[all-commits] [llvm/llvm-project] 7139d0: [AArch64] Add intrinsic support for Fdot instr. (#...

[Lukacma via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 7139d0b46f6d71334b74049bbd540ac4b18a1bc5
      https://github.com/llvm/llvm-project/commit/7139d0b46f6d71334b74049bbd540ac4b18a1bc5
  Author: Lukacma <Marian.Lukac at arm.com>
  Date:   2026-06-02 (Tue, 02 Jun 2026)

  Changed paths:
    M clang/include/clang/Basic/arm_neon.td
    M clang/lib/CodeGen/TargetBuiltins/ARM.cpp
    A clang/test/CodeGen/AArch64/f16f32dot-intrinsics.c
    M clang/test/Sema/aarch64-neon-immediate-ranges/dotprod.c
    M llvm/include/llvm/IR/IntrinsicsAArch64.td
    M llvm/lib/Target/AArch64/AArch64InstrFormats.td
    A llvm/test/CodeGen/AArch64/aarch64-f16f32dot-intrinsics.ll

  Log Message:
  -----------
  [AArch64] Add intrinsic support for Fdot instr. (#189987)

This patch adds intrinsics for new NEON Fdot instruction variants. The
implementation is based on the
[proposal](https://github.com/ARM-software/acle/pull/428) and adds these
ACLE builtins:

```
float32x2_t vdot_f32_f16(float32x2_t r, float16x4_t a, float16x4_t b);
float32x4_t vdotq_f32_f16(float32x4_t r, float16x8_t a, float16x8_t b);

float32x2_t vdot_lane_f32_f16(float32x2_t r, float16x4_t a, float16x4_t b, const int lane);
float32x4_t vdotq_laneq_f32_f16(float32x4_t r, float16x8_t a, float16x8_t b, const int lane);
float32x2_t vdot_laneq_f32_f16(float32x2_t r, float16x4_t a, float16x8_t b, const int lane);
float32x4_t vdotq_lane_f32_f16(float32x4_t r, float16x8_t a, float16x4_t b, const int lane);
```



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications