• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

PhysX Accelerators

Durzel said:
you especially can't call it "bad coding" when you have no idea what coding is

Something tells me you are in this bracket if you are comparing computer games to cars, lmao :D

@ sup3rc0w - say that again in a year's time. Physx cards will be everywhere soon enough as physics is the next step in game realism.
 
I think we'll have to agree to disagree here.

But no-one here can argue against the fact that if another year was added to the development time of games they would perform better and have less bugs.

Infact, I know its possible because they did it for GT4 on the PS2.

Nurburgring mapped with 4Mb ram with decent performance?

Well they managed it...
 
Robbie G said:
@ sup3rc0w - say that again in a year's time. Physx cards will be everywhere soon enough as physics is the next step in game realism.

They said that a year ago still hasn't happened, home users don't even know anything about physics and enthusiasts don't want to spend loads of money on yet another card that will need to be upgraded just for some better physics. Ati have already created a solution using graphics cards to calculate physics instead, onboard physics may happen, but I'd imagine ageia will be out of business before then, no games have been created with noteable physics affects using the physx card.
 
Robbie G said:
And, if one result yields better performance than the other, this is known as bad coding.

You either didn't read the rest of my previous post or did not understand it. It may not always be possible or desirable to use the fastest technique due to the techniques limitations or the fact it does not interact well with other effects that you are trying to apply to the scene. This does not equal bad coding.
 
Concorde Rules said:
I think we'll have to agree to disagree here.

But no-one here can argue against the fact that if another year was added to the development time of games they would perform better and have less bugs.

Infact, I know its possible because they did it for GT4 on the PS2.

Nurburgring mapped with 4Mb ram with decent performance?

Well they managed it...

What does the nurburgring and 4MB RAM have to do with being able to spot bad coding?

I would also be interested to know your answer to my question about which hardware you were comparing the perfomance of Total Annihilation to Supreme Commander on?
 
“Look at the Alan Wake vids. In stead of using PhysX, they use cores of c2d or x2 chips to do the same job... or am I being a nub ?”
I see no advance or high end physics in Alan Wake. If anything Alan Wake proves putting a complete CPU core to physics isn’t enough.

The physics in Alan Wake could be done on a PPU or GPU with ease freeing up a core for other stuff making the game even faster. Why waste a core on physics?





“but I'd imagine ageia will be out of business before then, no games have been created with noteable physics affects using the physx card.”
Its not just about new effects it’s also about offloading physics from the CPU so the CPU can work on other things and you get a speed boost. Ageia are not going to go out of business any time soon as they are doing to well.

As for no notable effects what about Cell Factor: Revolution or Warmonger?
 
Last edited:
i have to say i was mighty impressed by some of the videos out there, i hope some racing games use this, would it help developers make cars act more realistic i.e tyre deformation? im guessing it will improve damage effects by a long way!
 
Concorde Rules said:
I have none at all.

But tell me why Supreme Commander runs so poorly compared to total annihilation and when my GPU is hardly loaded (55C instead of 60+ in either game?)

i have no coding experience by the way but common sense tells me that

surely the fact that supreme commander runs poorly has to do with the volume of AI that the cpu has to calculate? which has been discussed in other threads. this should therefore confirm even more that the cpu could not add yet another process to its job list without being grinding to a halt?
 
Pottsey said:
Its not just about new effects it’s also about offloading physics from the CPU so the CPU can work on other things and you get a speed boost. Ageia are not going to go out of business any time soon as they are doing to well.

As for no notable effects what about Cell Factor: Revolution or Warmonger?

They can't be that good because I haven't even heard of those games anywhere.

Most games are gpu limited so offloading work from the cpu isn't going to make much of a difference, coh is a good example of this and it is an rts game, a genre which is supposed to rely heavily on the cpu, there was practically zero performance change when using a better cpu. I do remember seeing a video of a fps game though where ageia was enabled and the framerate dropped about 30fps. Ageia have a seriously long way to go before getting any decent results and customers.
 
Last edited:
skullman said:
What he said. You only have to look at cryengine2 vids to see that advanced physics is possible without the need of a dedicated processor such as AGEIA's offering.

Can you tell me what you think is so advanced about the physics in CryEngine2 that you have seen from those videos?
 
Marc Fraser said:
Can you tell me what you think is so advanced about the physics in CryEngine2 that you have seen from those videos?

Deformable vegetation and destructable buildings. Not hugely advanced physics, but adds to the game rather than more pointless liquid/cloth effects.
 
“Most games are gpu limited so offloading work from the cpu isn't going to make much of a difference, coh is a good example of this and it is an rts game,”
COH as in City of Heroes? Never heard of an RTS by the name of COH. The COH game I heard of is CPU limited as are lots of games and using a dual core speeds the game up a lot and adding a PPU on top of the dual core or without it also speeds up the game by 30% ish.






“What he said. You only have to look at cryengine2 vids to see that advanced physics is possible without the need of a dedicated processor such as AGEIA's offering.”
Last time I looked cryengine2 used hardware to accelerate physics not the CPU with a software engine which proves my point that the CPU is not enough.






”I do remember seeing a video of a fps game though where ageia was enabled and the framerate dropped about 30fps. Ageia have a seriously long way to go before getting any decent results and customers.”
That was the very first game with had some major performance issues that have since been fixed. Not only that but the 30fps drop was for a split of a second not even a full second. So short you wouldn’t notice. Why do so many people judge a PPU’s performance on 1 bad game? The game came out over a year ago with bugs that have since been fixed. You don’t judge 3dcards on year old buggy drivers. You use the newer games and new drivers. Just patch up Ghost Recon to the new version and the new Ageia drivers and the performance issues are mostly gone.







