https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80432
Markus Trippelsdorf <trippels at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |trippels at gcc dot gnu.org
--- Comment #4 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
(In reply to Hao Zhang from comment #3)
> (In reply to Hao Zhang from comment #2)
> > (In reply to Martin Liška from comment #1)
> > > Can't reproduce, can you please provide options you use to build the
> > > executable?
> >
> > I didn't use any extra options besides g++ pow.cpp, and both my CFLAGS and
> > CXXFLAGS environment variables are empty. And I just found out moments ago
> > compiling with -O1 gives the right result while compiling with -O0 gives the
> > wrong result. I also tried to turn on each of the 42 optimizations provided
> > by -O1 individually, and I found none of the 42 optimizations ALONE can get
> > rid of the problem. So there must be some combinations of optimizations that
> > can get rid of the problem.
> >
> > My glibc version is 2.25 and kernel version is 4.10.9. And I have reproduced
> > the error on two different computers. Is there any other information I can
> > provide to help pinpoint the problem? Thanks!
>
> I also found out that g++ -O1 pow.cpp gives the expected results, while the
> following command gives the wrong results:
>
> g++ -fauto-inc-dec -fbranch-count-reg -fcombine-stack-adjustments
> -fcompare-elim -fcprop-registers -fdce -fdefer-pop -fdelayed-branch -fdse
> -fforward-propagate -fguess-branch-probability -fif-conversion2
> -fif-conversion -finline-functions-called-once -fipa-pure-const
> -fipa-profile -fipa-reference -fmerge-constants -fmove-loop-invariants
> -freorder-blocks -fshrink-wrap -fshrink-wrap-separate -fsplit-wide-types
> -fssa-backprop -fssa-phiopt -ftree-bit-ccp -ftree-ccp -ftree-ch
> -ftree-coalesce-vars -ftree-copy-prop -ftree-dce -ftree-dominator-opts
> -ftree-dse -ftree-forwprop -ftree-fre -ftree-phiprop -ftree-sink -ftree-slsr
> -ftree-sra -ftree-pta -ftree-ter -funit-at-a-time -fomit-frame-pointer
> pow.cpp
>
> This is weird, isn't the two forms of command equivalent?
No.
And with -O0 it will just call powl. So if you get weird results your glibc is
to blame.
