[llvm-bugs] [Bug 139044] Calls to sin() don't get converted to sinf, leading to inefficient use of vector variants with -fveclib=ArmPL

[LLVM Bugs via llvm-bugs]

Issue	139044
Summary	Calls to sin() don't get converted to sinf, leading to inefficient use of vector variants with -fveclib=ArmPL
Labels	backend:AArch64, SVE
Assignees
Reporter	willlovett-arm

    https://godbolt.org/z/fE16nGPxP

```
#include <math.h>

void foo(float *p)
{
for (int x=0; x<999; ++x) {
    p[x] = sin(p[x]);
  }
}

```

generates calls to `armpl_vsinq_f64` instead of `armpl_vsinq_f32`.

If we replace that code with `sinf` it works fine:
```
#include <math.h>

void foo(float *p)
{
for (int x=0; x<999; ++x) {
    p[x] = sinf(p[x]);
  }
}
```

I'm reasonably sure it's legitimate to implicitly switch to sinf(), and this would lead to double the vector throughput.

_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs