[LLVMdev] Float undef value propagation

[Tim Northover]

On 11 December 2013 20:08, Raoux, Thomas F <thomas.f.raoux at intel.com> wrote:
> You are right some cases would definitely not be right like undef + Nan ->
> undef. For 2.0f case I’m not sure either if any bits could be known.

Do you have an example where that's unsafe? You usually have to do
something pretty bad to get an undef in the first place. Since
IEEE-754 doesn't have that concept I can't see how it could have
anything to say on how LLVM optimises it.

Cheers.

Tim.