CPU PhysX?

[drunkenmaster]

which no one would say didn't enchance the game considerably if they actually played such a game but won't happen because no developer wants to narrow down their potential market that much.

Yes, I most certainly would say it doesn't enhance a game, effects don't enhance a game, they enhance the effects.

Every last massively physx using game has sacrificed all game functionality, playability and story to focus on doing more effects.

Theres one very simply variable you CAN NOT AVOID. Time, if you could take 15 years on a game, sure you can add all the little effects and find some hardware to run it, really though, as certain HUGE game makers have learned, you take more than 4-5 years to make a game and you've got an obsolete title and have to start again.

You have limited time, money and man power, putting these people to work on the graphics and game world engine, story, texture design, animation will ALWAYS AND WITHOUT QUESTION yield a better game, than wasting half your team adding extra physics effects.

Games take a lot of programming, and wasting it on the tiny extra's, rather than the main game is beyond a waste of resources. Dark Void, Mirror's edge, to get a game working with lots of phsyx functionality, they sacrificed story, length, playability.

As for suggesting phsyx is capable of so much more earlier up, horse crap, the design and things that can be done are based on the designers thoughts, physx is merely a way to calculate things after the design. Everything must be designed, every breakable object in a game must be designed to be breakable, and this will always take FAR longer than making the same object thats not breakable.

You can program ANYTHING to use any physics effects, physx is not required to do anything, physx is merely a ultra accurate overly complicated engine, made that way so it only runs well on the hardware it was designed to sell.

As we've seen in each and every single physx title, needlessly accurate calculations don't somehow fix badly made physics effects. Batman AA, you've got a needlessly complicated and powerful physics engine being used to make an effect thats not accurate in the first place, you've got paper/leaves moving right through the characters. An estimation with a far more simply physics engine could have provided indistinguishable effects to the naked eye(and a better made effect could easily have been made more realistic aswell on any physics engine), but at massively less compute time.

We've seen pointless videos' where the whole screen is taken up by a cloth moving around ultra realistically(apparently realistically, whose to say), problem is, we haven't and won't see that in game, unless "Laundry Day" the ultimate game to give your wife for Xmas for the Wii comes out.

 

[Rroff]

Excuse me being to the point and frank

Yes, I most certainly would say it doesn't enhance a game, effects don't enhance a game, they enhance the effects.

Rubbish

Every last massively physx using game has sacrificed all game functionality, playability and story to focus on doing more effects.

Rubbish

Theres one very simply variable you CAN NOT AVOID. Time, if you could take 15 years on a game, sure you can add all the little effects and find some hardware to run it, really though, as certain HUGE game makers have learned, you take more than 4-5 years to make a game and you've got an obsolete title and have to start again.

You have limited time, money and man power, putting these people to work on the graphics and game world engine, story, texture design, animation will ALWAYS AND WITHOUT QUESTION yield a better game, than wasting half your team adding extra physics effects.

This is the point of having middleware physics APIs like PhysX

Games take a lot of programming, and wasting it on the tiny extra's, rather than the main game is beyond a waste of resources. Dark Void, Mirror's edge, to get a game working with lots of phsyx functionality, they sacrificed story, length, playability.

Physics was tacked on over the top of both of these games, more so mirrors edge than dark void. They didn't spend time on physics when they could have been working on addition substance to the games - the games just don't have that substance in the first place.

As for suggesting phsyx is capable of so much more earlier up, horse crap, the design and things that can be done are based on the designers thoughts, physx is merely a way to calculate things after the design. Everything must be designed, every breakable object in a game must be designed to be breakable, and this will always take FAR longer than making the same object thats not breakable.

You can program ANYTHING to use any physics effects, physx is not required to do anything, physx is merely a ultra accurate overly complicated engine, made that way so it only runs well on the hardware it was designed to sell.

As we've seen in each and every single physx title, needlessly accurate calculations don't somehow fix badly made physics effects. Batman AA, you've got a needlessly complicated and powerful physics engine being used to make an effect thats not accurate in the first place, you've got paper/leaves moving right through the characters. An estimation with a far more simply physics engine could have provided indistinguishable effects to the naked eye(and a better made effect could easily have been made more realistic aswell on any physics engine), but at massively less compute time.

We've seen pointless videos' where the whole screen is taken up by a cloth moving around ultra realistically(apparently realistically, whose to say), problem is, we haven't and won't see that in game, unless "Laundry Day" the ultimate game to give your wife for Xmas for the Wii comes out.

