new Physx

new Physx

Postby ikam » Wednesday, 25.June 2014, 20:12

Yop

Here it is, the first prototype is open to tests :)

I kept most of operators parameters but I completely rewrote the code architecture behind and so the operators class and behavior. Therefore nothing is compatible with the old system and even if doesn't seem to change on surface, all behind is new and so bugs can be found where before they was not... Why we need to test all...

This new architecture was necessary to fix most of problems of the old system and to create a more modular code to implant next features (like physx particles)...

so in brief what's new :

- all physx operators works entirely in graph mode. This means that all wz classic transformations are supported and correctly cumulated (transform, add, multiply) both for colliders and actors.

- all wz4Render can be binded to physic (in old system only meshes), so every special render can be used on rigidbody input (sprites, marchingcubes, ribbons, etc...), just create one or multiple colliders around the render node shape and go on...

- Entire physx node can be transformed and multiplied (with yellow operators under physx) for example to replicate the physx actors and simulation.

- Debris actor, to generate shards from a chunked mesh.

it remains to add a correct timestep simulation and the fromvertex build mode, but despite that all should be like the old one.

here the proto exe. Alternative is to get 'physx3' branch on github.
You do not have the required permissions to view the files attached to this post.
ikam
Operator
Operator
 
Posts: 911
Joined: Friday, 14.October 2011, 13:00
Location: France

Re: new Physx

Postby ikam » Wednesday, 25.June 2014, 20:34

a example file
You do not have the required permissions to view the files attached to this post.
ikam
Operator
Operator
 
Posts: 911
Joined: Friday, 14.October 2011, 13:00
Location: France

Re: new Physx

Postby Skinnytorus » Wednesday, 25.June 2014, 21:00

Yeah!!! You are awesome, man! =D> =D> =D>
Your power has reached unparallelled heights! Almighty Megatron is promoting you to the rank of the Grand Master Code Guru (GMCG)! =)
I will take a closer look tomorrow. Had a busy day... too tired but instantly cheered up by your good news! Many thanks as always.
Will get back to you tomorrow.
Skinnytorus
Operator
Operator
 
Posts: 1300
Joined: Monday, 06.February 2012, 17:46

Re: new Physx

Postby CybeREX » Thursday, 26.June 2014, 06:14

I have played with old phisics, but have much point that I don't like in how phisx working, I try to recreate my scene with this one new build and write feedback about tests.
PS. ikam, good coding as always =)
CybeREX
Demomaker
Demomaker
 
Posts: 154
Joined: Thursday, 03.April 2014, 07:44

Re: new Physx

Postby ikam » Thursday, 26.June 2014, 06:43

Skinnytorus : yeaaa my vauddo magic and my chiken sacrifice gave me a better understanding of werkkzeug internal :)

CybeREX : yes, in old system, it was difficult to build actors pose because of bad transformations cumulations, now it is fixed and it should be easier. I will also try to create a simpler joints builder. Don't hesiste to share your comments if you have ideas or suggestion about physx works.
ikam
Operator
Operator
 
Posts: 911
Joined: Friday, 14.October 2011, 13:00
Location: France

Re: new Physx

Postby CybeREX » Thursday, 26.June 2014, 07:52

Something wrong? I have 32 bit winXP here system and when I unpack this new archive with wz4 I have message when launch it
wz4 is not win 32 file...
CybeREX
Demomaker
Demomaker
 
Posts: 154
Joined: Thursday, 03.April 2014, 07:44

Re: new Physx

Postby Skinnytorus » Thursday, 26.June 2014, 08:52

CybeREX
Maybe it's a 64 bit release? You'd better compile the physx3 branch from github.

ikam
A simpler joints builder would be cool. Maybe use stacking order to simplify work with joint IDs? Parents/hooks: vertically - from top to bottom, siblings: horizontally - from left to right? Multiply and Add operators will then create sibling type object relationship... What do you think?
Skinnytorus
Operator
Operator
 
Posts: 1300
Joined: Monday, 06.February 2012, 17:46

Re: new Physx

Postby ikam » Thursday, 26.June 2014, 09:05

CybeREX : yes you're right, try this new file.

