2605 replies to this topic

[Adam]

Hmmm, Coderchris hasn't been on here for several days now, either he's working on the physics patch or this will be like Windows Vista (delays and more delays)

[jordanware]

Hmmm, Coderchris hasn't been on here for several days now, either he's working on the physics patch or this will be like Windows Vista (delays and more delays)

<{POST_SNAPBACK}>

Shizzastle, you're right. I was actualy gonna compare this to ."a physics engine called GMphysics..." But I'm talking about it. lol, stupid me.

[Metroid48]

Hey, how can I do a top-down car? Like, I'm not sure how to make acceleration and turning to work.

Also, I tried this and though the examples worked fine, the game froze when I tried making top-down cars.

[Adam]

Hmmm, Coderchris hasn't been on here for several days now, either he's working on the physics patch or this will be like Windows Vista (delays and more delays)

<{POST_SNAPBACK}>

Shizzastle, you're right. I was actualy gonna compare this to ."a physics engine called GMphysics..." But I'm talking about it. lol, stupid me.

<{POST_SNAPBACK}>

I'm practically thinking I could quit Gamemaker all together if I never see the patch materialize, I need the save/load function for my games and once you make a game with physics you almost never want to go back to GM without physics.

If I do quit Gamemaker i'll probably move to Blender, they have a new graphics engine coming for the game engine and they do have a nice physics engine, a few bugs left to work out but they're being fixed. (Blender's free by the way (www.blender.org))

[coderchris]

Iv honetsly been trying to get the save / load working, but it just doesnt want to work for me. Im not sure if ill be able to get it working or not, the problems seem to be in the loading. Theres something with physX that doesnt like creating tons of bodys at once, which is what the loading function does.

It is possible to make your own saving loading code in gamemaker thou it takes a bit of time. Basically what you do is use all the get_ functions to get the state of each body that you want to save. Then you write all these to a file and when you load it, open the file. (Your bodys should be created already when the room is loaded). Then you set the properties of each body using the data from the file.
This will actually work just fine for games that dont need to save hundreds of bodies. For example if you just have a platformer then you can just save the state of the main character

[Adam]

Mind you show us an example of this Coderchris?

When it comes to writing to files I haven't really touched that part of Gamemaker as i've only used the built in save/load functions.

It doesn't have to be fancy, maybe a character and a couple of boxes and static objects, saving variables with that method would be nice to show too.

Though if you find a way to load a bunch of bodies with PhysX that would be nice too
If you can't but you show us this alternative way that's fine I guess too.

Edited by Adam, 04 September 2006 - 05:04 PM.

[Potnop]

U actually might wanna get into makjing external files with GM. It's VERY useful soemtimes. One reason usually having persistent objects between rooms causes bugs, but with external files U can fix that. U can save the state of objects that need to be brought over to the next room the room itself would place them in their needed positions.

Or U can save game settings. Or even make custom levels and stuff. It would B very easy to make your own save system with external files. Just B sure no1 can hack it so either encrypt it or make sure the save's md5 hash matches or soemthing. Whatever U might want if U ever get around to doing that.

[coderchris]

Ill make a nice example of a room full of boxes that you can save/load since I cant get the function itself working.

[MatrixQuare]

Hey Coderchris, what about the "wall hugging" bug that me and Adam described a page or so back? And change of getting that fixed for V4? If not, would you still have V3.5 lying around that I you could possibly upload and PM me the link for? I need to get rid of the "wall hugging" syndrom, otherwise my game will be pretty darn boring (it involves a ball bouncing around the wall, but once in a while, it'll incorrect start rolling alon the wall instead of bouncing off realistically).

[jordanware]

Ill make a nice example of a room full of boxes that you can save/load since I cant get the function itself working.

<{POST_SNAPBACK}>

Well, having no save/load function is a bit of a bummer. But working hard to get nothing is much better then not working at all.....I guess...

So what else was planned for the patch...bug fixes?

[aemynzc]

how did he make that cool car in that pic! I demand rights to know!...[SIZE=1]please...

