm256_fmsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20562,7 +20785,7 @@ unsafe fn test_mm256_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_fmsub_ph() {
const unsafe fn test_mm256_mask_fmsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20574,7 +20797,7 @@ unsafe fn test_mm256_mask_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask3_fmsub_ph() {
const unsafe fn test_mm256_mask3_fmsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20586,7 +20809,7 @@ unsafe fn test_mm256_mask3_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_fmsub_ph() {
const unsafe fn test_mm256_maskz_fmsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20598,7 +20821,7 @@ unsafe fn test_mm256_maskz_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_fmsub_ph() {
const unsafe fn test_mm512_fmsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20608,7 +20831,7 @@ unsafe fn test_mm512_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_fmsub_ph() {
const unsafe fn test_mm512_mask_fmsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20621,7 +20844,7 @@ unsafe fn test_mm512_mask_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask3_fmsub_ph() {
const unsafe fn test_mm512_mask3_fmsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20634,7 +20857,7 @@ unsafe fn test_mm512_mask3_fmsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_fmsub_ph() {
const unsafe fn test_mm512_maskz_fmsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20711,7 +20934,7 @@ unsafe fn test_mm512_maskz_fmsub_round_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fmsub_sh() {
const unsafe fn test_mm_fmsub_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -20721,7 +20944,7 @@ unsafe fn test_mm_fmsub_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fmsub_sh() {
const unsafe fn test_mm_mask_fmsub_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -20734,7 +20957,7 @@ unsafe fn test_mm_mask_fmsub_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fmsub_sh() {
const unsafe fn test_mm_mask3_fmsub_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -20747,7 +20970,7 @@ unsafe fn test_mm_mask3_fmsub_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fmsub_sh() {
const unsafe fn test_mm_maskz_fmsub_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -20821,7 +21044,7 @@ unsafe fn test_mm_maskz_fmsub_round_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fnmadd_ph() {
const unsafe fn test_mm_fnmadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20831,7 +21054,7 @@ unsafe fn test_mm_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fnmadd_ph() {
const unsafe fn test_mm_mask_fnmadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20841,7 +21064,7 @@ unsafe fn test_mm_mask_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fnmadd_ph() {
const unsafe fn test_mm_mask3_fnmadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20851,7 +21074,7 @@ unsafe fn test_mm_mask3_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fnmadd_ph() {
const unsafe fn test_mm_maskz_fnmadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -20861,7 +21084,7 @@ unsafe fn test_mm_maskz_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_fnmadd_ph() {
const unsafe fn test_mm256_fnmadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20871,7 +21094,7 @@ unsafe fn test_mm256_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_fnmadd_ph() {
const unsafe fn test_mm256_mask_fnmadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20883,7 +21106,7 @@ unsafe fn test_mm256_mask_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask3_fnmadd_ph() {
const unsafe fn test_mm256_mask3_fnmadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20895,7 +21118,7 @@ unsafe fn test_mm256_mask3_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_fnmadd_ph() {
const unsafe fn test_mm256_maskz_fnmadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -20907,7 +21130,7 @@ unsafe fn test_mm256_maskz_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_fnmadd_ph() {
const unsafe fn test_mm512_fnmadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20917,7 +21140,7 @@ unsafe fn test_mm512_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_fnmadd_ph() {
const unsafe fn test_mm512_mask_fnmadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20930,7 +21153,7 @@ unsafe fn test_mm512_mask_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask3_fnmadd_ph() {
const unsafe fn test_mm512_mask3_fnmadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -20943,7 +21166,7 @@ unsafe fn test_mm512_mask3_fnmadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_fnmadd_ph() {
const unsafe fn test_mm512_maskz_fnmadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21021,7 +21244,7 @@ unsafe fn test_mm512_maskz_fnmadd_round_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fnmadd_sh() {
