Sorry your browser is not supported!

You are using an outdated browser that does not support modern web technologies, in order to use this site please update to a new browser.

Browsers supported include Chrome, FireFox, Safari, Opera, Internet Explorer 10+ or Microsoft Edge.

Dark Physics & Dark A.I. & Dark Dynamix / Hardware Vs Software no Difference

Author
Message
Codger
22
Years of Service
User Offline
Joined: 23rd Nov 2002
Location:
Posted: 24th Oct 2006 05:01
Well I purchased the BFG PPU and after finally getting the card to be recognized by the system .. I see little or no change in speed when running the Phyx demos. The demos report the FPS and number of objects. For example using the make particle demo the fps for hardware and software is within 5% (100 fps). The domino demo is also about the same fps.


What is everyone else with hardware getting on the dominos
My Results are

Software & hardware 115 FPS


Codger

System
PIV 2.8 MZ 512 Mem
FX 5600 256 mem
dark coder
22
Years of Service
User Offline
Joined: 6th Oct 2002
Location: Japan
Posted: 24th Oct 2006 08:03
Well most of the demos are made with software in-mind and don't use much physics, such as the dominos example, if you look in the 'Physx' folder of the demos you can see some demos which have alot more objects, and don't forget the fluids .

Hallowed are the ori.
Chenak
22
Years of Service
User Offline
Joined: 13th Sep 2002
Location: United Kingdom
Posted: 24th Oct 2006 14:23
If that was the case then TGC would not have the license to sell it...

The hardware demos, the particles and cloth run significantly faster, especially with interactions.
david w
18
Years of Service
User Offline
Joined: 18th Dec 2005
Location: U.S.A. Michigan
Posted: 24th Oct 2006 23:08
I have to agree with codger on this one. On my system I see almost no distinct difference between hardware and software mode's. This doesnt mean there isnt any, I just dont see it. Makes me wonder if I wasted my money here. Fluids are not even really useable with/without hardware. I think somethings got to be up.
Argoon
18
Years of Service
User Offline
Joined: 16th Oct 2006
Location:
Posted: 25th Oct 2006 01:51
Yes i also can´t see the diference with or without the PPU, in all the demos that have the two options the fps are almost the same( i,m using the free version), i know that PhysX can do more as it is seing in CellFactor, so here most be a PPU implementation problem.

Cheers
Argoon
18
Years of Service
User Offline
Joined: 16th Oct 2006
Location:
Posted: 25th Oct 2006 02:09
I forgot to post my dominos results.

60 fps for the software the first domino falling to the side contrary of the motion.

60 fps for the hardware all the dominos fall to the same side and the falling motion is visualy more fast, but the fps´s were always the same on the entire simulation.

This on a AMD64 x2 4200+ and the Asus P1.

codger 115 fps , what CPU do you have? For the DP developers and Ageia- Is the hardware result somewhat tied to the CPU result? If not why codger have 115 fps with the PPU and hi only 60, don´t all the cards perform the same?

Cheers
Codger
22
Years of Service
User Offline
Joined: 23rd Nov 2002
Location:
Posted: 25th Oct 2006 04:08
Argoon
I rechecked my fps and it ranged from a low of 105 when Dominos were falling to 130 when all dominos were still. My CPU is just an Intel Pentium 4 2.8. The biggest speed benifits come from memory both system & Graphics cards and the graphics card itself.
256 for the GPU and 512 for the system

Codger

System
PIV 2.8 MZ 512 Mem
FX 5600 256 mem
bosskeith
18
Years of Service
User Offline
Joined: 5th Dec 2005
Location:
Posted: 25th Oct 2006 04:09
no because you still rely on your gpu to render the scene...the PPU reduces the work of the CPU. Simple demos may not have as much of an impact, however in a FPS what happens when the CPU has to handle all the AI along with the physics they would bottleneck...with the PPU it only has to handle the AI therefore allowing it to execute faster.The demos dont have to handle so much.Therefore the cpu is available to focus it's calculations on the physics.

this is my understanding of how it all works.

