Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
Still no luck here.Could anyone push it so i can do a git pull? Thanks --- On Fri, 9/23/11, Thomas Albrecht wrote: From: Thomas Albrecht Subject: Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects To: flightgear-devel@lists.sourceforge.net Date: Friday, September 23, 2011, 11:24 PM Hey Csaba, hey Geoff, thanks a lot for looking into this! > void SGPagedLOD::forceLoad(osgDB::DatabasePager *dbp, FrameStamp* > framestamp, ? ? ? ? ? ? ? ? ? ? ? ? ???NodePath& path) > { > ? ? //SG_LOG(SG_GENERAL, SG_ALERT, "SGPagedLOD::forceLoad(" << > ? ? //getFileName(getNumChildren()) << ")"); > > And now the crazy part! If I uncomment this logging, everything > suddenly works, even with 20k objects: Same here -- uncommenting this part in simgear/scene/model/SGPagedLOD.cxx fixes the issue for me, too. 20k objects were loaded within 10 sec on my P4 2.4 GHz, though frame rate is ~5 then... Cheers, Tom -Inline Attachment Follows- -- All of the data generated in your IT infrastructure is seriously valuable. Why? It contains a definitive record of application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-d2dcopy2 -Inline Attachment Follows- ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- All of the data generated in your IT infrastructure is seriously valuable. Why? It contains a definitive record of application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-d2dcopy2___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
Hey Csaba, hey Geoff, thanks a lot for looking into this! > void SGPagedLOD::forceLoad(osgDB::DatabasePager *dbp, FrameStamp* > framestamp, ? ? ? ? ? ? ? ? ? ? ? ? ???NodePath& path) > { > ? ? //SG_LOG(SG_GENERAL, SG_ALERT, "SGPagedLOD::forceLoad(" << > ? ? //getFileName(getNumChildren()) << ")"); > > And now the crazy part! If I uncomment this logging, everything > suddenly works, even with 20k objects: Same here -- uncommenting this part in simgear/scene/model/SGPagedLOD.cxx fixes the issue for me, too. 20k objects were loaded within 10 sec on my P4 2.4 GHz, though frame rate is ~5 then... Cheers, Tom signature.asc Description: This is a digitally signed message part. -- All of the data generated in your IT infrastructure is seriously valuable. Why? It contains a definitive record of application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-d2dcopy2___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
Hi cannot find it. grep -r nor Qtcreator find it, which files are those or how should i search? I gets worse on my PC. I've several sceneries that hang so I need to investigate.Thanks --- On Thu, 9/22/11, Csaba Halász wrote: From: Csaba Halász Subject: Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects To: "FlightGear developers discussions" Date: Thursday, September 22, 2011, 3:13 AM On Wed, Sep 21, 2011 at 11:07 PM, Thomas Albrecht wrote: > > On machine (much slower than yours: Pentium 4 2.4 GHz, 1.5GB DDR), FG hangs > when using ~2300 objects. Here, with AMD 605e 2.3GHz, it hangs at around 5800 objects, with the CPU behaviour you described. Going slightly higher, from around 6000, FG starts to burn CPU again, but nevertheless won't get any result. Looking into the problem, seems the scenery is loaded eventually, but the fdm is not initialized so this check never passes: if (globals->get_tile_mgr()->isSceneryLoaded() && fgGetBool("sim/fdm-initialized")) { Now, the FDM init code has this: if (globals->get_scenery()->scenery_available(geod, range)) { SG_LOG(SG_FLIGHT, SG_INFO, "Scenery loaded, will init FDM"); That in turn ends up at: simgear::CheckSceneryVisitor csnv(getPagerSingleton(), toOsg(p), range_m, framestamp); // currently the PagedLODs will not be loaded by the DatabasePager // while the splashscreen is there, so CheckSceneryVisitor force-loads // missing objects in the main thread get_scene_graph()->accept(csnv); if(!csnv.isLoaded()) Finally we arrive at: void SGPagedLOD::forceLoad(osgDB::DatabasePager *dbp, FrameStamp* framestamp, NodePath& path) { //SG_LOG(SG_GENERAL, SG_ALERT, "SGPagedLOD::forceLoad(" << //getFileName(getNumChildren()) << ")"); And now the crazy part! If I uncomment this logging, everything suddenly works, even with 20k objects: http://i53.tinypic.com/wwn12f.png Sounds like some timing/threading issue to me. -- Csaba/Jester -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- All of the data generated in your IT infrastructure is seriously valuable. Why? It contains a definitive record of application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-d2dcopy2___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
