Skip to content

Use SIMD intrinsics for vfmaddsubph and vfmsubaddph #1932

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Oct 4, 2025
Merged

Use SIMD intrinsics for vfmaddsubph and vfmsubaddph #1932

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 29 additions & 10 deletions crates/core_arch/src/x86/avx512fp16.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7184,7 +7184,11 @@ pub fn _mm_maskz_fnmsub_round_sh<const ROUNDING: i32>(
#[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 {
unsafe { vfmaddsubph_128(a, b, c) }
unsafe {
let add = simd_fma(a, b, c);
let sub = simd_fma(a, b, simd_neg(c));
simd_shuffle!(sub, add, [0, 9, 2, 11, 4, 13, 6, 15])
}
}

/// Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
Expand Down Expand Up @@ -7235,7 +7239,15 @@ pub fn _mm_maskz_fmaddsub_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) ->
#[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 {
unsafe { vfmaddsubph_256(a, b, c) }
unsafe {
let add = simd_fma(a, b, c);
let sub = simd_fma(a, b, simd_neg(c));
simd_shuffle!(
sub,
add,
[0, 17, 2, 19, 4, 21, 6, 23, 8, 25, 10, 27, 12, 29, 14, 31]
)
}
}

/// Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
Expand Down Expand Up @@ -7286,7 +7298,18 @@ pub fn _mm256_maskz_fmaddsub_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h
#[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 {
_mm512_fmaddsub_round_ph::<_MM_FROUND_CUR_DIRECTION>(a, b, c)
unsafe {
let add = simd_fma(a, b, c);
let sub = simd_fma(a, b, simd_neg(c));
simd_shuffle!(
sub,
add,
[
0, 33, 2, 35, 4, 37, 6, 39, 8, 41, 10, 43, 12, 45, 14, 47, 16, 49, 18, 51, 20, 53,
22, 55, 24, 57, 26, 59, 28, 61, 30, 63
]
)
}
}

/// Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively add and
Expand Down Expand Up @@ -7459,7 +7482,7 @@ pub fn _mm512_maskz_fmaddsub_round_ph<const ROUNDING: i32>(
#[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 {
unsafe { vfmaddsubph_128(a, b, simd_neg(c)) }
_mm_fmaddsub_ph(a, b, unsafe { simd_neg(c) })
}

/// Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
Expand Down Expand Up @@ -7510,7 +7533,7 @@ pub fn _mm_maskz_fmsubadd_ph(k: __mmask8, a: __m128h, b: __m128h, c: __m128h) ->
#[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 {
unsafe { vfmaddsubph_256(a, b, simd_neg(c)) }
_mm256_fmaddsub_ph(a, b, unsafe { simd_neg(c) })
}

/// Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
Expand Down Expand Up @@ -7561,7 +7584,7 @@ pub fn _mm256_maskz_fmsubadd_ph(k: __mmask16, a: __m256h, b: __m256h, c: __m256h
#[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 {
_mm512_fmsubadd_round_ph::<_MM_FROUND_CUR_DIRECTION>(a, b, c)
_mm512_fmaddsub_ph(a, b, unsafe { simd_neg(c) })
}

/// Multiply packed half-precision (16-bit) floating-point elements in a and b, alternatively subtract
Expand Down Expand Up @@ -16409,10 +16432,6 @@ unsafe extern "C" {
#[link_name = "llvm.x86.avx512fp16.vfmadd.f16"]
fn vfmaddsh(a: f16, b: f16, c: f16, rounding: i32) -> f16;

#[link_name = "llvm.x86.avx512fp16.vfmaddsub.ph.128"]
fn vfmaddsubph_128(a: __m128h, b: __m128h, c: __m128h) -> __m128h;
#[link_name = "llvm.x86.avx512fp16.vfmaddsub.ph.256"]
fn vfmaddsubph_256(a: __m256h, b: __m256h, c: __m256h) -> __m256h;
#[link_name = "llvm.x86.avx512fp16.vfmaddsub.ph.512"]
fn vfmaddsubph_512(a: __m512h, b: __m512h, c: __m512h, rounding: i32) -> __m512h;

Expand Down