This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug optimization/10679] [3.3/3.4] --param min-inline-insns not honoured


PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10679



------- Additional Comments From rguenth at tat dot physik dot uni-tuebingen dot de  2003-07-19 15:02 -------
Subject: Re:  [3.3/3.4] --param min-inline-insns not
 honoured

On 18 Jul 2003, pinskia at physics dot uc dot edu wrote:

> The check is still there but if the function that gcc is inlining is empty (which only happens with the
> new inlining counting mechanism), gcc will now always inline that function so that part is fixed, if
> on the other hand the function is not empty, you can and will hit that limit which you pointed out.

Yes, constructing a testcase for this bug is not easy, but placing a
warning inside the check causes >10000 hits on a medium size POOMA based
application. So this bug really triggers and hurts both runtime and
codesize.

Richard.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]