Thread: [TuxKart-devel] unloadRace() function

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

If someone else could write it that would be very nice of them, I don't
know how all that materials stuff etc works so I don't know how to go
about unloading it.

Thanks,

	James

Hi there guys,
  i'm just confirming that tuxkart you can play with 2 joysticks.

It's almost 100% playable, the only annoying thing is that lag prevents you
from turning. When I face a complex and/or deep part of the track, it lags,
and i can't turn. So it's impossible to really race against each other in "On
the beach".

Has it something to do with the time delta ? (refresh intervals)

My neighbor and i had _alot_ of fun in "Race Track" in which the gameplay is
very smooth/perfect. :)

-Pascal
--
Projet MoviXMaker (http://sv.gnu.org/projects/movixmaker)
Projet [e]MoviX[2] (http://movix.sf.net) 
Debian Project (http://www.debian.org)
TuxKart (Wiki (GOTM): http://netpanzer.berlios.de/tuxkart/index.php)

On Sun, 2004-08-08 at 08:06, Pascal Giard wrote:
> Hi there guys,
>   i'm just confirming that tuxkart you can play with 2 joysticks.

Cool, thanks for testing it.

> 
> It's almost 100% playable, the only annoying thing is that lag prevents you
> from turning. When I face a complex and/or deep part of the track, it lags,
> and i can't turn. So it's impossible to really race against each other in "On
> the beach".
> 
> Has it something to do with the time delta ? (refresh intervals)
> 
> My neighbor and i had _alot_ of fun in "Race Track" in which the gameplay is
> very smooth/perfect. :)

Yeah, my computer totally can't cope with "On the beach" either. I've
also noticed the frame rate drops hugely each time a Tux comes into view
- maybe the model is slightly too detailed?

	James

James Gregory <jam...@bt...> writes:

> Yeah, my computer totally can't cope with "On the beach" either.
> I've also noticed the frame rate drops hugely each time a Tux comes
> into view - maybe the model is slightly too detailed?

Tux has around ~3200 triangles, the track itself around ~10000, in
comparism the old geekokart has 870, the old tracks around ~1000.

No idea exactly how much polys todays gfx cards can handle, but due to
the horrible performance on startup I think the gfx card itself isn't
yet the problem, but more the collision code or something like that.

I am also not sure how plib internally works, so pumping the track as
one huge mesh to it might not be a good idea. Some of the other tracks
where broken up in little segments, which might make it easier to cull
these when they are out of the view.

-- 
WWW:      http://pingus.seul.org/~grumbel/ 
JabberID: gr...@ja... 
ICQ:      59461927

Ingo Ruhnke wrote:

> I am also not sure how plib internally works, so pumping the track as
> one huge mesh to it might not be a good idea. Some of the other tracks
> where broken up in little segments, which might make it easier to cull
> these when they are out of the view.

Both collision code and rendering code will benefit IMMENSELY from having
the track broken up into short segments.  Without that, there is no way
to cull by field of view - and the collision detection code will have to
do really costly testing for just about every polygon.

---------------------------- Steve Baker -------------------------
HomeEmail: <sjb...@ai...>    WorkEmail: <sj...@li...>
HomePage : http://www.sjbaker.org
Projects : http://plib.sf.net    http://tuxaqfh.sf.net
            http://tuxkart.sf.net http://prettypoly.sf.net
-----BEGIN GEEK CODE BLOCK-----
GCS d-- s:+ a+ C++++$ UL+++$ P--- L++++$ E--- W+++ N o+ K? w--- !O M-
V-- PS++ PE- Y-- PGP-- t+ 5 X R+++ tv b++ DI++ D G+ e++ h--(-) r+++ y++++
-----END GEEK CODE BLOCK-----

If you read the section on "How Steve Baker builds tracks" here:

http://netpanzer.berlios.de/tuxkart/index.php/Procedure%20For%20Building%20Tracks

...you'll see that the procedure I recommend just 'naturally' makes it
Cull and collision-detect nicely because each 10m to 20m section ends up
being a separate object.

