Issue 84152
Summary [AArch64] there is redundant neg instruction for mull with complex type
Labels new issue
Assignees
Reporter vfdff
    * **test**: https://gcc.godbolt.org/z/enYjsjGfE
```
define ptr @_ZNSt7complexIiEmLIiEERS0_RKS_IT_E(ptr noundef nonnull align 4 dereferenceable(8) %a, 
                                               ptr noundef nonnull align 4 dereferenceable(8) %b) align 2 {
entry:
  %0 = load i32, ptr %a, align 4
  %1 = load i32, ptr %b, align 4
  %mul = mul nsw i32 %1, %0
  %_M_imag = getelementptr inbounds i8, ptr %a, i64 4
 %2 = load i32, ptr %_M_imag, align 4
  %_M_imag.i = getelementptr inbounds i8, ptr %b, i64 4
  %3 = load i32, ptr %_M_imag.i, align 4
  %mul3 = mul nsw i32 %3, %2
  %sub = sub nsw i32 %mul, %mul3
  %mul6 = mul nsw i32 %3, %0
  %mul9 = mul nsw i32 %2, %1
  %add = add nsw i32 %mul6, %mul9
  store i32 %add, ptr %_M_imag, align 4
  store i32 %sub, ptr %a, align 4
  ret ptr %a
}
```

* **llvm**:  transform the madd into fmls, then the **neg instruction**  is redundant 
```
std::complex<int>& std::complex<int>::operator*=<int>(std::complex<int> const&):     // @std::complex<int>& std::complex<int>::operator*=<int>(std::complex<int> const&)
        ldp     w11, w8, [x0]
        ldp     w10, w9, [x1]
 mul     w12, w9, w8
        mul     w8, w8, w10
        madd w8, w9, w11, w8
        neg     w13, w12
        madd    w9, w10, w11, w13  ; w9 = w10 * w11 + w13 = w9 = w10 * w11 - w12 =  w10 * w11 - w9 * w8 = fmls w10 * w11, w9, 28
        stp     w9, w8, [x0]
 ret
```

* the IR is simlplied from https://gcc.godbolt.org/z/fPnnxGhb3
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs

Reply via email to