BatVink
Moderator
21
Years of Service
User Offline
Joined: 4th Apr 2003
Location: Gods own County, UK
Posted: 25th Oct 2006 19:00
BossKeith has a point. A purely physics demo is not a test. Wrap a game around the physics and you should see the difference.



Hoozer
18
Years of Service
User Offline
Joined: 8th Sep 2006
Location: Bremerhaven (Germany)
Posted: 25th Oct 2006 19:33
I would say the amount of the physics plays also into the use of such a card!

If you can organize your game just to use about 20 to 40 objects, the CPU may be able to handle all this without an addon-card! BUT when you decide to play it BIG and use up to 1000s of objects the CPU will mostly get crazy with such an amount of PhysiX to compute! Then a PhysiX-Card would make sense!

Sadly I have not such a card at hand, but it would be very interesting to see, at what a point that card would slow down (reaches it's limits)!


Hoozer

Athlon 64X2 4800+ (939);
2GB DDR-400;
GeForce 6800LE (@12 Pixel-Shader-Units, 6 Vertex-Shader-Units; GPU: @ 380 MHz, RAM: 434 MHz)
Chenak
22
Years of Service
User Offline
Joined: 13th Sep 2002
Location: United Kingdom
Posted: 25th Oct 2006 20:14
Again, using a large amount of interactive particals and using detailed cloth you will see the difference in FPS, particulally if you have a crappy computer You prolly wont see much of a difference just using a static rigid body mesh and a few dynamic boxes.
bosskeith
18
Years of Service
User Offline
Joined: 5th Dec 2005
Location:
Posted: 25th Oct 2006 23:04 Edited at: 25th Oct 2006 23:07
also another thing is if you can run the program at your graphics cards maximum proficiency (she's giving it all she's got captin!) in software mode, having the PPU will not be able to make the GPU push out any more FPS.

Alfred
18
Years of Service
User Offline
Joined: 22nd Jun 2006
Location: Jalisco, Mexico
Posted: 26th Oct 2006 00:50 Edited at: 26th Oct 2006 00:58
I agree with bosskeith . You should be able to notice the difference even more with a good video card. For example, I get a low 387 fps (no joking) when the domino´s demo ball is rolling and a high 452 fps when all the dominos are still and this is with no PhysX card. I have an ATI Radeon 9950 and an Athlon XP 2700 MHz processor. If you lower the screen resolution ( your app´s resolution) you can get ever more fps. This means that the final game can really benefit from not having to wait for the CPU to process all the collisions and have as many fps as the video card can produce as the GPU is mainly the cause of the slowdown. (of course, the main CPU also matters).

Hallowed are the ori.
Benjamin
21
Years of Service
User Offline
Joined: 24th Nov 2002
Location: France
Posted: 26th Oct 2006 07:33
Quote: "Then a PhysiX-Card would make sense!"

And then a more powerful graphics card makes sense, and then a more powerful CPU makes sense.

Tempest - P2P UDP Multiplayer Plugin (DBP/DBCe)
Download the free version
Codger
22
Years of Service
User Offline
Joined: 23rd Nov 2002
Location:
Posted: 26th Oct 2006 09:21
The point of this thread is to see if there is a benefit to purchasing a PPU. We all know that some machines run faster and that by changing the screen resolution the FPS can be modified.
When I spend $300.00 for a card whether it is a new GPU of a PPU I expect to see a noticeable change in performance.
I am very enthusiastic about Dark Physics but if the PPU adds no performance increase then we would get more "Bang for the buck" from a new GPU or CPU.