---------------------------- Steve Baker -------------------------
HomeEmail: <sjb...@ai...>    WorkEmail: <sj...@li...>
HomePage : http://www.sjbaker.org
Projects : http://plib.sf.net    http://tuxaqfh.sf.net
            http://tuxkart.sf.net http://prettypoly.sf.net
-----BEGIN GEEK CODE BLOCK-----
GCS d-- s:+ a+ C++++$ UL+++$ P--- L++++$ E--- W+++ N o+ K? w--- !O M-
V-- PS++ PE- Y-- PGP-- t+ 5 X R+++ tv b++ DI++ D G+ e++ h--(-) r+++ y++++
-----END GEEK CODE BLOCK-----

Steve Baker <sjb...@ai...> writes:

> Both collision code and rendering code will benefit IMMENSELY from
> having the track broken up into short segments. Without that, there
> is no way to cull by field of view - and the collision detection
> code will have to do really costly testing for just about every
> polygon.

Tried that now and at least the time it took until the track gets
useable went down to zero (aka. the slow-on-startup problem went
completly away).

When I am in a position where only a small portion of the track is
visible the framerate is also a better, in other section it is however
still unusable slow. Tweaking the Near/Far clipping planes helps here
quite a bit, even so that would be more an ugly workaround than a
solution. wolfman8k mentioned VBOs, which might provide some serious
speedup over vertex arrays which are currently used in plib (correct
me if I am wrong here). However I have no idea how good that speed
boost would be.

Another problem is that breaking the track up into pieces results in
seams all over the track, this might be due to the way the AC3d
exporter works which might result in some nummerical errors, however
havn't checked the details.

-- 
WWW:      http://pingus.seul.org/~grumbel/ 
JabberID: gr...@ja... 
ICQ:      59461927

Ingo Ruhnke wrote:

> When I am in a position where only a small portion of the track is
> visible the framerate is also a better, in other section it is however
> still unusable slow. Tweaking the Near/Far clipping planes helps here
> quite a bit, even so that would be more an ugly workaround than a
> solution. wolfman8k mentioned VBOs, which might provide some serious
> speedup over vertex arrays which are currently used in plib (correct
> me if I am wrong here). However I have no idea how good that speed
> boost would be.

It varies *tremendously* from one graphics card to another and even
from one driver version to another.   I'd be suprised if that turned
out to help much.

We could create ssgRangeSelector (aka level-of-detail) nodes to toss
out track segments that are further than some specific range.

> Another problem is that breaking the track up into pieces results in
> seams all over the track, this might be due to the way the AC3d
> exporter works which might result in some nummerical errors, however
> havn't checked the details.

I can't explain that problem - it doesn't happen with files created
within AC3D.  Since AC3D files are ASCII text, it should be really
easy to figure it out with a two polygon test model.

---------------------------- Steve Baker -------------------------
HomeEmail: <sjb...@ai...>    WorkEmail: <sj...@li...>
HomePage : http://www.sjbaker.org
Projects : http://plib.sf.net    http://tuxaqfh.sf.net
            http://tuxkart.sf.net http://prettypoly.sf.net
-----BEGIN GEEK CODE BLOCK-----
GCS d-- s:+ a+ C++++$ UL+++$ P--- L++++$ E--- W+++ N o+ K? w--- !O M-
V-- PS++ PE- Y-- PGP-- t+ 5 X R+++ tv b++ DI++ D G+ e++ h--(-) r+++ y++++
-----END GEEK CODE BLOCK-----

For the first time, tuxkart crashed on me.

i built from CVS at 04:04:51 -0400.

here's the output:

[...]
WARNING: Speed too high for collision detect!
WARNING: Nsteps=115, Speed=45.969669!
WARNING: Speed too high for collision detect!
WARNING: Nsteps=116, Speed=46.016544!
WARNING: Speed too high for collision detect!
WARNING: Nsteps=116, Speed=46.051704!
More than 200 hooknodes in database!

than i get back to X, mouse no longer works and i'm in 800x600 instead of
1024x768.

i then use the same old trick, i fire up quake3 then quit, it brings back the
mouse and the screen resolution.

