Techdemo, Step #2 (indoor part)

A forum about everything related to the Cafu World Editor CaWE: mapping, editing, scripting, making GUIs and fonts, using the BSP, PVS and Light compilers.
User avatar
Carsten
Site Admin
Posts:2170
Joined:2004-08-19, 13:46
Location:Germany
Contact:

Post by Carsten » 2006-05-29, 11:03

n321 wrote:About the menu system, I will wait then! :) If it is ok I would like to start on a basic save load system tho! AND SP aspects. (I will just add to the mod code! )
Starting with the SP aspects is certainly a good idea, because I think that I'll have the GUI stuff ready for use soon enough, and about the load/save code, I'll look into it myself as soon as the GUI code is done. It's nontrivial.
As far a my part in the techdemo...I will wait until I get a for sure thing to do...like someone says "hey n321 start on a boat "...then I'll start!
Okay. Thrawn, Kai: Any suggestions?
About the user group! when I go to your link and select tech demo dev team It says it is a closed group and I have no way to join!
Sorry, I just updated your access rights. Please try again.
Best regards,
Carsten
User avatar
Carsten
Site Admin
Posts:2170
Joined:2004-08-19, 13:46
Location:Germany
Contact:

Post by Carsten » 2006-05-29, 11:18

n321 wrote:Ok just got the D3 techdemo map shots and they are great! Whoever is doing the brushwork on this is amazing! I love it....It looked better then D3's maps!
Yep! This is why we have chosen it for the TechDemo...
And were those custom weapons! They were great too!
I don't think so, iirc these were the standard D3 weapons (or maybe from the expansion pack?). Anyways, we will not be able to convert them for Ca3DE.
Will your gui system be ingame like D3 as well!?
Yes.
I hope the PPL will be the same too...bright is ok but I like the dark look...I hope it wont be as light as the old ca3de demo!
The lighting will definitively be brighter than D3 on average. I just played a few hours the first levels of Q4 yesterday, and must say that I don't like the very dark look at all - it's both highly unrealistic (how comes nobody complains though...??) and all the pitch-black areas provide no depth cue at all, so it's easy to get spatially disoriented with just the few lit spots that are sometimes on the screen.
Well if everyone is interested in physics and ca3de's physics are in the mod code I can work on that a bit...quick question tho...How many FPS are you getting in ca3de with ppl and D3 sized maps? because ODE has a speed requirement! if it is 20 or lower we may have to use newton! Let me know if there is a preference...I have put ode into xreal (the old engine i used) and I did not like it as much as I thought i would have!
To be honest, I never deeply looked into physics right now, so I can hardly comment on details like performance. Anyways, the bottleneck of Ca3DE on my computer is currently the per-pixel GPU performance. On other computers it may be the CPU. In general, I think that CPU consumption is much less a problem than one thinks, so I would not pay too much attention on it, especially as subjects like dual-processing and physics accelerators are getting more and more popular. (E.g. why not base the physics system on http://www.ageia.com/developers/api.html ?)
I am also starting SP game code to the DM mod code...I am starting with monsters and AI...wish me luck!
Cool!! :D
Best regards,
Carsten
n321
Posts:28
Joined:2006-05-12, 18:58

Post by n321 » 2006-05-29, 18:57

E.g. why not base the physics system on http://www.ageia.com/developers/api.html ?)
well I could but I was looking for the less costly approach!!! I will link you to some of the ones I was looking at later today!
PhysX looks awsome tho! :)
User avatar
Thrawn
Posts:302
Joined:2004-08-30, 10:38
Location:NRW, Germany
Contact:

Post by Thrawn » 2006-05-29, 22:20

Sry but a physics cart is just crap, I think the approaches to calculate AI on GPU's is much more promising when it comes to it's potential to be spread in public!
Image
User avatar
Carsten
Site Admin
Posts:2170
Joined:2004-08-19, 13:46
Location:Germany
Contact:

Post by Carsten » 2006-05-30, 11:11

Well, the last issue of the c't magazine (or the one before that) had an article that discussed the computation of game physics a) on multi-processor machines, b) on machines with more than one graphics processor where the second is (mis-)used for the physics, and c) dedicated physics accelerators as the one by Ageia.

In my opinion, using a graphics processor for anything else than graphics is just toying around, you also don't use a cellular phone to listen to your latest CDs/mp3s either (even though you probably could use it as a replacement for a HiFi station; the phone probably even has more CPU power these days).
It's nice how imaginative people are when it comes to using GPUs for general-purpose processing, but I really think that these applications are highly limited due to the specialized nature of the GPUs.

The best solution in my opinion is using multiple processors (of the generic CPU type) for this purpose, because they are universally programmable etc.

Dedicated PPUs (Physics Processing Units) may or may not have their purpose - time will tell. If I could choose, I would prefer everyone gets a multi-CPU computer rather than a PPU card, but for me, the same if true for GPU cards. (The ideal computer needs no accelerators for anything... ;) )

To summarize, I think that physics in Ca3DE should be computed on the CPU. If we have the option to have it also run on a PPU, being able to fall-back on the CPU if none is present, I'd prefer that even more.

