Speed !!

Blitz3D Forums/Blitz3D Userlibs/Speed !!

Pete Carter(Posted 2005) [#1]
So out of JV-ode and EasyTOK which one gets the highest FPS ive tryed both demos and Jv-ode seems much quicker but is this just on my machine? also which is easyer to get working?


OJay(Posted 2005) [#2]
easytok is easier to use (maybe) so if you want quick and simple physics, this is yours. but since easytok uses the tokamak physics engine, their will be some slight issues, when moving to more complex simulations! also, the development of tokamak (and especially the wrapper, which you need too) is slow as hell (2 or 3 updates this year?). so keep this in mind.

speaking about jv-ode: it uses the opensource engine ODE, so there're quite often pretty large updates and bugfixes. ODE is used in a wide variety of games, some even professional (e.g. expand rally). but there's (not yet) a simple library to setup ode physics, so this is a bit more complicated (although viper and the other guys are giving excellent support and help), but there are some very helpful samples you get with it (terrain, cars and all that fency stuff).

to sum it up:

easytok: simple to use, but you have to live with the tokamak bugs and the development-freeze. no support from the tokamak-dev's at all.

jv-ode: more complicated, but rocksolid physics engine with all features you ever would need for games. excellent support and regular updates.

the prizing is nearly the same for both.


you decide. :)


big10p(Posted 2005) [#3]
Thanks for the info OJay - I've been meaning to ask a similar question to Pete's, myself.

Mmm. I'm liking the sound of this jv-ode thingie more and more. I think I'll check it out more closely when I finish this damn app I've been working on forever, and can get back to actually writing a game... maybe. :)


Shifty Geezer(Posted 2005) [#4]
I produced an engine in Tokamak. When I switched to JV-ODE I was getting about a 5x speed improvement, plus LOTS of extra features like joints. ODE takes a bit of getting used to, but it's versatile and quick, and AFAIK the only suitable physics engine to consider for Blitz until perhaps someone produces a Newton wrapper and that can be considered (dunno how good it is though!).


Damien Sturdy(Posted 2005) [#5]
I remember the old days- when ODE was slower than Toka.

Tell you what, it toka is the slower system now, i need to change! ;)


KuRiX(Posted 2005) [#6]
Cygnus... are you using toka now? i thought you were still with ode...


Vorderman(Posted 2005) [#7]
I'm using Tokamak for Stunt Race Xtreme, so I can safely say it works really well for some stuff. I've achieved a much better car physics model using Tokamak than ODE, despite ODE having an inbuilt car-wheel joint - ODE's friction seems to very iffy - all the car demos I've tried, and all the tests I've done end up with a vehicle that is very twitchy around the rear end - it either grips or it slides, and holding a drift is impossible.

I spent a year or two playing with ODE and car physics and could never get the result I wanted - with Tokamak I have the SRX cars drifting around beautifully in exactly the right way, so I much prefer Tokamak to ODE at this point.


Shifty Geezer(Posted 2005) [#8]
Vorderman : Have you checked my ODE friction research in the ODE thread? I found a formula for defining friction between a 0 to 100% range and it works very well. eg. Apply backspin to a 50% friction sphere and give it a push, it travels forwards, sliding, gradually gains traction, and returns on the roll, just as it should.


Pete Carter(Posted 2005) [#9]
Thanks for your help! i think jv-ode seems to be the best for blitz at the mo. plus i remembered that my mate ricky uses it with Pacemaker and im sure he can give me some tips if i get stuck :OP


Vorderman(Posted 2005) [#10]
Shifty Geezer : no, I haven't seen that, but I'll have a look at it, thanks.

Have you looked at the Pacejka formula to simulate a tyre? - http://www.racer.nl/reference/pacejka.htm


Shifty Geezer(Posted 2005) [#11]
No, and thankfully I'm not writing anything simulating cars so I don't need to worry about getting all that working!


KuRiX(Posted 2005) [#12]
You can get cars working very well with ode if you make lot of tuning (see the last jv-ode thread for my post).

I will try to upload a free demo later to show a good car working with ode!


Damien Sturdy(Posted 2005) [#13]
Hi Kurix.


Im using ODE for the racer projects and Toka for another networking mp game. At the time i started, Toka was faster than the available ODE wrapper :)

I still much prefer ODE :D


I spent a year or two playing with ODE and car physics and could never get the result I wanted - with Tokamak I have the SRX cars drifting around beautifully in exactly the right way, so I much prefer Tokamak to ODE at this point.


heh, You can drift in my racer. thats using YOUR wrapper ;)


KuRiX(Posted 2005) [#14]
Where is your racer, i want to play it now!!!


Vorderman(Posted 2005) [#15]
No, and thankfully I'm not writing anything simulating cars so I don't need to worry about getting all that working!

Yes, it is a bit deep isn't it?

I will try to upload a free demo later to show a good car working with ode!


Cool - I'd like to see your demo.

heh, You can drift in my racer. thats using YOUR wrapper ;)

At the time i started, Toka was faster than the available ODE wrapper :)

Using my SLLLOOOOWWW wrapper eh? :) Anyway, pipe down, Tokamak is far better.


Sweenie(Posted 2005) [#16]
I guess you never thought I'd say this but... Newton rulez!

Ok, it may be slower than ODE n Tokamak, but when it comes to precision, features and stability ODE and Tokamak has nothing compared to Newton.


KuRiX(Posted 2005) [#17]
A small video of Fast Racing in Ode:

www.lcuriel.arrakis.es/kracing1.avi

Hope you like it.


Shifty Geezer(Posted 2005) [#18]
Sweenie : Are you using Newton with Blitz? It's something I'd be keen to have a look at as I'd love to be able to use Newtonian values instead of arbitary 1's and 10's that don't really mean anything.


Sweenie(Posted 2005) [#19]
Since Newton makes heavily use of callbacks, using it in Blitz3D is a bit difficult.
I've been using it alot in Bmax & C++ though.


Damien Sturdy(Posted 2005) [#20]
I'll see what i can do Kurix. Maybe later today!


Damien Sturdy(Posted 2005) [#21]
Or maybe not. I'll do it when we get the time to make a build!