-Pascal
PS: what are hooknodes ?
PS²: i noticed that it doesn't lag during your first lap but then it gets
worst and worst... is it because of skidrows or there's more to it ?
--
Projet MoviXMaker (http://sv.gnu.org/projects/movixmaker)
Projet [e]MoviX[2] (http://movix.sf.net) 
Debian Project (http://www.debian.org)
TuxKart (Wiki (GOTM): http://netpanzer.berlios.de/tuxkart/index.php)

"Pascal Giard" <pa...@gu...> writes:

> i then use the same old trick, i fire up quake3 then quit, it brings
> back the mouse and the screen resolution.

That should be avoidable by doing atexit(SDL_QUIT); or something along
the lines, which forces a clean shutdown even on segfault.

> PS: what are hooknodes ?
> PS=B2: i noticed that it doesn't lag during your first lap but then it ge=
ts
> worst and worst... is it because of skidrows or there's more to it ?

The skidmarks currently just grow and grow all the time, so yep, if
you do a lots of laps you will reach a point where it will get quite
slow, especially with multiple vehicles, this should however change as
soon as we start placing skidmarks only when you skid and not all the
time. Expiring skidmarks might also help, even so this might not
really be necesarry.

--=20
WWW:      http://pingus.seul.org/~grumbel/=20
JabberID: gr...@ja...=20
ICQ:      59461927

On Sat, 2004-08-14 at 12:47 +0200, Ingo Ruhnke wrote:
> The skidmarks currently just grow and grow all the time, so yep, if
> you do a lots of laps you will reach a point where it will get quite
> slow, especially with multiple vehicles, this should however change as
> soon as we start placing skidmarks only when you skid and not all the
> time. Expiring skidmarks might also help, even so this might not
> really be necesarry.

Regardless, I think skidmarks should be optional, and optionally expire.
Don't forget the users with crap hardware - they need a 'back to basics'
graphics mode or a compromise inbetween.
-- 
- Charlie

Charles Goodwin <ch...@ve...>
Online @ www.charlietech.com

Good job! It's now fully playable even when lagging!

Even at 15fps, i get to turn and i can't do incredibly long and high jumps
anymore ;)

The incomplete parts, i notice the more now, are:
- all parts of the track have the same coefficient of friction, i mean, you
can go as fast on grass than on asphalt.
- floating skidrows, i can jump and burn "the air"
- the AI

-Pascal
--
Projet MoviXMaker (http://sv.gnu.org/projects/movixmaker)
Projet [e]MoviX[2] (http://movix.sf.net) 
Debian Project (http://www.debian.org)
TuxKart (Wiki (GOTM): http://netpanzer.berlios.de/tuxkart/index.php)

Em S=E1bado, 14 de Agosto de 2004 15:14, o Charles Goodwin escreveu:
> On Sat, 2004-08-14 at 12:47 +0200, Ingo Ruhnke wrote:
> > The skidmarks currently just grow and grow all the time, so yep, if
> > you do a lots of laps you will reach a point where it will get quite
> > slow, especially with multiple vehicles, this should however change as
> > soon as we start placing skidmarks only when you skid and not all the
> > time. Expiring skidmarks might also help, even so this might not
> > really be necesarry.
>
> Regardless, I think skidmarks should be optional, and optionally expire.
> Don't forget the users with crap hardware - they need a 'back to basics'
> graphics mode or a compromise inbetween.

 What about adding an option to lower the quality of the graphics. Somethin=
g=20
like: Graphical quality: Low/Medium/High.

Cheers,
 Ricardo

=2D-=20
Q:	How many gradual (sorry, that's supposed to be "graduate") students
	does it take to screw in a lightbulb?
A:	"I'm afraid we don't know, but make my stipend tax-free, give my
	advisor a $30,000 grant of the taxpayer's money, and I'm sure he
	can tell me how to do the shit work for him so he can take the
	credit for answering this incredibly vital question."

