Function core::arch::x86_64::_mm512_cvtne2ps_pbh[−][src]

pub unsafe fn _mm512_cvtne2ps_pbh(a: __m512, b: __m512) -> __m512bh

🔬 This is a nightly-only experimental API. (stdsimd #48556)

This is supported on x86-64 and target feature avx512bf16,avx512f only.

Expand description

将两个 512 位 vectors a 和 b 中的包装的单精度 (32-bit) 浮点元素转换为包装的 BF16 (16-bit) 浮点元素，并将结果存入 a