https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85048
Devin Hussey <husseydevin at gmail dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |husseydevin at gmail dot com
--- Comment #5 from Devin Hussey <husseydevin at gmail dot com> ---
ARM/AArch64 NEON use these:
From To Intrinsic ARMv7-a AArch64
intXxY_t -> int2XxY_t vmovl_sX vmovl.sX sshll #0?
uintXxY_t. -> uint2XxY_t vmovl_uX vmovl.uX ushll #0?
[u]int2XxY_t -> [u]intXxY_t vmovn_[us]X vmovn.iX xtn
floatXxY_t -> intXxY_t vcvt[q]_sX_fX vcvt.sX.fX fcvtzs
floatXxY_t -> uintXxY_t vcvt[q]_uX_fX vcvt.uX.fX fcvtzu
intXxY_t -> floatXxY_t vcvt[q]_fX_sX vcvt.fX.sX scvtf
uintXxY_t -> floatXxY_t vcvt[q]_fX_uX vcvt.fX.uX ucvtf
float32x2_t -> float64x2_t vcvt_f32_f64 2x vcvt.f64.f32 fcvtl
float64x2_t -> float32x2_t vcvt_f64_f32 2x vcvt.f32.f64 fcvtn
Clang optimizes vmovl to vshll by zero for some reason.
float32x2_t <-> float64x2_t requires 2 VFP instructions on ARMv7-a.
