[Bug tree-optimization/24333] missed div optimizations

[gabravier at gmail dot com via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=24333

Gabriel Ravier  changed:

   What|Removed |Added

 CC||gabravier at gmail dot com

--- Comment #8 from Gabriel Ravier  ---
Seems to be fixed since GCC 7

[Bug tree-optimization/24333] missed div optimizations

[pluto at agmk dot net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24333

--- Comment #7 from Pawel Sikora  2011-06-22 21:56:27 
UTC ---
4.6 still affected:

f:  movl%edi, %edx
movl%edi, %eax
sarl$31, %edx
idivl   %edi
ret

[Bug tree-optimization/24333] missed div optimizations

[pluto at agmk dot net]

--- Comment #6 from pluto at agmk dot net  2007-03-19 09:15 ---
`int f( int x ) { return x / x; }' still not optimized on 4.2/4.3


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24333

[Bug tree-optimization/24333] missed div optimizations

[pinskia at gcc dot gnu dot org]

--- Comment #5 from pinskia at gcc dot gnu dot org  2006-04-08 01:41 ---
x/x is done for fp now.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24333

[Bug tree-optimization/24333] missed div optimizations

[pinskia at gcc dot gnu dot org]

--- Comment #4 from pinskia at gcc dot gnu dot org  2005-12-30 18:33 ---
Confirmed.


-- 

pinskia at gcc dot gnu dot org changed:

   What|Removed |Added

 Status|UNCONFIRMED |NEW
 Ever Confirmed|0   |1
   Keywords||missed-optimization
   Last reconfirmed|-00-00 00:00:00 |2005-12-30 18:33:47
   date||
Summary|missed div optimizations?   |missed div optimizations


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24333

[Bug tree-optimization/24333] missed div optimizations?

[pinskia at gcc dot gnu dot org]

--- Comment #3 from pinskia at gcc dot gnu dot org  2005-10-12 16:04 ---
(In reply to comment #2)
> basing on the http://en.wikipedia.org/wiki/L'hospital we get:
> 
> lim  (x / x) = lim (x' / x') = lim (1 / 1) = 1
> x->0   x->0x->0

for float/float, we should get NANs with -ffast-math which assumes
-ffinite-math, for long/long and unsigned/unsigned, we have this thing called
undefinedness. BUT for java, 0/0 is defined to throw IIRC so we can only do
that for signed/signed and with -fno-wrapv.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24333

[Bug tree-optimization/24333] missed div optimizations?

[pluto at agmk dot net]

--- Comment #2 from pluto at agmk dot net  2005-10-12 15:57 ---
basing on the http://en.wikipedia.org/wiki/L'hospital we get:

lim  (x / x) = lim (x' / x') = lim (1 / 1) = 1
x->0   x->0x->0

should we trap/check anything?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24333

[Bug tree-optimization/24333] missed div optimizations?

[pinskia at gcc dot gnu dot org]

--- Comment #1 from pinskia at gcc dot gnu dot org  2005-10-12 15:42 ---
(In reply to comment #0)

This one, we have to deal with x = 0, what should we do, trap or not to trap.
> unsigned foo(const unsigned x) {  return (x / x); }


For the following one:
> double bar(const double x) { return (x / x); }

Only with -ffast-math.


-- 

pinskia at gcc dot gnu dot org changed:

   What|Removed |Added

   Severity|normal  |enhancement


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24333