Alright, so there isn't a clear culprit, but there's a good chance it's the heuristics for STUN - after 1.8 a conversation lookup was added (r44068), and that uses a g_hash_table. The only major difference between the two profiles that I found was a significant increase in calls to g_hash_table_lookup from the conversation code.
If you don't need STUN decoding, try disabling its heuristic (there's no preference, but just commenting out the heur_dissector_add call should work) and see if that improves things. If that doesn't work I'm not sure what the problem could be. Evan PS the new name-resolution code also uses g_hash_tables now, but the call source is comparatively small next to the conversation code, so I don't think that's the problem. It may be worth trying a 1.10 build as well though, just for comparison (since that will have the STUN changes but not the name-resolution changes). On Thu, Sep 12, 2013 at 9:40 AM, Anders Broman <anders.bro...@ericsson.com>wrote: > File from tshark 1.8.2 > /Anders > -----Original Message----- > From: wireshark-dev-boun...@wireshark.org [mailto: > wireshark-dev-boun...@wireshark.org] On Behalf Of Anders Broman > Sent: den 12 september 2013 14:54 > To: Developer support list for Wireshark > Subject: Re: [Wireshark-dev] Performance degradation in trunk > > > > -----Original Message----- > From: wireshark-dev-boun...@wireshark.org [mailto: > wireshark-dev-boun...@wireshark.org] On Behalf Of Evan Huus > Sent: den 12 september 2013 13:15 > To: bal...@balintreczey.hu; Developer support list for Wireshark > Subject: Re: [Wireshark-dev] Performance degradation in trunk > > On 2013-09-12, at 5:40 AM, Bálint Réczey <bal...@balintreczey.hu> wrote: > > >> Could they analyze perf results or run git bisect to find the point > >> when the degradation started? > > >Since this is on Ubuntu (which can run valgrind) just use the -p option > to tools/valgrind-wireshark.sh and it will produce a performance profile. > Do that for >both builds and we'll have something to work with. > > > >Evan > > Here's the output from "trunk", how do you get anything useful from it? > /Anders > > >PS I suspect some new heuristic dissector has been introduced, since none > of the other changes from 1.8 to trunk are likely to have such a > significant >performance impact. > > > Cheers, > > Balint > > > > 2013/9/12 Anders Broman <anders.bro...@ericsson.com>: > >> Hi, > >> > >> I recently supplied some people with an internal build from recent > >> trunk and got complaints on performance. Switching back > >> > >> To the default Wireshark on the Ubuntu 13.04 system 1.8.2 reading of > >> a 400M trace file with SIP Diameter etc traffic is more than twice as > fast. > >> > >> Regards > >> > >> Anders > > > ___________________________________________________________________________ > > Sent via: Wireshark-dev mailing list <wireshark-dev@wireshark.org> > > Archives: http://www.wireshark.org/lists/wireshark-dev > > Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev > > > > mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe > ___________________________________________________________________________ > Sent via: Wireshark-dev mailing list <wireshark-dev@wireshark.org> > Archives: http://www.wireshark.org/lists/wireshark-dev > Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev > mailto:wireshark-dev-requ...@wireshark.org > ?subject=unsubscribe > > ___________________________________________________________________________ > Sent via: Wireshark-dev mailing list <wireshark-dev@wireshark.org> > Archives: http://www.wireshark.org/lists/wireshark-dev > Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev > mailto:wireshark-dev-requ...@wireshark.org > ?subject=unsubscribe >
___________________________________________________________________________ Sent via: Wireshark-dev mailing list <wireshark-dev@wireshark.org> Archives: http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
