IrrPhysx 0.2 - Nvidia Physx 2.8.1 wrapper
Thanks JP,
I'm writing my third year university project using irrlicht and the day I look for a physics engine to integrate you release this - perfect timing!
Do you have a SVN server with the code in, because there are some features that I require which you have yet to implement as far as I can see (haven't gone through the whole code yet) which I plan to implement. I would like to create patch files based of your code so keep downloading your newest version and readding my changes...
Also I could contribute the patch files if you wanted...
I'm writing my third year university project using irrlicht and the day I look for a physics engine to integrate you release this - perfect timing!
Do you have a SVN server with the code in, because there are some features that I require which you have yet to implement as far as I can see (haven't gone through the whole code yet) which I plan to implement. I would like to create patch files based of your code so keep downloading your newest version and readding my changes...
Also I could contribute the patch files if you wanted...
When I was trying to add irrPhysX to my project Visual Studio 2005 kept reporting the following error on compile.
1>fatal error C1007: unrecognized flag '-typedil' in 'p2'
After much searching on the internet I found nothing useful reguarding the error, so I recompiled irrPhysX using VS2005 instead of 2008 and used that lib. This fixed the issue.
So just a heads up for anybody having a similar error.
1>fatal error C1007: unrecognized flag '-typedil' in 'p2'
After much searching on the internet I found nothing useful reguarding the error, so I recompiled irrPhysX using VS2005 instead of 2008 and used that lib. This fixed the issue.
So just a heads up for anybody having a similar error.
Taonas, yeh I'm pretty sure that just in general VS2008 DLL's wont work with projects being compiled by VS2005 and vice versa. I had a similar problem trying to use boost::python with python 2.6 in my VS2005 project. Python 2.6 and later are all compiled in VS2008 on windows, so my .pyd file compiled in VS2005 would not load. GOing back to python 2.5.2 (compiled in VS2005) fixed it.
My irrlicht-based projects have gone underground for now, but if you want, check out my webcomic instead! http://brokenboomerang.net
Yep if people have weird problems and are using a different compiler then it's a good idea to recompile the library. It's a bit of an annoyance, and maybe there's some solution... Irrlicht works fine for everyone doesn't it? Maybe that's because it's a DLL or something... Whereas IrrPhysx is just a lib.
i've added convex mesh, triangle mesh, tearable cloth and capsule shape support for the next release which should be coming sometime next week.
i've got a short list of things which are really necessary for developing games with physics and they include:
joints,
ragdolls,
raycasting
character controller
and then there's things like soft body and visual debugger and no doubt more things which i'll be supporting fairly soon too hopefully.
basically anything you see in the Physx samples i'll eventually wrap up and support so that includes things like AI sensors and force fields.
i've got a short list of things which are really necessary for developing games with physics and they include:
joints,
ragdolls,
raycasting
character controller
and then there's things like soft body and visual debugger and no doubt more things which i'll be supporting fairly soon too hopefully.
basically anything you see in the Physx samples i'll eventually wrap up and support so that includes things like AI sensors and force fields.
http://developer.nvidia.com/object/physx_features.html
Very cool. This looks like a perfect match for what I'm working with, nice work JP.![Smile :)](./images/smilies/icon_smile.gif)
Very cool. This looks like a perfect match for what I'm working with, nice work JP.
![Smile :)](./images/smilies/icon_smile.gif)
-
- Posts: 153
- Joined: Mon Mar 03, 2008 8:42 am
- Location: Suceava - Romania
- Contact:
Very cool.
I'm looking forward to the character controller support. Are you planning vehicle support (does PhysX have vehicles)?
I have nvidia's latest drivers enabling hardware PhysX support on my graphics card. It's tricky to determine how much benefit I get from that as Irrlicht doesn't have instancing support for objects - the cpu cost of handling and rendering 1000's of objects becomes a bottleneck.
BTW Is it just me or do objects seem to fall slower than you'd expect with Earth gravity?
![Cool 8)](./images/smilies/icon_cool.gif)
I'm looking forward to the character controller support. Are you planning vehicle support (does PhysX have vehicles)?
