[RFA 2/4] dwarf2_physname

Tom Tromey tromey@redhat.com
Thu Dec 17 20:19:00 GMT 2009

>>>>> "Keith" == Keith Seitz <keiths@redhat.com> writes:

Keith> Okay, so I have some performance numbers for my box using CVS HEAD
Keith> patched and unpatched with my dwarf2_physname patches.


It looks to me like dwarf2_physname is not significantly slower in the
normal (attach and bt) cases -- only in the -readnow case.

Keith> dwarf2_realname:
Keith> real    774m37.623s

On irc Keith suggested that this may be due to swapping.

I probably wouldn't have let this test finish, myself :-)

Keith> As you can see, --readnow (and by extrapolation "info func" and
Keith> "info type") take an enormous performance hit.

I am not concerned about this.  I don't think -readnow is actually used
-- I can't imagine why anyone would want to use it.  And, on a big
program, an unrestricted "info func" or "info type" is probably just a
mistake, like hitting enter too soon.

I think we should wait a bit to see if there are other comments,
particularly because I imagine a lot of people are going on vacation


More information about the Gdb-patches mailing list