[LLVMdev] Representing -ffast-math at the IR level
Renato Golin
rengolin at systemcall.org
Mon Apr 16 05:29:53 PDT 2012
On 16 April 2012 13:23, Duncan Sands <baldrick at free.fr> wrote:
> what if Accuracy is NaN?
D'oh. ;)
--
cheers,
--renato
http://systemcall.org/
More information about the llvm-dev
mailing list