const unsafe fn test_mm_fnmadd_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -21031,7 +21254,7 @@ unsafe fn test_mm_fnmadd_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fnmadd_sh() {
const unsafe fn test_mm_mask_fnmadd_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -21044,7 +21267,7 @@ unsafe fn test_mm_mask_fnmadd_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fnmadd_sh() {
const unsafe fn test_mm_mask3_fnmadd_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -21057,7 +21280,7 @@ unsafe fn test_mm_mask3_fnmadd_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fnmadd_sh() {
const unsafe fn test_mm_maskz_fnmadd_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -21131,7 +21354,7 @@ unsafe fn test_mm_maskz_fnmadd_round_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fnmsub_ph() {
const unsafe fn test_mm_fnmsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21141,7 +21364,7 @@ unsafe fn test_mm_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fnmsub_ph() {
const unsafe fn test_mm_mask_fnmsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21151,7 +21374,7 @@ unsafe fn test_mm_mask_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fnmsub_ph() {
const unsafe fn test_mm_mask3_fnmsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21161,7 +21384,7 @@ unsafe fn test_mm_mask3_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fnmsub_ph() {
const unsafe fn test_mm_maskz_fnmsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21171,7 +21394,7 @@ unsafe fn test_mm_maskz_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_fnmsub_ph() {
const unsafe fn test_mm256_fnmsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21181,7 +21404,7 @@ unsafe fn test_mm256_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_fnmsub_ph() {
const unsafe fn test_mm256_mask_fnmsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21193,7 +21416,7 @@ unsafe fn test_mm256_mask_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask3_fnmsub_ph() {
const unsafe fn test_mm256_mask3_fnmsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21205,7 +21428,7 @@ unsafe fn test_mm256_mask3_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_fnmsub_ph() {
const unsafe fn test_mm256_maskz_fnmsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21217,7 +21440,7 @@ unsafe fn test_mm256_maskz_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_fnmsub_ph() {
const unsafe fn test_mm512_fnmsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21227,7 +21450,7 @@ unsafe fn test_mm512_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_fnmsub_ph() {
const unsafe fn test_mm512_mask_fnmsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21240,7 +21463,7 @@ unsafe fn test_mm512_mask_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask3_fnmsub_ph() {
const unsafe fn test_mm512_mask3_fnmsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21253,7 +21476,7 @@ unsafe fn test_mm512_mask3_fnmsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_fnmsub_ph() {
const unsafe fn test_mm512_maskz_fnmsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21331,7 +21554,7 @@ unsafe fn test_mm512_maskz_fnmsub_round_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fnmsub_sh() {
const unsafe fn test_mm_fnmsub_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -21341,7 +21564,7 @@ unsafe fn test_mm_fnmsub_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fnmsub_sh() {
const unsafe fn test_mm_mask_fnmsub_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -21354,7 +21577,7 @@ unsafe fn test_mm_mask_fnmsub_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fnmsub_sh() {
const unsafe fn test_mm_mask3_fnmsub_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -21367,7 +21590,7 @@ unsafe fn test_mm_mask3_fnmsub_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fnmsub_sh() {
const unsafe fn test_mm_maskz_fnmsub_sh() {
let a = _mm_setr_ph(1.0, 10., 11., 12., 13., 14., 15., 16.);
let b = _mm_setr_ph(2.0, 20., 21., 22., 23., 24., 25., 26.);
let c = _mm_setr_ph(3.0, 30., 31., 32., 33., 34., 35., 36.);
@@ -21441,7 +21664,7 @@ unsafe fn test_mm_maskz_fnmsub_round_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fmaddsub_ph() {
const unsafe fn test_mm_fmaddsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21451,7 +21674,7 @@ unsafe fn test_mm_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fmaddsub_ph() {
const unsafe fn test_mm_mask_fmaddsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21461,7 +21684,7 @@ unsafe fn test_mm_mask_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fmaddsub_ph() {