System
PIV 2.8 MZ 512 Mem
FX 5600 256 mem
david w
18
Years of Service
User Offline
Joined: 18th Dec 2005
Location: U.S.A. Michigan
Posted: 26th Oct 2006 13:15
I have had the ppu for a couple months now and I havent seen any noticeable difference with any application with/without hardware. Sometimes I wonder if the commands "phystart" is not enabling it. I also have a radeon x1950xtx 512mb, x2 3800, and 1.5mb system ram. So the graphics are not a problem. I just dont seen any difference on any demo/game or anything. Again I wonder if there is a bug somewhere.
Codger
22
Years of Service
User Offline
Joined: 23rd Nov 2002
Location:
Posted: 27th Oct 2006 03:39 Edited at: 27th Oct 2006 03:41
Quote: "I just dont seen any difference on any demo/game or anything. Again I wonder if there is a bug somewhere. "

That's my point exactly I am really hoping that "The Game Creators" have not set a variable correctly and that it can be fixed soon.

Is the only function of the hardware to act as a expensive dongle to allow the software to switch on fluids?

How can you make a demo or game for the contest if the hardware has no effect?

I also wish that Mike Johnson and or Richard Davey would chime in with thier FPS for "Dominos" or any other demo software so we can see if the pro's are getting a benifit from the hardware.

Codger

System
PIV 2.8 MZ 512 Mem
FX 5600 256 mem
dark coder
22
Years of Service
User Offline
Joined: 6th Oct 2002
Location: Japan
Posted: 27th Oct 2006 08:03
The dominos demo only has around 50 objects and only few collisions, try the bricks demo, I get double the fps in hardware mode.

Hallowed are the ori.
Codger
22
Years of Service
User Offline
Joined: 23rd Nov 2002
Location:
Posted: 31st Oct 2006 04:15 Edited at: 31st Oct 2006 04:16
I came across couple of links and a demo by AGEIA "carwash" again I notice no difference between hardware and software. I am sure my card is not working correctly.


http://www.planetphysx.com/weblog/archives/2006/10/ageia_physx_carwash_demo_1.html#more

Here are a number of additional links that others may find interesting.

http://personal.inet.fi/atk/kjh2348fs/ageia_physx.html

codger

System
PIV 2.8 MZ 512 Mem
FX 5600 256 mem
dark coder
22
Years of Service
User Offline
Joined: 6th Oct 2002
Location: Japan
Posted: 31st Oct 2006 06:12
So fluids do work in software , and on the carwash demo, I get around 450fps with nothing, 40-50 with fluids 15-16 with fluids in software, so your slowdown issue is likely your graphics card.

Hallowed are the ori.
Codger
22
Years of Service
User Offline
Joined: 23rd Nov 2002
Location:
Posted: 31st Oct 2006 07:13
dark coder

What graphics card / cpu etc is on your system?

System
PIV 2.8 MZ 512 Mem
FX 5600 256 mem
dark coder
22
Years of Service
User Offline
Joined: 6th Oct 2002
Location: Japan
Posted: 31st Oct 2006 07:27
2X 7800gtx, x2 4400+, 2gigs ram, bfg ppu.

Hallowed are the ori.
BatVink
Moderator
21
Years of Service
User Offline
Joined: 4th Apr 2003
Location: Gods own County, UK
Posted: 31st Oct 2006 10:16
I would imagine that in future mesh deformation may get included, in which case the PPU will become even more essential.



Hoozer
18
Years of Service
User Offline
Joined: 8th Sep 2006
Location: Bremerhaven (Germany)
Posted: 31st Oct 2006 14:12
Somehow the fluids do not work on my machine!
(At first there is no real scene (an empty scene), but when I change to Software-Mode, I get the car and the room! I can throw balls, but no fluids get produced when pressing the "f"-key, regardless of the selected fluid!))

Seems like a mixture of Hard-Soft-Ware-Mode!
So I think for fluids it still needs (or at least in the code it is limited to) a PhysX-Card!


Hoozer

Athlon 64X2 4800+ (939);
2GB DDR-400;
GeForce 6800LE (@12 Pixel-Shader-Units, 6 Vertex-Shader-Units; GPU: @ 380 MHz, RAM: 434 MHz)

Login to post a reply

Server time is: 2024-11-24 14:28:38
Your offset time is: 2024-11-24 14:28:38