PhysX is capable of a LOT more than any implementation we've seen so far - it can do far more complex soft body effects, integrated/jointed rigid body effects, and so on - making entire environments much more interactive and immersive. While there is still a certain amount of design aspect to it still its a lot less than creating your own physics implementation from scratch - you can create "template" functions so to speak for your commonly used physics effects and simply tag your assets to be run through that part of the code.*

The days of PhysX being ultra realistic complex physics are long ago over in its novodex days. Its entirely focused on game physics now trading accurate and realistic simulation for simulations that better suit a gaming environment.



*It took me all of 15 minutes to create some functions for handling the world gun models, then I simply tagged up a low poly collision tri-mesh with parts for the gun and magazine (so the magazine can break away from the gun for various reasons) and dropped it into my world - every single gun can now be tagged up the same way and automatically has the right physics running through the same functions.

 

[Final8y]

Physics was tacked on over the top of both of these games, more so mirrors edge than dark void. They didn't spend time on physics when they could have been working on addition substance to the games - the games just don't have that substance in the first place.



PhysX is capable of a LOT more than any implementation we've seen so far - it can do far more complex soft body effects, integrated/jointed rigid body effects, and so on - making entire environments much more interactive and immersive.

What it has been used for up until now has been nothing more than waste of resources.

 

[RavenXXX2]

Yes, I most certainly would say it doesn't enhance a game, effects don't enhance a game, they enhance the effects.

Every last massively physx using game has sacrificed all game functionality, playability and story to focus on doing more effects.

Theres one very simply variable you CAN NOT AVOID. Time, if you could take 15 years on a game, sure you can add all the little effects and find some hardware to run it, really though, as certain HUGE game makers have learned, you take more than 4-5 years to make a game and you've got an obsolete title and have to start again.

You have limited time, money and man power, putting these people to work on the graphics and game world engine, story, texture design, animation will ALWAYS AND WITHOUT QUESTION yield a better game, than wasting half your team adding extra physics effects.

Games take a lot of programming, and wasting it on the tiny extra's, rather than the main game is beyond a waste of resources. Dark Void, Mirror's edge, to get a game working with lots of phsyx functionality, they sacrificed story, length, playability.

As for suggesting phsyx is capable of so much more earlier up, horse crap, the design and things that can be done are based on the designers thoughts, physx is merely a way to calculate things after the design. Everything must be designed, every breakable object in a game must be designed to be breakable, and this will always take FAR longer than making the same object thats not breakable.

You can program ANYTHING to use any physics effects, physx is not required to do anything, physx is merely a ultra accurate overly complicated engine, made that way so it only runs well on the hardware it was designed to sell.

As we've seen in each and every single physx title, needlessly accurate calculations don't somehow fix badly made physics effects. Batman AA, you've got a needlessly complicated and powerful physics engine being used to make an effect thats not accurate in the first place, you've got paper/leaves moving right through the characters. An estimation with a far more simply physics engine could have provided indistinguishable effects to the naked eye(and a better made effect could easily have been made more realistic aswell on any physics engine), but at massively less compute time.

We've seen pointless videos' where the whole screen is taken up by a cloth moving around ultra realistically(apparently realistically, whose to say), problem is, we haven't and won't see that in game, unless "Laundry Day" the ultimate game to give your wife for Xmas for the Wii comes out.

Quoted for truth.

 

[Final8y]

Quoted for truth.

Damn straight.

 

[xBLiNG]

PhysX won't be worth much until it's on ATi cards along with nVidia ones. Seems suicidal for a dev to cut out a large part of their audience by only allowing nVidia uses with dedicated PhysX cards to play with decent settings or at all.

 

[Rroff]

What it has been used for up until now has been nothing more than waste of resources.

On that note I wouldn't disagree with you.

 

[Rroff]

Quoted for truth.

Damn straight.

Guy most of what he said is based on some hysterical anti-physx rhetoric and watching videos on youtube... theres very little truth to it - and what there is is mostly out of date information.

 

[reflux]

I can only think of three games that implement meaningful physics - HL2, Company of Heroes and Battlefield Bad Company 2. The hardware dependency of PhysX ensures it stays a product based on gimmickry.

 

[Strawberry]

As for suggesting phsyx is capable of so much more earlier up, horse crap, the design and things that can be done are based on the designers thoughts, physx is merely a way to calculate things after the design. Everything must be designed, every breakable object in a game must be designed to be breakable, and this will always take FAR longer than making the same object thats not breakable.

I wouldn't readily agree with that. We've seen more than enough techdemos of realistically self-breakable objects that require no user input (Crysis, The Force Unleashed, etc), instead using their own physical properties to decide where to break. Even if this would be too computationally expensive, such things could be pre-computed.