[Hach-Que]

Need Help : I have a game, where you roll around as a red ball, and I am trying to create terrain for it (by using 128x128 sprites), however, create_body_auto() doesn't seem to work properly on any terrain that isn't a block. How can I fix this?

LINK (doesn't contain physics DLLS)

[stickman_638]

So, chris you want to start something like newton playground?

[MrYano]

You....are....a....god....
Nice job btw
I was using the ackack engine for my Lymphs game, but now.... well, I just don't know.. I think you might take the cake. Who should I acredit this engine to anyway? Just coderchris or another name? Anyway, pm me or just post =P.
-Ryan

Edited by MrYano, 12 September 2006 - 10:38 PM.

[jordanware]

Coderchris? Nata update let alone reply to keep us happy?

By the way, chris, are you still working on the game with GMphysics?



And to the people at the GMPC:

I might be resigning as an admin soon....maybe. If I do, "mrgibblet" (mod) will take my place.

[yo1dog]

In the create event of an object I have:

b=create_body(x,y,1,SHAPE_ELLIPSE,23,23)

but when I change the width and height to be less than 23 or greater than 15, it shows " unexpected game error", why? Also the "circles" seem to be more like ovals. There off balance and not centered.


HERE is the gm6 file.

BTW, I do have the sprite's orgin at 16x16 and the sprite's width and height is 32x32.


EDIT: I will use SHAPE_SPHERE but that problem still needs to be fixed.


Also How does the set_hinge_limit work? I tried changeing it so that the ragdoll's arm could only be striaght in front of him but it doesnt seem to do anything.

Edited by yo1dog, 14 September 2006 - 10:48 AM.

[RiChArD_fLoOd]

any chance of a c++ release?

[coderchris]

sorry for not being around lately and for all the delays; have been really busy with things and school just started up so I have quite a bit of work to do but I promise ill get this stuff done. I just need to put aside a few hours so I can work on the save/ load example and the c++ one. The game I was working on is currently on hold until I can get more time to finish it.

One of the things iv been busy with is experimenting with v1 of gmphysics, the really old version that was first released a loooong time ago. Iv been studying alot more recently and have come up with some ways to make a faster, more stable, and more accurate simulation than PhysX can provide (in 2d anyway )
Iv come up with a broadphase collision detection algorithm that detects all possible collisions in O(1) time. This coupled with seperate threads for the dynamics and collision detection make it extremely fast, able to simulate literally thousands of colliding bodies with 30+ fps (in c++). If there arent too many pairs of colliding bodies and some are sleeping, you could easily get up to 10,000 without slowing down the simulation due to my new broadphase collision method.
Iv also come across a technical paper that describes the simulation of granular material (sand, small rocks, ect..) Iv implemented this sucessfully and im also adapting it for use with fluid. What this means is you wont have particles to represent fluid but rather whole polygons that group the particles together and allow for much greater performance and visual appearance as well as things like correct viscocity and fluid flow. With this approach to fluid, water bodies and fluid become one in the same and you could for example dip a bucket into a body of water and it will accurately fill up and stay there as you lift the bucket out of the water, allowing you to pour and manipulate it as it should.

Iv also been experimenting with GPGPU processing (i.e. processing physics on the GPU) I havent been able to port all aspects of it to the gpu yet but from what im seeing, it will be possible to simulate a rediculous number of bodies (possibly over 20,000 active bodies!) without slowing down the CPU at all whatsoever, leaving you with crisp framerates and more processing time to do things like AI. What I would like to do is something similar to what directx does: have two modes, referance and hardware. If there is hardware that would benefit the physics simulation, it will automaticall use it (i.e use the GPU). If not it will use reference mode which implements everything in the CPU.

Another advantage to re-writing v1 is that you will be able to use it in fully commercial projects because all the code will be my own and I have no restrictions on its usage (other than the prefered credit). I will also keep 99% of the functions, parameters and constants the same so that you wont have to do much to make it compatable.

