-
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.
R.I.P Physx Hardware
- Thread starter Tom|Nbk
- Start date
Associate
- Joined
- 19 Jun 2007
- Posts
- 1,498
people actually bought them?
Last edited:
I have to say Tom, you do really like to bait people
Matthew
Great contribution to the thread subject thanks.
Thankyou Tom. I was thought it was slightly hypocritical considering how you go at other people for this sort of thing, and then you (ok, I admit he seems like the ONLY physx supporter 
) have a bait at Pottsey. Your right, it wasn't exactly a decent contribution.... baiter 
It is common business practice to assimilate other companies for their tech, just to get it out the way. It makes this purchase curious, especially considering they are only going to use the software side of it. I personally haven't looked into what the CUDA is used for, but do wonder where it will go from here. Will it be used to try and get an edge over AMD's GPU programming you think?
Matthew
) have a bait at Pottsey. Your right, it wasn't exactly a decent contribution.... baiter It is common business practice to assimilate other companies for their tech, just to get it out the way. It makes this purchase curious, especially considering they are only going to use the software side of it. I personally haven't looked into what the CUDA is used for, but do wonder where it will go from here. Will it be used to try and get an edge over AMD's GPU programming you think?
Matthew
Thankyou Tom. I was thought it was slightly hypocritical considering how you go at other people for this sort of thing, and then you (ok, I admit he seems like the ONLY physx supporter
It is common business practice to assimilate other companies for their tech, just to get it out the way. It makes this purchase curious, especially considering they are only going to use the software side of it. I personally haven't looked into what the CUDA is used for, but do wonder where it will go from here. Will it be used to try and get an edge over AMD's GPU programming you think?
Matthew
CUDA is the programming language used by the NVIDIA GPU's
Certainly allready gives Nvidia and edge over AMD.
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
AS i've said before Nvidia simply bought out Physx so they could attach concrete boots and let it lay with the fishes. Its cheaper for Nvidia to pay a little money now and let the whole idea die, than have to compete purely because someone else was doing it over the course of several years developement would have cost a lot more than it would to simply buy them out and dump the whole idea.
Everyone on earth with sense can see why hardware physics sucks, extra particles in an explosion moving perfectly are inperceptable to the human eye. When somethign explodes, say a bomb on the right hand side of the wall and the debrie blows left we have no inherent mathematical ability to judge where each piece should fall. THerefore a system based on estimates, which is easily easily easily doable on cpu as it is now, is more than fine, and a hardware based over accurate system will look identical to us as we could never physically know which is accurate based on just watching. To our eyes the wall will blow off in the right direction in both situations, and debrie will maybe be different, might even not be different, but the simple fact is anyone claiming they could tell which is real is lying.
Completely accurate physics calculations are 100% and completely not needed.
They pottsey will say, but look at the video, the video for the love of god look at the video of the cloth moving, how can you say thats not the best thing ever, in any game ever, that changes the way we'll play games forever. To which i'll say, get over yourself, its a piece of cloth. I'm shooting the guy behind the cloth, with a sniper rifle at 400 yards(to scale) and i can barely make out what he's wearing, i don't care if the cloth is rippling perfectly or not.
Then i'll also point out it CAN be done on cpu's it just hasn't. The wrote a demo to show how they can do cloth, which is in their code for their hardware and then they say a cpu can't do it. Sure it can't its code written for and optimised for the PPU, running code for a K10 phenom and running it on an itanium would give horrific results despite the itanium being incredibly fast, at CERTAIN things. IF someone could be bothered to spend half a year coming up with a demo to show how cloth can be done to look identical on the CPU either through full calculations or via a estimating physics engine using 1/10th of the power to do it could be done. But who wants to, Physx did it to show off their hardware, no one needs to show a cpu doing that as no ones going to buy a CPU based on some cloth demo which won't actually be used in game, as the demo is using shedloads of power to do one thing, which in a game it can't do unless the game is called "hanging washing on the line, the first truly PPU only title, excitement to the max" .
You can do all kinds of things in demo's turning full attention to one thing that simply aren't feasable in the real world. Thats why PPU's last game thing that came out the cloth in its own game looked significantly worse than in the demo because the cloth wasn't the only thing available. Its exactly the same as crysis adverts showing the very very very highest settings possible, probably rendered at 0.5fps but show in full speed and it looks simply stunning, but the game comes out, though the engine can cope with those settings hardware can't. PPU hardware can't do what it shows in demo's in real games, as it needs to do more than render one piece of cloth at a time, demo's are for showcasing best case scenario's which are so far removed from real world application that you can't base how good PPU is based on one tiny demo.
Everything Pottsey has ever said would be good in games from a physics standpoint are in games already to a lesser degree, it has nothing whatsoever to do with available gpu/cpu power, but everything to do with man power. every level thats fully destructable needs designers to program it to be fully destructable, its as completely simple as that. You could make a game right now where ever last piece of cloth, and wall, ceiling weapon, car, barrel, EVERYTHING is destructable, and everything blows up nicely and looks realistic, but that game would take 10 years and incredible man power to create and debug. SO ok, i guess theres a slim chance Duke Nukem Forever might infact do that as they have had that long
Hardware physics was dead from the start because it needlessly added complexity when it wasn't in any way needed. You can't tell real from estimated physics, you simply can't, its impossible we don't see the world in complex mathematical equations, we can't tell what is the right and wrong way for something to explode. Physics only dictates where the debrie will fly, not if the wall is destructable thats a pure and simply design choice and time requirement on if it can be done.