“but adds to the game rather than more pointless liquid/cloth effects.“
Liquid and cloth are not more pointless though it depends on the game and where its based. That’s like saying glass windows that break and you can shoot though are pointless. Right now most games with camping tents, clown tents, trucks with cloth(*), curtains over a windows use cloth as an solid object that are normally indestructible.

With real cloth you could shoot though the tent. How many times in games have you had to move around cloth as you cannot shoot though it? I have been in many tents and had to move outside to shoot the bad guys as my gun cannot go though a tent!
Same for a house if the curtains are shut tough luck you cannot do anything with real cloth you can bend, move or shoot though the cloth. As for troop transports right now they have fake cloth that you have to move around. All the little things as well like cloths lying around are solid objects.

* What do you call those military troop vehicles with cloth around the back? Like in the 20second point in this vid? http://www.youtube.com/watch?v=fjB4EzbIgSw The vid also shows one use of cloth.

http://www.planetphysx.com/weblog/archives/2006/12/holiday_update_from_chimera_is.html This is a moded level to show off cloth. Noticed the cloth blimp and everything else cloth is used for. You really telling me that’s pointless?

As for liquids surely you can see the use for decent liquids?
 
Last edited:
Marc Fraser said:
What does the nurburgring and 4MB RAM have to do with being able to spot bad coding?

I would also be interested to know your answer to my question about which hardware you were comparing the perfomance of Total Annihilation to Supreme Commander on?

1. It doesn't, but Im sure you can testify how difficult it its to get 6 cars, physics and textures into 4mb of ram. Therefore they refined the code enough to load and unload what it doesn't need perfectly...

2. My 3.456ghz Conroe rig. I think as someone said before is due to the new weaponry calculations, im still extremely disappointed it doesn't use the 2nd core much or at all :(



Until I see a game thats been made for both quad cores and agie from an independant party which hasn't been payed off with decent physics and decent graphics then im afraid im gonna keep saying that the card is a complete waste of money.

£150 = the difference between a E6600 and a QX6600 in 3 months time. I know what I'd rather have... QX6600 Vs. Physics card *hmmm* :rolleyes:
 
Last edited:
Concorde Rules said:
2. My 3.456ghz Conroe rig. I think as someone said before is due to the new weaponry calculations, im still extremely disappointed it doesn't use the 2nd core much or at all :(

Are you seriously telling me that someone had to explain to you why a game designed in 1997 for the hardware available then (please note that it wouldn't have just been designed to run well on the top level hardware either) runs better on your 3.456ghz Conroe than a game designed in 2006 for current hardware.

If you can't work this one out for yourself then you really are in no position to be having a discussion about what is bad coding.
 
Marc Fraser said:
Are you seriously telling me that someone had to explain to you why a game designed in 1997 for the hardware available then (please note that it wouldn't have just been designed to run well on the top level hardware either) runs better on your 3.456ghz Conroe than a game designed in 2006 for current hardware.

If you can't work this one out for yourself then you really are in no position to be having a discussion about what is bad coding.

Errrm, your missing the point.

SC is based on TA, in fact its pretty much the same except for better graphics and the new physics calculations which apparently have been used.

If these new physics calculations are really complex then I understand why it gets slow, but if they arn't that complex (I can't actually remember what the battles are like - aint played for weeks) why is it so slow?

Maybe it isn't a case of bad coding, but the game still has bugs, and again if it had had another year of development the performance could be better.

Im not argueing anymore. FACT = MORE TIME = BETTER PRODUCT.
 
Robbie G said:
Something tells me you are in this bracket if you are comparing computer games to cars, lmao :D

@ sup3rc0w - say that again in a year's time. Physx cards will be everywhere soon enough as physics is the next step in game realism.
lmao :D :D :D I do programming for a living m8 so don't presume to know anything about me.
 
jesus,people are talking about two entirely, utterly different things in this thread.

being able to shoot through the tent and the tent looking and moving much more realistically are two utterly and entirely different ends of the physics question spectrum. you can shoot through a glass window because the game engine lets you, you can shoot through a tent if a game engine lets you, it has nothing to do with being cloth, the cloth reacting exactly like cloth when you shoot through it is what the ageia card is supposed to let systems do more easily. but it CAN still be done on the cpu. shooting through stuff, letting bullets pass through certain types of object are all really design and engine questions, extra effects through a ppu to model the cloth moving realistically don't effect gameplay really at all. you can shoot through it and it looks good or it looks crap, you still shot through it.

but the point i raised earlier is, do you have to have the water swirling around in a pond so realistically that you have to buy a £150 ppu to do it, or are you happy with that water splashing around using much more basic, guessed patterns that repeat over and over. the fact is that if you shoot into water in HL2 or something you get a little splash of water, do you care if that splash looked slightly more realistic?
 
Pottsey said:
Last time I looked cryengine2 used hardware to accelerate physics not the CPU with a software engine which proves my point that the CPU is not enough.

Software engine or not, it specifically says on the CryEngine 2 website "without the need of specialized coprocessing hardware" - which is what matters. It also says that the physics engine will benefit from multicore cpu's.
 
Back
Top Bottom