Hi Tom, Thanks for adding the png texture... Now they look like trees ;=)) not square buildings... http://geoffair.org/tmp/fgfs-screen-003.png http://geoffair.org/tmp/fgfs-screen-004.png But still no problem loading in my machine... And switching to your second test, EHLE, using 10K trees, again no problems loading - http://geoffair.org/tmp/fgfs-screen-005.png And actually I chose to ADD the 10,000 python generated objects into the existing 3040154.stg... Naturally the scenery loading stage seemed longer, but no freeze... and my frame rate was down to about 15 fps... I even tried a _MAD_ example of 100,000 trees... This certainly took lots longer to load... the fading of the splash alone took about a minute... reduced my frame rate to 1, and the rendering engine refused to paint them all... http://geoffair.org/tmp/fgfs-screen-006.png Well the size of the block was continuing to grow all the time - I could see the rendering thread add trees, horizontal, row by row, so maybe if I had waited the time... Still no freeze exactly ;=() unless you count this very-slow-rendering... maybe that is what you are experiencing... with a Pentium 4... I had fgfs running for 20+ minutes to get that far even with my Core 2 Quad cpu... Or maybe, as Csaba points out, there may be some timing / threading issue, but nothing shows up on my machine... Enough FUN with trees ;=)) Now to remember to put my scenery back to what it was... Sorry can not help more... Regards, Geoff. On Wed, 2011-09-21 at 23:07 +0200, Thomas Albrecht wrote: > Hi Geoff, > > thanks for testing! Indeed, I forgot the texture, sorry about that. It is > included in a new package: http://www.mediafire.com/?q99zyzkyu2tw04w > > For further testing, I wrote a small python script which fills a rectangular > area at EHLE (because it's mostly flat there, so I can use hardcoded > elevation) with an arbitraty number of objects (all copies of that coco > palm). Could you please give it a try and see if you find any strange > behaviour with, say, 5k, 10k, 20k objects? > > Usage: > > backup Objects/e000n50/e005n52/3040154.stg first! > > tar xzf Objects.tar.gz > cd Objects/e000n50/e005n52 > ./place_objs.py 5000 > then start fgfs at EHLE > > On machine (much slower than yours: Pentium 4 2.4 GHz, 1.5GB DDR), FG hangs > when using ~2300 objects. > > Thanks, > Tom -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
Yes, I only found out that sometimes it can load after a restart... Mostly it hangs in conjunction with terragear trying to load terrain that I've flown multiple times already. So apart as Csaba mentioned, there might be some terragear code to improve as well? --- On Thu, 9/22/11, Csaba Halász wrote: From: Csaba Halász Subject: Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects To: "FlightGear developers discussions" Date: Thursday, September 22, 2011, 3:13 AM On Wed, Sep 21, 2011 at 11:07 PM, Thomas Albrecht wrote: > > On machine (much slower than yours: Pentium 4 2.4 GHz, 1.5GB DDR), FG hangs > when using ~2300 objects. Here, with AMD 605e 2.3GHz, it hangs at around 5800 objects, with the CPU behaviour you described. Going slightly higher, from around 6000, FG starts to burn CPU again, but nevertheless won't get any result. Looking into the problem, seems the scenery is loaded eventually, but the fdm is not initialized so this check never passes: if (globals->get_tile_mgr()->isSceneryLoaded() && fgGetBool("sim/fdm-initialized")) { Now, the FDM init code has this: if (globals->get_scenery()->scenery_available(geod, range)) { SG_LOG(SG_FLIGHT, SG_INFO, "Scenery loaded, will init FDM"); That in turn ends up at: simgear::CheckSceneryVisitor csnv(getPagerSingleton(), toOsg(p), range_m, framestamp); // currently