This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: Very slow link times on debug versions of C++ program on x86_64. Includes oprofile of ld...
- From: Daniel Jacobowitz <drow at false dot org>
- To: "Talbot, George" <Gtalbot at locuspharma dot com>
- Cc: "H. J. Lu" <hjl at lucon dot org>, binutils at sourceware dot org
- Date: Wed, 20 Jun 2007 11:19:44 -0400
- Subject: Re: Very slow link times on debug versions of C++ program on x86_64. Includes oprofile of ld...
- References: <BAE72CC195BC2C45A07D1C1C161F16662562DA@ld3.discovery.com>
On Wed, Jun 20, 2007 at 09:58:42AM -0400, Talbot, George wrote:
> Thanks for your quick reply.
>
> I'll see if I can come up with a testcase.
>
> Based upon my examination of oprofile results, I can say that ld is
> _not_ spending all of its time in I/O, like the previous message
> suggests. Looking at the oprofile results, it's spending it's time
> matching wildcards, though of what I'm not exactly sure.
>
> What exactly does the "wild_section"/"match_simple_wild" stuff do in
> there? I'm not sure that I fully understand it. It appears to be
> matching section names, yes? That seems kinda weird--does a GCC debug
> compile have a lot more sections in the .o file?
Have you tried building a HEAD checkout or snapshot? If not, I highly
recommend it. I think this has been fixed already.
--
Daniel Jacobowitz
CodeSourcery