Skinnytorus : yes using stacking order should be possible now to build joints. I think to abandon the join ID system and adopt a join operator than simply take two rigidbody on input and join them according stacking hierarchy. Multiply will multiply object + joint. for chained joint, simply joint the same rigidbody but at 2 different joint pose.
You do not have the required permissions to view the files attached to this post.
ikam
Operator
Operator
 
Posts: 911
Joined: Friday, 14.October 2011, 13:00
Location: France

Re: new Physx

Postby Skinnytorus » Thursday, 26.June 2014, 09:23

ikam
Yep, joint ID system is crying out for visual approach. I like your ideas, but they would be best seen in action. I hope you'll come up with a proto soon.
Skinnytorus
Operator
Operator
 
Posts: 1300
Joined: Monday, 06.February 2012, 17:46

Re: new Physx

Postby CybeREX » Thursday, 26.June 2014, 10:02

Thx ikam, now wz4 work in 32 bit =)

Here is first example, look in both collider and mesh. I use same y=1.5000 but two branch is different.
You do not have the required permissions to view the files attached to this post.
CybeREX
Demomaker
Demomaker
 
Posts: 154
Joined: Thursday, 03.April 2014, 07:44

Re: new Physx

Postby ikam » Thursday, 26.June 2014, 13:26

yes, it was a bug. it is fixed now on "physx3" github branch.
thanks ;)

(ps: if you use the physx3 branch and updates, when building an example remind you to re-create all operators at each time or start a new file, all previous example that contain physx operators may have change and result unwanted bugs (even copy /paste of operators from different files will fail), for example the first example that I send on top of this thread is no more working correctly with lastest update in physx3 branch. It's because, during dev, I often modify parameters orders (or add and delete others) that result incompatibles operators between updates.)
ikam
Operator
Operator
 
Posts: 911
Joined: Friday, 14.October 2011, 13:00
Location: France

Re: new Physx

Postby CybeREX » Thursday, 26.June 2014, 14:27

mmm.... I just know about this branch
https://github.com/wzman/werkkzeug4CE
where I can find physX?))

PS. I play a little with new build (I use ikam 32 bit build)
and now, it's look more friendly, I try to create some scene and share with results.
CybeREX
Demomaker
Demomaker
 
Posts: 154
Joined: Thursday, 03.April 2014, 07:44

Re: new Physx

Postby ikam » Thursday, 26.June 2014, 14:58

It is on the same project but on another branch

are you using the git command line bash tool ? if yes :

git fetch => get updates
git checkout physx3 => switch to physx3 branch
(and compile wz)

with git gui you also can switch to a branch but don't remmeber how and depend og which git client you are using
ikam
Operator
Operator
 
Posts: 911
Joined: Friday, 14.October 2011, 13:00
Location: France

Re: new Physx

Postby Skinnytorus » Thursday, 26.June 2014, 18:22

Alternatively: Go there: https://github.com/wzman/werkkzeug4CE/tree/physx3
click 'Download ZIP' and you are done.
Skinnytorus
Operator
Operator
 
Posts: 1300
Joined: Monday, 06.February 2012, 17:46

Re: new Physx

Postby Skinnytorus » Thursday, 26.June 2014, 19:24

Some more bugs here.
Suggestion: Add Empty physx operator for RigidBoby input (no need to go to Render section for Null). Or even better: allow RigidBoby to have just one Collider input if a primitive collider is output (box, sphere, plane)
You do not have the required permissions to view the files attached to this post.
Skinnytorus
Operator
Operator
 
Posts: 1300
Joined: Monday, 06.February 2012, 17:46


Return to Development and contribution



Who is online

Users browsing this forum: No registered users and 1 guest

cron