the PagedLODs will not be loaded by the DatabasePager // while the splashscreen is there, so CheckSceneryVisitor force-loads // missing objects in the main thread get_scene_graph()->accept(csnv); if(!csnv.isLoaded()) Finally we arrive at: void SGPagedLOD::forceLoad(osgDB::DatabasePager *dbp, FrameStamp* framestamp, NodePath& path) { //SG_LOG(SG_GENERAL, SG_ALERT, "SGPagedLOD::forceLoad(" << //getFileName(getNumChildren()) << ")"); And now the crazy part! If I uncomment this logging, everything suddenly works, even with 20k objects: http://i53.tinypic.com/wwn12f.png Sounds like some timing/threading issue to me. -- Csaba/Jester -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
On Wed, Sep 21, 2011 at 11:07 PM, Thomas Albrecht wrote: > > On machine (much slower than yours: Pentium 4 2.4 GHz, 1.5GB DDR), FG hangs > when using ~2300 objects. Here, with AMD 605e 2.3GHz, it hangs at around 5800 objects, with the CPU behaviour you described. Going slightly higher, from around 6000, FG starts to burn CPU again, but nevertheless won't get any result. Looking into the problem, seems the scenery is loaded eventually, but the fdm is not initialized so this check never passes: if (globals->get_tile_mgr()->isSceneryLoaded() && fgGetBool("sim/fdm-initialized")) { Now, the FDM init code has this: if (globals->get_scenery()->scenery_available(geod, range)) { SG_LOG(SG_FLIGHT, SG_INFO, "Scenery loaded, will init FDM"); That in turn ends up at: simgear::CheckSceneryVisitor csnv(getPagerSingleton(), toOsg(p), range_m, framestamp); // currently the PagedLODs will not be loaded by the DatabasePager // while the splashscreen is there, so CheckSceneryVisitor force-loads // missing objects in the main thread get_scene_graph()->accept(csnv); if(!csnv.isLoaded()) Finally we arrive at: void SGPagedLOD::forceLoad(osgDB::DatabasePager *dbp, FrameStamp* framestamp, NodePath& path) { //SG_LOG(SG_GENERAL, SG_ALERT, "SGPagedLOD::forceLoad(" << //getFileName(getNumChildren()) << ")"); And now the crazy part! If I uncomment this logging, everything suddenly works, even with 20k objects: http://i53.tinypic.com/wwn12f.png Sounds like some timing/threading issue to me. -- Csaba/Jester -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
Hi Jon! > I've seen something similar before, it's incredibly annoying, and I suspect > you'll eventually track it down to a typo in an stg file. I've also had typos/missing objects in .stgs before, causing similiar effects. However, for testing, I'm creating the .stg with a very simple python script, placing the same object over and over at different locations. I am pretty sure it's not a typo or such. Cheers, Tom #!/usr/bin/env python import numpy as np import sys obj="Juliana-CocoPalmSeedling.ac" alt=-5 hdg=0 mid_x=5.519707 mid_y=52.456022 size_x=0.0050 size_y=0.0050 num_x=num_y=int(sys.argv[1])**0.5 X = np.linspace(-0.5*size_x, 0.5*size_x, num_x) + mid_x Y = np.linspace(-0.5*size_y, 0.5*size_y, num_y) + mid_y f = open("3040154.stg", "w") for y in Y: for x in X: f.write("OBJECT_STATIC %s %1.10g %1.10g %g %g\n" % (obj, x, y, alt, hdg)) f.close() signature.asc Description: This is a digitally signed message part. -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
Hi Geoff, thanks for testing! Indeed, I forgot the texture, sorry about that. It is included in a new package: http://www.mediafire.com/?q99zyzkyu2tw04w For further testing, I wrote a small python script which fills a rectangular area at EHLE (because it's mostly flat there, so I can use hardcoded elevation) with an arbitraty number of objects (all copies of that coco palm). Could you please give it a try and see if you find any strange behaviour with, say, 5k, 10k, 20k objects? Usage: backup Objects/e000n50/e005n52/3040154.stg first! tar xzf Objects.tar.gz cd Objects/e000n50/e005n52 ./place_objs.py 5000 then start fgfs at EHLE On machine (much slower than yours: Pentium 4 2.4 GHz, 1.5GB DDR), FG hangs when using ~2300 objects. Thanks, Tom signature.asc Description: This is a digitally signed message part. -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
