sqrt

What is the state of sqrt in LLVM?

It was an intrinsic but there are no OCaml bindings for it and, last I looked,
it generated inefficient code on Linux due to this bug:

  http://www.llvm.org/PR3219

Is the intrinsic deprecated? Am I losing a lot of performance by calling sqrt
from libm instead of using the intrinsic?

There is a fundamental difference between sqrt() and llvm.sqrt: the former is defined on negative values and sets errno (on linux). The later is undefined. Both work well for their stated purpose, llvm.sqrt should not be slower than sqrt even on linux. Both llvm.sqrt and sqrt could be much better on linux, but noone seems compelled to do the work.

-Chris

Many platforms could also benefit from recognizing 1.0/sqrt() as rsqrt().

deep

What exactly is the work for llvm.sqrt on Linux?

Ah sorry, llvm.sqrt works fine on linux. The issue is that a raw call to sqrt() in a C program doesn't typically compile to llvm.sqrt on linux, because it sets errno. This can be controlled with -fmath-errno.

-Chris

It is more than errno. sqrt() should't compile to llvm.sqrt on any platform that uses IEEE754 math, because IEEE sqrt() is well defined on negative arguments and llvm.sqrt isn't. There are currently no command line arguments to override this.

Hi Dale,

It is more than errno. sqrt() should't compile to llvm.sqrt on any platform that uses IEEE754 math, because IEEE sqrt() is well defined on negative arguments and llvm.sqrt isn't. There are currently no command line arguments to override this.

-funsafe-math-optimizations?

Ciao,

Duncan.