This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: Test gcc.c-torture/compile/20001226-1.c takes too long to compile
- To: "Billinghurst, David (CRTS)" <David dot Billinghurst at riotinto dot com>
- Subject: Re: Test gcc.c-torture/compile/20001226-1.c takes too long to compile
- From: Jeffrey Oldham <oldham at codesourcery dot com>
- Date: Thu, 17 May 2001 20:55:46 -0700
- Cc: "'gcc-bugs at gcc dot gnu dot org'" <gcc-bugs at gcc dot gnu dot org>
- Organization: CodeSourcery LLC
- References: <8D00C32549556B4E977F81DBC24E985D1B19F8@crtsmail1.technol_exch.corp.riotinto.org>
On Fri, May 18, 2001 at 12:57:57AM -0000, Billinghurst, David (CRTS) wrote:
> Test gcc.c-torture/compile/20001226-1.c is causing me some pain on
> mips-sgi-irix6.5:
> * it takes around 300s to compile on a R10k with lots of RAM
> * with 2 ABIs to test, this test alone takes an hour
> * it consistently fails due to timeouts
>
> I wonder if a test that takes so long is appropriate.
I raised this issue with Geoff Keating 2000Dec. See
http://gcc.gnu.org/ml/gcc-patches/2000-12/msg01481.html and associated
links.
Also, I added some parameters to cut back on delayed branch running
time until someone rewrites the very inefficient code. Perhaps, you
can play with the default constants to reduce the running time. See
gcc/invoke.texi's max-delay-slot-insn-search and
max-delay-slot-live-search.
(David, would you be willing to send more information about the g++
execution failures? I will try to investigate, assuming someone
unbreaks the Irix computer I use, but I will be gone all weekend.
Thanks.)
Thanks,
Jeffrey D. Oldham
oldham@codesourcery.com