I now realize that although top down mode could be usefull, the way I have implemented it doesnt allow for many things like vehicles and rolling spheres. Because of this, in the next version Ill be seperating top down and regular mode completely so that things like vehicles and rolling spheres could easily be added

So sometime this year or early next year you might be seeing GmPhysics v6 featuring things i just explained.
Hopefully all that goodness keeps your interests in this project flowing
I have yet to make an official website for it, but eventually, my plan is to create 3d physics middleware such as havok and physx for use in commercial games and applications

Edited by coderchris, 18 September 2006 - 07:33 PM.

[stickman_638]

sorry for not being around lately and for all the delays; have been really busy with things and school just started up so I have quite a bit of work to do but I promise ill get this stuff done. I just need to put aside a few hours so I can work on the save/ load example and the c++ one. The game I was working on is currently on hold until I can get more time to finish it.

One of the things iv been busy with is experimenting with  v1 of gmphysics, the really old version that was first released a loooong time ago. Iv been studying alot more recently and have come up with some ways to make a faster, more stable, and more accurate simulation than PhysX can provide (in 2d anyway )
Iv come up with a broadphase collision detection algorithm that detects all possible collisions in O(1) time. This coupled with seperate threads for the dynamics and collision detection make it extremely fast, able to simulate literally thousands of colliding bodies with 30+ fps (in c++). If there arent too many pairs of colliding bodies and some are sleeping, you could easily get up to 10,000 without slowing down the simulation due to my new broadphase collision method.
Iv also come across a technical paper that describes the simulation of granular material (sand, small rocks, ect..) Iv implemented this sucessfully and im also adapting it for use with fluid. What this means is you wont have particles to represent fluid but rather whole polygons that group the particles together and allow for much greater performance and visual appearance as well as things like correct viscocity and fluid flow. With this approach to fluid, water bodies and fluid become one in the same and you could for example dip a bucket into a body of water and it will accurately fill up and stay there as you lift the bucket out of the water, allowing you to pour and manipulate it as it should.

Iv also been experimenting with GPGPU processing (i.e. processing physics on the GPU) I havent been able to port all aspects of it to the gpu yet but from what im seeing, it will be possible to simulate a rediculous number of bodies (possibly over 20,000 active bodies!) without slowing down the CPU at all whatsoever, leaving you with crisp framerates and more processing time to do things like AI. What I would like to do is something similar to what directx does: have two modes, referance and hardware. If there is hardware that would benefit the physics simulation, it will automaticall use it (i.e use the GPU). If not it will use reference mode which implements everything in the CPU.

Another advantage to re-writing v1 is that you will be able to use it in fully commercial projects because all the code will be my own and I have no restrictions on its usage (other than the prefered credit). I will also keep 99% of the functions, parameters and constants the same so that you wont have to do much to make it compatable.

I now realize that although top down mode could be usefull, the way I have implemented it doesnt allow for many things like vehicles and rolling spheres. Because of this, in the next version Ill be seperating top down and regular mode completely so that things like vehicles and rolling spheres could easily be added

So sometime this year or early next year you might be seeing GmPhysics v6 featuring things i just explained.
Hopefully all that goodness keeps your interests in this project flowing
I have yet to make an official website for it, but eventually, my plan is to create 3d physics middleware such as havok and physx for use in commercial games and applications

<{POST_SNAPBACK}>

AHHH!!! OMGOMGOMGOMGMOMGOGMOGMGOMGOMGM OMFG!!!! NEED DROOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOL!!!!!
CANT WAIT!! YOU ARE THE BEST OF THE BEST! I CANOT WAIT NOW YOU ARE GOING TO HAVE ME AWAKE FOR MONTHS WAITING....(sorry for cap's but i'm not talking it off im to exited!)

[jordanware]

Wow, seems like you've been spending alot of time on this (and, typing up that reply ).
I really like what I'm hearing....er, reading....sounds like a GMphysics Revolution!
lol.

But seriously, I liked everything you said. Just keep at it and I'm sure this'll be great

And good luck with school