I have nvidia's latest drivers enabling hardware PhysX support on my graphics card. It's tricky to determine how much benefit I get from that as Irrlicht doesn't have instancing support for objects - the cpu cost of handling and rendering 1000's of objects becomes a bottleneck.
BTW Is it just me or do objects seem to fall slower than you'd expect with Earth gravity?
Irrlicht Demos: http://irrlicht.sourceforge.net/forum/viewtopic.php?f=6&t=45781
PhysX has a simple raycast vehicle controller (as does Bullet), should be pretty interesting to play with if and when JP decides to add it to irrPhysX ![Razz :P](./images/smilies/icon_razz.gif)
Its not just you, using real gravity values doesn't seem to have the same effect in PhysX, not really sure why :\
setting really high values for mass/desnsities also similarly doesn't actually change much either, so it seems to be a niggle with the solver they use.
Either up the gravity value (2x, 4x, 10x), or use use a small range of masses (1-10) to get the desired results.
bizzare (not to mention annoying)![Rolling Eyes :roll:](./images/smilies/icon_rolleyes.gif)
![Razz :P](./images/smilies/icon_razz.gif)
Its not just you, using real gravity values doesn't seem to have the same effect in PhysX, not really sure why :\
setting really high values for mass/desnsities also similarly doesn't actually change much either, so it seems to be a niggle with the solver they use.
Either up the gravity value (2x, 4x, 10x), or use use a small range of masses (1-10) to get the desired results.
bizzare (not to mention annoying)
![Rolling Eyes :roll:](./images/smilies/icon_rolleyes.gif)
Really very good work.
Just a question JP, do you know the package 'Trainingprogram for PhysX'.
This package contains more than 80 examples and tutotials for PhysX, which describes really all types of physical possible with PhysX.
This is really a excellent tools to their function of PhysX. I unfortunately failed to find the link on the site of NVIDIA.
somes screenshots:
![Image](http://www.portail-irrlicht.fr/_upload/images/physXTraining.jpg)
Just a question JP, do you know the package 'Trainingprogram for PhysX'.
This package contains more than 80 examples and tutotials for PhysX, which describes really all types of physical possible with PhysX.
This is really a excellent tools to their function of PhysX. I unfortunately failed to find the link on the site of NVIDIA.
somes screenshots:
![Image](http://www.portail-irrlicht.fr/_upload/images/physXTraining.jpg)
Strength and wisdom.
Admin of the http://www.irrlicht.fr and dad of N3xtD.
Sorry for my poor English.
Admin of the http://www.irrlicht.fr and dad of N3xtD.
Sorry for my poor English.
tmyke, the training programme thing used to be included with the PhysX SDK whilst it was still Ageia's, but ever since nvidia took over, they've scrapped it all for some reason.
I think the last version to include them was v2.7.2, not sure if nvidia even bothers keeping those (really) old versions anymore...
seems to be ~33 MB (the entire SDK including the training programme docs, code and binaries);
I'll upload them for you (and anyone else) sometime later today, if you can't find the original (Ageia) ones that is.
keep in mind that a lot of changes have been made since nvidia took over, and some stuff may actually now not be compatible anymore (pmaps for one... or am I getting confused with Blender again?
)
EDIT:
you can get it here from the PhysX Archives page; direct download link:
http://developer.download.nvidia.com/Ph ... K_Core.exe
I think the last version to include them was v2.7.2, not sure if nvidia even bothers keeping those (really) old versions anymore...
seems to be ~33 MB (the entire SDK including the training programme docs, code and binaries);
I'll upload them for you (and anyone else) sometime later today, if you can't find the original (Ageia) ones that is.
keep in mind that a lot of changes have been made since nvidia took over, and some stuff may actually now not be compatible anymore (pmaps for one... or am I getting confused with Blender again?
![Embarassed :oops:](./images/smilies/icon_redface.gif)
EDIT:
you can get it here from the PhysX Archives page; direct download link:
http://developer.download.nvidia.com/Ph ... K_Core.exe