This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Basic block profiling (-a) does produce wrong code for insnprologues
- To: Richard Henderson <rth at redhat dot com>
- Subject: Re: Basic block profiling (-a) does produce wrong code for insnprologues
- From: Scott A Crosby <crosby at qwes dot math dot cmu dot edu>
- Date: Mon, 12 Nov 2001 20:56:59 -0500 (EST)
- Cc: Hartmut Penner <HPENNER at de dot ibm dot com>, <gcc at gcc dot gnu dot org>
On Sun, 4 Nov 2001, Richard Henderson wrote:
> On Mon, Nov 05, 2001 at 08:32:22AM +0100, Hartmut Penner wrote:
> > Shouldn'd it be better to delete this options in general?
>
> Yes. I keep meaning to do that.
>
Will it be replaced with an equivalent option that does the same thing?
Because it is *really* useful when profiling or analyzing code. I've used
it to determine the dynamic behaivor of running code; for example the
hit-rate of caches.
I run 2.95.2, which seems to work OK with this option, I would be unhappy
to see it go, without being replaced by something equally useful.
Or, are you proposing to depreciate '-a' and instead move to using
'-fprofile-arcs' and get gprof to understand that output format?
Scott