const unsafe fn test_mm_mask3_fmaddsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21471,7 +21694,7 @@ unsafe fn test_mm_mask3_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fmaddsub_ph() {
const unsafe fn test_mm_maskz_fmaddsub_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21481,7 +21704,7 @@ unsafe fn test_mm_maskz_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_fmaddsub_ph() {
const unsafe fn test_mm256_fmaddsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21493,7 +21716,7 @@ unsafe fn test_mm256_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_fmaddsub_ph() {
const unsafe fn test_mm256_mask_fmaddsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21505,7 +21728,7 @@ unsafe fn test_mm256_mask_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask3_fmaddsub_ph() {
const unsafe fn test_mm256_mask3_fmaddsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21517,7 +21740,7 @@ unsafe fn test_mm256_mask3_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_fmaddsub_ph() {
const unsafe fn test_mm256_maskz_fmaddsub_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21529,7 +21752,7 @@ unsafe fn test_mm256_maskz_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_fmaddsub_ph() {
const unsafe fn test_mm512_fmaddsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21542,7 +21765,7 @@ unsafe fn test_mm512_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_fmaddsub_ph() {
const unsafe fn test_mm512_mask_fmaddsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21555,7 +21778,7 @@ unsafe fn test_mm512_mask_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask3_fmaddsub_ph() {
const unsafe fn test_mm512_mask3_fmaddsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21568,7 +21791,7 @@ unsafe fn test_mm512_mask3_fmaddsub_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_fmaddsub_ph() {
const unsafe fn test_mm512_maskz_fmaddsub_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21649,7 +21872,7 @@ unsafe fn test_mm512_maskz_fmaddsub_round_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_fmsubadd_ph() {
const unsafe fn test_mm_fmsubadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21659,7 +21882,7 @@ unsafe fn test_mm_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_fmsubadd_ph() {
const unsafe fn test_mm_mask_fmsubadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21669,7 +21892,7 @@ unsafe fn test_mm_mask_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask3_fmsubadd_ph() {
const unsafe fn test_mm_mask3_fmsubadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21679,7 +21902,7 @@ unsafe fn test_mm_mask3_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_maskz_fmsubadd_ph() {
const unsafe fn test_mm_maskz_fmsubadd_ph() {
let a = _mm_set1_ph(1.0);
let b = _mm_set1_ph(2.0);
let c = _mm_set1_ph(3.0);
@@ -21689,7 +21912,7 @@ unsafe fn test_mm_maskz_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_fmsubadd_ph() {
const unsafe fn test_mm256_fmsubadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21701,7 +21924,7 @@ unsafe fn test_mm256_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_fmsubadd_ph() {
const unsafe fn test_mm256_mask_fmsubadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21713,7 +21936,7 @@ unsafe fn test_mm256_mask_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask3_fmsubadd_ph() {
const unsafe fn test_mm256_mask3_fmsubadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21725,7 +21948,7 @@ unsafe fn test_mm256_mask3_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_maskz_fmsubadd_ph() {
const unsafe fn test_mm256_maskz_fmsubadd_ph() {
let a = _mm256_set1_ph(1.0);
let b = _mm256_set1_ph(2.0);
let c = _mm256_set1_ph(3.0);
@@ -21737,7 +21960,7 @@ unsafe fn test_mm256_maskz_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_fmsubadd_ph() {
const unsafe fn test_mm512_fmsubadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21750,7 +21973,7 @@ unsafe fn test_mm512_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_fmsubadd_ph() {
const unsafe fn test_mm512_mask_fmsubadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21763,7 +21986,7 @@ unsafe fn test_mm512_mask_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask3_fmsubadd_ph() {
const unsafe fn test_mm512_mask3_fmsubadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -21776,7 +21999,7 @@ unsafe fn test_mm512_mask3_fmsubadd_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_maskz_fmsubadd_ph() {
const unsafe fn test_mm512_maskz_fmsubadd_ph() {
let a = _mm512_set1_ph(1.0);
let b = _mm512_set1_ph(2.0);
let c = _mm512_set1_ph(3.0);
@@ -23741,42 +23964,42 @@ unsafe fn test_mm_maskz_reduce_round_sh() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_reduce_add_ph() {
const unsafe fn test_mm_reduce_add_ph() {
let a = _mm_set1_ph(2.0);
let r = _mm_reduce_add_ph(a);