On Sat, 14 Aug 2004 16:01:53 +0100, Ricardo Cruz <ri...@ae...> wrote:
> Em S=E1bado, 14 de Agosto de 2004 15:14, o Charles Goodwin escreveu:
> > On Sat, 2004-08-14 at 12:47 +0200, Ingo Ruhnke wrote:
> > > The skidmarks currently just grow and grow all the time, so yep, if
> > > you do a lots of laps you will reach a point where it will get quite
> > > slow, especially with multiple vehicles, this should however change a=
s
> > > soon as we start placing skidmarks only when you skid and not all the
> > > time. Expiring skidmarks might also help, even so this might not
> > > really be necesarry.
> >
> > Regardless, I think skidmarks should be optional, and optionally expire=
.
> > Don't forget the users with crap hardware - they need a 'back to basics=
'
> > graphics mode or a compromise inbetween.

the skid makrs should also be more realistic cause you would not skid
on a little turn would you?

>=20
>  What about adding an option to lower the quality of the graphics. Someth=
ing
> like: Graphical quality: Low/Medium/High.
>=20
> Cheers,
>  Ricardo
>=20
> --
> Q:      How many gradual (sorry, that's supposed to be "graduate") studen=
ts
>         does it take to screw in a lightbulb?
> A:      "I'm afraid we don't know, but make my stipend tax-free, give my
>         advisor a $30,000 grant of the taxpayer's money, and I'm sure he
>         can tell me how to do the shit work for him so he can take the
>         credit for answering this incredibly vital question."
>=20
>=20
>=20
>=20
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Tuxkart-devel mailing list
> Tux...@li...
> https://lists.sourceforge.net/lists/listinfo/tuxkart-devel
>

Em S=E1bado, 21 de Agosto de 2004 00:53, o Caleb Sawtell escreveu:
> On Sat, 14 Aug 2004 16:01:53 +0100, Ricardo Cruz <ri...@ae...> wrote:
> > Em S=E1bado, 14 de Agosto de 2004 15:14, o Charles Goodwin escreveu:
> > > On Sat, 2004-08-14 at 12:47 +0200, Ingo Ruhnke wrote:
> > > > The skidmarks currently just grow and grow all the time, so yep, if
> > > > you do a lots of laps you will reach a point where it will get quite
> > > > slow, especially with multiple vehicles, this should however change
> > > > as soon as we start placing skidmarks only when you skid and not all
> > > > the time. Expiring skidmarks might also help, even so this might not
> > > > really be necesarry.
> > >
> > > Regardless, I think skidmarks should be optional, and optionally
> > > expire. Don't forget the users with crap hardware - they need a 'back
> > > to basics' graphics mode or a compromise inbetween.
>
> the skid makrs should also be more realistic cause you would not skid
> on a little turn would you?
>
 They should also be graphically more realistic. They should look more like=
 a=20
piece of burned wheel, rather than just a transparent color. Using a textur=
e=20
might help a lot.
 Also make it starting right below the wheel. Just making the marks startin=
g a=20
few pixels before will do the trick.

Cheers,
 Ricardo

> >  What about adding an option to lower the quality of the graphics.
> > Something like: Graphical quality: Low/Medium/High.
> >
> > Cheers,
> >  Ricardo
> >
> > --
> > Q:      How many gradual (sorry, that's supposed to be "graduate")
> > students does it take to screw in a lightbulb?
> > A:      "I'm afraid we don't know, but make my stipend tax-free, give my
> >         advisor a $30,000 grant of the taxpayer's money, and I'm sure he
> >         can tell me how to do the shit work for him so he can take the
> >         credit for answering this incredibly vital question."
> >
> >
> >
> >
> > -------------------------------------------------------
> > SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> > 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> > Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> > http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> > _______________________________________________
> > Tuxkart-devel mailing list
> > Tux...@li...
> > https://lists.sourceforge.net/lists/listinfo/tuxkart-devel
>
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Tuxkart-devel mailing list
> Tux...@li...
> https://lists.sourceforge.net/lists/listinfo/tuxkart-devel

=2D-=20
You own a dog, but you can only feed a cat.