+++ Geoff McLane [21/09/11 13:38 +0200]: >If you add that png, I would give it a try with 5000, >but the problems does seem to be in your machine at the >moment ;=(( I've seen something similar before, it's incredibly annoying, and I suspect you'll eventually track it down to a typo in an stg file. What happens is that the scenery engine never finishes loading the startup scenery, so everything hangs. When you fly *to* the airport in question flightgear is already running in its normal state and so doesn't have that "all the scenery is loaded" hurdle to get over. Start it up with some very verbose logging and you'll track it down. -- Jon Stockill li...@stockill.net -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FG hangs on "loading scenery" when using many objects
Hi Tom, He, he, just for ***FUN*** I tried with 4000 objects... See the 'crowded' runway on startup - http://geoffair.org/tmp/fgfs-screen-001.png but more spectacular from the air - http://geoffair.org/tmp/fgfs-screen-002.png and had no problems loading ;=)) But you seem to have missed the png texture file from your zip, and maybe that would make a difference? Console output - several times - osgDB ac3d reader: could not find texture "CocoPalmSeedling.png" Did not check the cpu usage, and the total load time was well under a minute... about 30 seconds... If you add that png, I would give it a try with 5000, but the problems does seem to be in your machine at the moment ;=(( HTH. Regards, Geoff. System Information: - SG/FG/FGDATA - Git of 2011/09/14 12:30:29 - OSG 3.0.1, PLIB 1.8.5, boost 1.40.00 - Ubuntu 10.04 (lucid) linux - 2.6.32-33-generic - Video: ATI Technologies Inc RV630 [Radeon HD 2600XT] - ATI/AMD proprietary fglrx driver - $ modinfo fglrx srcversion: 33640324E1E3A5E6A7234AC - CPU: Intel Core 2 Quad Q6600 2.40 GHz - RAM: 2 GB - Box: Dell XPS 420 - Dell monitor at 1440x900, 60 Hz Scripts used - after failing to load the new stg the first time, I rename the 1907463.stg in my Scenery-1.0.1, and copied in 1907463.stg.4000, plus the *.ac file... then ran - #!/bin/sh #< run_test.sh #SCENERY="--fg-scenery=/home/downloads/temp:/home/geoff/Scenery-1.0.1" SCENERY="--fg-scenery=/home/geoff/Scenery-1.0.1" OPTS="--prop:/sim/frame-rate-throttle-hz=30 --disable-random-objects --geometry=1920x1190+0+0" OPTS="$OPTS --atlas=socket,out,1,localhost,5500,udp --season=summer $SCENERY" OPTS="$OPTS --aircraft=c172p --airport=TNCM --log-level=alert" OPTS="$OPTS --prop:/sim/rendering/multi-sample-buffers=true --prop:/sim/rendering/multi-samples=2" OPTS="$OPTS --prop:/sim/ai-traffic/enabled=false --prop:/sim/traffic-manager/enabled=false --prop:/sim/atc/enabled=false" OPTS="$OPTS --timeofday=dawn --enable-real-weather-fetch --control=joystick --disable-auto-coordination" ./run_fgfs.sh $OPTS which runs - #!/bin/sh #< run_fgfs.sh version=1.3.4, 2011-09-12 - run from anywhere BN=`basename $0` cd /home/geoff/fg/fg16 HERE=$PWD cd install/fgfs/bin export LD_LIBRARY_PATH=/home/geoff/fg/fg16/install/OSG301/lib echo "$BN: Running: ./fgfs --fg-root=/home/geoff/fg/fg16/fgfs/data $@" ./fgfs --fg-root=/home/geoff/fg/fg16/fgfs/data $@ On Tue, 2011-09-20 at 23:00 +0200, Thomas Albrecht wrote: > Hey group, > > I've come across a problem with FG when many (static) objects are to be > loaded > on FG startup. > > Usually, on my (faily old) PC FG loads for about 20 seconds, then > says "loading scenery" for about 6 seconds, and places me in the c172 ready > for takeoff. During all this CPU load is at 100%. > > Now if I start FG with scenery that contains many (> 1000) objects, I get the > following: > - FG still loads for about 20 sec, then says "loading scenery" > - about 3 seconds later, CPU load drops to ~20% and stays there > - and FG never finishes startup > --log-level=debug shows the main loop is running, I can use the menu, but I > never end up in the c172, nor see anything else but the splash screen > > Starting at a nearby airport and flying into said scenery works. I can also > teleport to this nearby airport while FG 'hangs', and then fly into said > scenery flawlessly. > > I've created a test scenery [1] which uses TNCM terrain and 5000 instances of > one object, furthermore, a script which lets me reduce the number of objects > in the .stg. If I use 3100 objects, everything is fine. 3200 objects, and FG > hangs. 