[Xastir-Dev] Re: Re: CPU still too busy for the wrong reasons

[Jerry Chamberlin]

I see all as USER CPU Here RH 7.2
All Maps off All interfaces off.

Still a big lag when centering the screen.
Suspect the number of calls is the table, but the bottom line says loading maps
Clear all stations takes forever (Long time < 5 minutes ) uptime:about a week
No Stations No CPU or Lag.


On Fri, 28 Nov 2003, Tom Russo wrote:

> On Fri, Nov 28, 2003 at 05:59:28PM -0800, a Mr. Richard Feyler of Fort Lee, New Jersey <archer at eskimo.com> writes 'Dear Rosanne Rosannadanna':
> > On Fri, 28 Nov 2003, Jerry Chamberlin wrote:
> > 
> > > Mine is RH 6.2 and RG 7.2 both do it.
> > 
> > Is there a case where it doesn't do it?  For instance try shutting
> > off all interfaces, then try turning off dead-reckoning, then try
> > turning off weather alerts, etc.  All the things I mentioned in my
> > other e-mail (plus weather alerts).
> 
> In my case, none of these suggestions has any effect.  If I am running
> xastir, it is taking all available CPU.  Right now, with all interfaces
> turned off, dead reckoning off, weather alerts off, all networked maps 
> disabled, basically just sitting there displaying static maps and no
> stations, it's taking 97% of the laptop's CPU.
> 
> 
> I did build xastir with profiling turned on and checked what's taking the
> time.  In a 3 minute run, UpdateTime was called once.  During that
> call, there were some amazingly huge numbers:  sec_now was called 2.3million
> times.  check_statusline_timout, popup_timeout_check, 
> check_and_transmit_objects_items, and several other routines had 236,626 calls 
> each. 
> 
> There were also a very huge number of calls to various pthread-related
> calls:  I'm out of practice reading gprof output, but it looks like
> over 70 million calls to pthread_mutex_lock.  Remember, this is in a 3 minute
> run of xastir, with one internet server interface turned on.
> 
> > If Xastir does it in all conditions, it's most likely that the
> > UpdateTime() routine is running too often.  
> 
> In my case it's doing this under all conditions.  Gprof shows only a single
> UpdateTime call in the whole 3 minutes, but it and its children took up 
> 63.15 seconds and 76.2% of the run time.  (not sure how gprof gets that number 
> --- it says main and *its* decendents took only 67.44 seconds and accounted 
> for 81% of the time)
> 
> -- 
> Tom Russo    KM5VY   SAR502   DM64ux          http://www.swcp.com/~russo/
> Tijeras, NM  QRPL#1592 K2#398  SOC#236 AHTB#1 http://www.qsl.net/~km5vy/
>  echo "prpv_a'rfg_cnf_har_cvcr" | sed -e 's/_/ /g' | tr [a-m][n-z] [n-z][a-m]
> 

The Net Lab year 2000 and beyond Internet Education is Science
http://www.netlab.org
WA0JRJ - Jerry
used to be ICQ 6408731
used to be AIM PappyJerry