assert_eq!(r, 16.0);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_reduce_add_ph() {
const unsafe fn test_mm256_reduce_add_ph() {
let a = _mm256_set1_ph(2.0);
let r = _mm256_reduce_add_ph(a);
assert_eq!(r, 32.0);
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_reduce_add_ph() {
const unsafe fn test_mm512_reduce_add_ph() {
let a = _mm512_set1_ph(2.0);
let r = _mm512_reduce_add_ph(a);
assert_eq!(r, 64.0);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_reduce_mul_ph() {
const unsafe fn test_mm_reduce_mul_ph() {
let a = _mm_set1_ph(2.0);
let r = _mm_reduce_mul_ph(a);
assert_eq!(r, 256.0);
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_reduce_mul_ph() {
const unsafe fn test_mm256_reduce_mul_ph() {
let a = _mm256_set1_ph(2.0);
let r = _mm256_reduce_mul_ph(a);
assert_eq!(r, 65536.0);
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_reduce_mul_ph() {
const unsafe fn test_mm512_reduce_mul_ph() {
let a = _mm512_set1_ph(2.0);
let r = _mm512_reduce_mul_ph(a);
assert_eq!(r, 16777216.0);
@@ -24013,7 +24236,7 @@ unsafe fn test_mm_mask_fpclass_sh_mask() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm_mask_blend_ph() {
const unsafe fn test_mm_mask_blend_ph() {
let a = _mm_set_ph(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0);
let b = _mm_set_ph(-1.0, -2.0, -3.0, -4.0, -5.0, -6.0, -7.0, -8.0);
let r = _mm_mask_blend_ph(0b01010101, a, b);
@@ -24022,7 +24245,7 @@ unsafe fn test_mm_mask_blend_ph() {
}
#[simd_test(enable = "avx512fp16,avx512vl")]
unsafe fn test_mm256_mask_blend_ph() {
const unsafe fn test_mm256_mask_blend_ph() {
let a = _mm256_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -24039,7 +24262,7 @@ unsafe fn test_mm256_mask_blend_ph() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_mask_blend_ph() {
const unsafe fn test_mm512_mask_blend_ph() {
let a = _mm512_set_ph(
1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -27310,14 +27533,14 @@ unsafe fn test_mm_maskz_cvt_roundsh_sd() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm_cvtsh_h() {
const unsafe fn test_mm_cvtsh_h() {
let a = _mm_setr_ph(1.0, 2.0, 3.0, 42.0, 5.0, 6.0, 7.0, 8.0);
let r = _mm_cvtsh_h(a);
assert_eq!(r, 1.0);
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm256_cvtsh_h() {
const unsafe fn test_mm256_cvtsh_h() {
let a = _mm256_setr_ph(
1.0, 2.0, 3.0, 42.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
);
@@ -27326,7 +27549,7 @@ unsafe fn test_mm256_cvtsh_h() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm512_cvtsh_h() {
const unsafe fn test_mm512_cvtsh_h() {
let a = _mm512_setr_ph(
1.0, 2.0, 3.0, 42.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0,
@@ -27337,14 +27560,14 @@ unsafe fn test_mm512_cvtsh_h() {
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm_cvtsi128_si16() {
const unsafe fn test_mm_cvtsi128_si16() {
let a = _mm_setr_epi16(1, 2, 3, 4, 5, 6, 7, 8);
let r = _mm_cvtsi128_si16(a);
assert_eq!(r, 1);
}
#[simd_test(enable = "avx512fp16")]
unsafe fn test_mm_cvtsi16_si128() {
const unsafe fn test_mm_cvtsi16_si128() {
let a = 1;
let r = _mm_cvtsi16_si128(a);
let e = _mm_setr_epi16(1, 0, 0, 0, 0, 0, 0, 0);
library/stdarch/crates/core_arch/src/x86/f16c.rs
+7 -4
View File
@@ -25,7 +25,8 @@
#[target_feature(enable = "f16c")]
#[cfg_attr(test, assert_instr("vcvtph2ps"))]
#[stable(feature = "x86_f16c_intrinsics", since = "1.68.0")]
pub fn _mm_cvtph_ps(a: __m128i) -> __m128 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm_cvtph_ps(a: __m128i) -> __m128 {
unsafe {
let a: f16x8 = transmute(a);
let a: f16x4 = simd_shuffle!(a, a, [0, 1, 2, 3]);
@@ -41,7 +42,8 @@ pub fn _mm_cvtph_ps(a: __m128i) -> __m128 {
#[target_feature(enable = "f16c")]
#[cfg_attr(test, assert_instr("vcvtph2ps"))]
#[stable(feature = "x86_f16c_intrinsics", since = "1.68.0")]
pub fn _mm256_cvtph_ps(a: __m128i) -> __m256 {
#[rustc_const_unstable(feature = "stdarch_const_x86", issue = "149298")]
pub const fn _mm256_cvtph_ps(a: __m128i) -> __m256 {
unsafe {
let a: f16x8 = transmute(a);
simd_cast(a)
@@ -103,6 +105,7 @@ pub fn _mm256_cvtps_ph<const IMM_ROUNDING: i32>(a: __m256) -> __m128i {
#[cfg(test)]
mod tests {
use crate::core_arch::assert_eq_const as assert_eq;
use crate::{core_arch::x86::*, mem::transmute};
use stdarch_test::simd_test;
@@ -116,7 +119,7 @@ mod tests {
const F16_EIGHT: i16 = 0x4800;
#[simd_test(enable = "f16c")]
unsafe fn test_mm_cvtph_ps() {
const unsafe fn test_mm_cvtph_ps() {
let a = _mm_set_epi16(0, 0, 0, 0, F16_ONE, F16_TWO, F16_THREE, F16_FOUR);
let r = _mm_cvtph_ps(a);
let e = _mm_set_ps(1.0, 2.0, 3.0, 4.0);
@@ -124,7 +127,7 @@ unsafe fn test_mm_cvtph_ps() {
}
#[simd_test(enable = "f16c")]
unsafe fn test_mm256_cvtph_ps() {
const unsafe fn test_mm256_cvtph_ps() {
let a = _mm_set_epi16(
F16_ONE, F16_TWO, F16_THREE, F16_FOUR, F16_FIVE, F16_SIX, F16_SEVEN, F16_EIGHT,
);