100% repeatable, though I did not narrow down the threshold number > further. > > However, the threshold number seems to depend on > - the object(s) loaded > - CPU load: If I have another process running (mplayer, for example) which > consumes some CPU, FG now also hangs for the 3100 objects case (which would > otherwise load fine if there was no other demanding process). > > It appears as if FG somewhat locks up if the initial scenery is not loaded > within a certain wall clock time. > > Any ideas? > > Cheers, > Tom > > - Git from 5 Sep 2011 > - Gentoo Linux > - GeForce 7600 GS (running ancient nvidia drivers 180.29) > > fgfs --prop:/sim/frame-rate-throttle-hz=30 --disable-random-objects > --geometry=1920x1190+0+0 --atlas=socket,out,1,localhost,5500,udp > --fg-root=/home/tom/daten/fgfs/src/fgdata --season=summer > --fg-scenery=/home/tom/fgfs/home/Scenery-Manual:/home/tom/fgfs/home/Scenery-TerraSync:/home/tom/daten/fgfs/src/fgdata/Scenery:/home/tom/fgfs/home/Scenery-1.0.1 > --aircraft=c172p --airport=TNCM --log-level=alert > --prop:/sim/rendering/multi-sample-buffers=true > --prop:/sim/rendering/multi-samples=2 --prop:/sim/ai-traffic/enabled=false > --prop:/sim/traffic-manager/enabled=false --prop:/sim/atc/enabled=false > --timeofday=dawn --enable-real-weather-fetch --control=joystick > --disable-auto-coordination > > [1] http://www.mediafire.com/?n9uftx7vi
[Flightgear-devel] FG hangs on "loading scenery" when using many objects
Hey group, I've come across a problem with FG when many (static) objects are to be loaded on FG startup. Usually, on my (faily old) PC FG loads for about 20 seconds, then says "loading scenery" for about 6 seconds, and places me in the c172 ready for takeoff. During all this CPU load is at 100%. Now if I start FG with scenery that contains many (> 1000) objects, I get the following: - FG still loads for about 20 sec, then says "loading scenery" - about 3 seconds later, CPU load drops to ~20% and stays there - and FG never finishes startup --log-level=debug shows the main loop is running, I can use the menu, but I never end up in the c172, nor see anything else but the splash screen Starting at a nearby airport and flying into said scenery works. I can also teleport to this nearby airport while FG 'hangs', and then fly into said scenery flawlessly. I've created a test scenery [1] which uses TNCM terrain and 5000 instances of one object, furthermore, a script which lets me reduce the number of objects in the .stg. If I use 3100 objects, everything is fine. 3200 objects, and FG hangs. 100% repeatable, though I did not narrow down the threshold number further. However, the threshold number seems to depend on - the object(s) loaded - CPU load: If I have another process running (mplayer, for example) which consumes some CPU, FG now also hangs for the 3100 objects case (which would otherwise load fine if there was no other demanding process). It appears as if FG somewhat locks up if the initial scenery is not loaded within a certain wall clock time. Any ideas? Cheers, Tom - Git from 5 Sep 2011 - Gentoo Linux - GeForce 7600 GS (running ancient nvidia drivers 180.29) fgfs --prop:/sim/frame-rate-throttle-hz=30 --disable-random-objects --geometry=1920x1190+0+0 --atlas=socket,out,1,localhost,5500,udp --fg-root=/home/tom/daten/fgfs/src/fgdata --season=summer --fg-scenery=/home/tom/fgfs/home/Scenery-Manual:/home/tom/fgfs/home/Scenery-TerraSync:/home/tom/daten/fgfs/src/fgdata/Scenery:/home/tom/fgfs/home/Scenery-1.0.1 --aircraft=c172p --airport=TNCM --log-level=alert --prop:/sim/rendering/multi-sample-buffers=true --prop:/sim/rendering/multi-samples=2 --prop:/sim/ai-traffic/enabled=false --prop:/sim/traffic-manager/enabled=false --prop:/sim/atc/enabled=false --timeofday=dawn --enable-real-weather-fetch --control=joystick --disable-auto-coordination [1] http://www.mediafire.com/?n9uftx7vil98btz signature.asc Description: This is a digitally signed message part. -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel