This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Temporarily enable ac_fold_checking for --enable-checking=yes.
- From: "Kaveh R. GHAZI" <ghazi at caip dot rutgers dot edu>
- To: Sebastian Pop <sebpop at gmail dot com>
- Cc: Uros Bizjak <ubizjak at gmail dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Fri, 22 Jun 2007 12:12:58 -0400 (EDT)
- Subject: Re: [PATCH] Temporarily enable ac_fold_checking for --enable-checking=yes.
- References: <467AF510.3010907@gmail.com> <cb9d34b20706211533w1fb02383vefb73f056eb9edbe@mail.gmail.com>
On Fri, 22 Jun 2007, Sebastian Pop wrote:
> Hi,
>
> just some numbers from my bootstrap and testing. Currently, when
> configuring without touching to the enable-checking and with all
> default languages, I have to wait for a bootstrap for:
>
> Elapsed (wall clock) time (h:mm:ss or m:ss): 1:26:33
>
> with the following extra arguments to configure,
> --enable-checking=assert,fold,gc,misc,rtl,rtlflag,runtime,tree
> bootstrap finishes after:
>
> Elapsed (wall clock) time (h:mm:ss or m:ss): 3:02:26
Note if your goal is to use existing checking flags plus N extra ones,
it's cleaner IMHO to use: --enable-checking=yes,rtl,fold. Aside from
better readability, using "yes" rather than enumerating them avoids
accidentally forgetting one. E.g. in this case you left out the "df" flag
from the default option list. This makes your comparison invalid because
it compares default to +rtl+fold-df, not +rtl+fold as you meant it. I'd
like to see a comparison of default vs. "yes,rtl" vs. "yes,fold" vs.
"yes,rtl,fold" (if you have the free cpu time). That way we could see
which one yields the worst time regression.
On whether to include extra flags in the default, I've proposed this in
the past and was rejected for precisely the bootstrap time issue.
http://gcc.gnu.org/ml/gcc-patches/2000-10/msg00756.html
However IMHO I feel the sanity checking is worth it.
--Kaveh
--
Kaveh R. Ghazi ghazi@caip.rutgers.edu