Therefore, I only suggested to use Ageia because I believe that I've read somewhere that they run their physics on the CPU if no PPU is present (it's probably easy to verify that on their website), so using the Ageia SDK might be a good choice, pending further and more detailed investigations.
Best regards,
Carsten
n321
Posts:28
Joined:2006-05-12, 18:58

Post by n321 » 2006-05-30, 16:24

Ya, I think that if I am to add physics to ca3de I will have to do something a little more simple...If what I do is not advanced enough for what your looking for You can add stuff later! All I am looking for is physics close to what was seen in doom 3! Like vehicles, barrels that move when shot ect...things like advanced ragdoll physics are very possible but not for me! :)
n321
Posts:28
Joined:2006-05-12, 18:58

Post by n321 » 2006-05-30, 22:41

Sorry, I just updated your access rights. Please try again.
ummm still the same!
Sindwiller
Posts:108
Joined:2006-04-14, 21:11
Location:Zürich, Switzerland

Post by Sindwiller » 2006-05-31, 19:25

Thrawn wrote:Sry but a physics cart is just crap, I think the approaches to calculate AI on GPU's is much more promising when it comes to it's potential to be spread in public!
Yawn... decentralisation is here the keyword. Physics and AI on a GPU *are* crap. Because, a GPU can't give information back to the processor. And the capabilities of PhysX are very impressive thought. It shall calculate around 20k physic objects at a time. No more problems about killing Physic in a not rendered node ;)

Wfr, Sindwiller

EDIT:
Ya, I think that if I am to add physics to ca3de I will have to do something a little more simple...If what I do is not advanced enough for what your looking for You can add stuff later! All I am looking for is physics close to what was seen in doom 3! Like vehicles, barrels that move when shot ect...things like advanced ragdoll physics are very possible but not for me!
To add: The Physics in D3 are really crap :roll: These day's, it's not enough to have some boxes and barrels moven around (HL1) or a little bit "realistic" vehicle physics and pseudo ragdoll that kills collision detection at the end-point (UT2004). These days, you need much objects that fly around 'cause of explosions and for shooting on enemys (HL2) or even Physics, who are capable of calculating the destruction of static brushes (Black). Well, you need advanced Physics these days... and some games are showing: Physics are'nt just Eyecandy! (Gmod, HL2DM)
Im Working on:
- Some Linux Bash-Scripts for installing stuff. Dont ask further questions, because i can't explain that more simple ^^
User avatar
Carsten
Site Admin
Posts:2170
Joined:2004-08-19, 13:46
Location:Germany
Contact:

Post by Carsten » 2006-05-31, 21:33

n321 wrote:ummm still the same!
:oopsp: :sorry:
I should read the docs more thoroughly... :book:
You're a member of the TD Dev Team group now, I just made the membership assignment manually.
Best regards,
Carsten
User avatar
Carsten
Site Admin
Posts:2170
Joined:2004-08-19, 13:46
Location:Germany
Contact:

Post by Carsten » 2006-05-31, 21:44

Sindwiller wrote: a GPU can't give information back to the processor.
Of course it can: The return result is the generated image, whose pixels may be considered as a 2D array of values. This array of pixels may be interpreted to encode just about anything. For example, the colors of a normal-map image are interpreted as an array of vectors...
This concept can of course be generalized, allowing GPUs to solve general-purpose problems.
As with traditional parallelization on CPUs, it's easy to see that not all algorithms are easily amenable to implementation on a GPU.
Best regards,
Carsten
Sindwiller
Posts:108
Joined:2006-04-14, 21:11
Location:Zürich, Switzerland

Post by Sindwiller » 2006-05-31, 22:20

Well, it can do that. But how you want to calculate the whole physics within the GPU, when the GPU is also overheated by high detailed geometry, per-pixel lighting and shadowing, massive use of Normal-Mapping/Parallax-Mapping, HDRR and other brainless techniques for increase of reality (maybe SHL, SSS, PRT etc.)? This is senseless and doesnt have future. Even if you know: HavokFX and the ATI-Pendant calculate only small parts of physics like particles etc. The main part is still calculated by the CPU.

When the first GPU was released (Voodo), it revolutionised the world of computer graphics. Now PhysX can do that. Decentralisation is here the keyword.

Wfr, Sindwiller
Im Working on:
- Some Linux Bash-Scripts for installing stuff. Dont ask further questions, because i can't explain that more simple ^^
User avatar
Thrawn
Posts:302
Joined:2004-08-30, 10:38
Location:NRW, Germany
Contact:

Post by Thrawn » 2006-05-31, 23:00

Here is a German news article about an approach by ATI concerning physics calculation by ATI. Click me

I also don't think that this works well but who the hell buys a physics cart? I think cpu is the best approach and Doom 3 are good enough - and if not, at least a very good start ;)
Image
n321
Posts:28
Joined:2006-05-12, 18:58

Post by n321 » 2006-06-01, 03:19

I may try to add more after I get the basic doom 3 style pysics in...still checking best options...I'm thinking Newton just because I did not like ode when I put it in xreal...it was slow and unrealistic! Newton looks better for games! I will add my own custom features to it once its in...so it will be much like doom 3/halo :)
scott
Posts:173
Joined:2004-08-23, 09:11

Post by scott » 2006-06-01, 08:40

Newton is nice, I have been using it at school with the Blitzmax language and Ogre renderer, It has some great features even though its a free system. The continuous collision is damn handy and the geometry destruction is cool as well.

Carsten, with the physics have you thought of implmenting a universal Physics interface so users can easily drop in which ever sdk they want to use?
User avatar
Carsten
Site Admin
Posts:2170
Joined:2004-08-19, 13:46
Location:Germany
Contact:

Post by Carsten » 2006-06-01, 09:32

Thrawn wrote:I also don't think that this works well but who the hell buys a physics cart?
They said that about the first Voodoo 3D accelerators, too. ;) Besides (@Sindwiller), running physics on the GPU means to buy a second graphics card for that very purpose. I wonder who is going to do that. ;)
I think cpu is the best approach and Doom 3 are good enough - and if not, at least a very good start ;)
I think so, too.
Best regards,
Carsten
Post Reply

Who is online

Users browsing this forum: No registered users and 14 guests