Its been suggested by many that the entire idea was to eventually (sooner rather than later) be bought out by a company if they can simply advertise well enough to make people think they need super realistic physics, on that level they succeeded perfectly, so at some point someone would buy them out to save developing a competing product. Physx won, they set out to achieve exactly what they have, they've made a killing(i would hope) and thats the end of it.
IF nvidia ever do port the API to be usable on their gpu's is questionable at best, a PR stunt at worst. Whose going to buy a second 8800gt for tiny increase in fps in 3 games, when they can buy a 2 8800 gt to give a 30-80% boost in fps in a damn site more games. no one, thats who.
Everyone on earth with sense can see why hardware physics sucks, extra particles in an explosion moving perfectly are inperceptable to the human eye. When somethign explodes, say a bomb on the right hand side of the wall and the debrie blows left we have no inherent mathematical ability to judge where each piece should fall. THerefore a system based on estimates, which is easily easily easily doable on cpu as it is now, is more than fine, and a hardware based over accurate system will look identical to us as we could never physically know which is accurate based on just watching. To our eyes the wall will blow off in the right direction in both situations, and debrie will maybe be different, might even not be different, but the simple fact is anyone claiming they could tell which is real is lying.
Completely accurate physics calculations are 100% and completely not needed.
They pottsey will say, but look at the video, the video for the love of god look at the video of the cloth moving, how can you say thats not the best thing ever, in any game ever, that changes the way we'll play games forever. To which i'll say, get over yourself, its a piece of cloth. I'm shooting the guy behind the cloth, with a sniper rifle at 400 yards(to scale) and i can barely make out what he's wearing, i don't care if the cloth is rippling perfectly or not.
Then i'll also point out it CAN be done on cpu's it just hasn't. The wrote a demo to show how they can do cloth, which is in their code for their hardware and then they say a cpu can't do it. Sure it can't its code written for and optimised for the PPU, running code for a K10 phenom and running it on an itanium would give horrific results despite the itanium being incredibly fast, at CERTAIN things. IF someone could be bothered to spend half a year coming up with a demo to show how cloth can be done to look identical on the CPU either through full calculations or via a estimating physics engine using 1/10th of the power to do it could be done. But who wants to, Physx did it to show off their hardware, no one needs to show a cpu doing that as no ones going to buy a CPU based on some cloth demo which won't actually be used in game, as the demo is using shedloads of power to do one thing, which in a game it can't do unless the game is called "hanging washing on the line, the first truly PPU only title, excitement to the max" .
You can do all kinds of things in demo's turning full attention to one thing that simply aren't feasable in the real world. Thats why PPU's last game thing that came out the cloth in its own game looked significantly worse than in the demo because the cloth wasn't the only thing available. Its exactly the same as crysis adverts showing the very very very highest settings possible, probably rendered at 0.5fps but show in full speed and it looks simply stunning, but the game comes out, though the engine can cope with those settings hardware can't. PPU hardware can't do what it shows in demo's in real games, as it needs to do more than render one piece of cloth at a time, demo's are for showcasing best case scenario's which are so far removed from real world application that you can't base how good PPU is based on one tiny demo.
Everything Pottsey has ever said would be good in games from a physics standpoint are in games already to a lesser degree, it has nothing whatsoever to do with available gpu/cpu power, but everything to do with man power. every level thats fully destructable needs designers to program it to be fully destructable, its as completely simple as that. You could make a game right now where ever last piece of cloth, and wall, ceiling weapon, car, barrel, EVERYTHING is destructable, and everything blows up nicely and looks realistic, but that game would take 10 years and incredible man power to create and debug. SO ok, i guess theres a slim chance Duke Nukem Forever might infact do that as they have had that long
Hardware physics was dead from the start because it needlessly added complexity when it wasn't in any way needed. You can't tell real from estimated physics, you simply can't, its impossible we don't see the world in complex mathematical equations, we can't tell what is the right and wrong way for something to explode. Physics only dictates where the debrie will fly, not if the wall is destructable thats a pure and simply design choice and time requirement on if it can be done.
Its been suggested by many that the entire idea was to eventually (sooner rather than later) be bought out by a company if they can simply advertise well enough to make people think they need super realistic physics, on that level they succeeded perfectly, so at some point someone would buy them out to save developing a competing product. Physx won, they set out to achieve exactly what they have, they've made a killing(i would hope) and thats the end of it.
IF nvidia ever do port the API to be usable on their gpu's is questionable at best, a PR stunt at worst. Whose going to buy a second 8800gt for tiny increase in fps in 3 games, when they can buy a 2 8800 gt to give a 30-80% boost in fps in a damn site more games. no one, thats who.
Last edited:
Some people here jump all over people like a bad rash. Especially to defend certain purchases *cough*physx*cough*. 
To be fair, it was a nice idea, it just didn't take off, especially with the up rise of quad cores. The way things work is that when people don't buy them, it means there's not much point in them, look ad HDDVD, It died because more people bought blu-ray and personally never saw a point n HDDVD my self with blu-ray's superior capacity, I think I'll end this here before I create a HDDVD versus Blu-ray "war"
To be fair, it was a nice idea, it just didn't take off, especially with the up rise of quad cores. The way things work is that when people don't buy them, it means there's not much point in them, look ad HDDVD, It died because more people bought blu-ray and personally never saw a point n HDDVD my self with blu-ray's superior capacity, I think I'll end this here before I create a HDDVD versus Blu-ray "war"
Soldato
- Joined
- 19 Feb 2007
- Posts
- 15,325
- Location
- Northampton
ive got a ppu and ive noticed a considerable change in the performance in games, better fps and stuff just moves more realisticly, thats from my own experience as i do have one in my build.
Last edited:
Drunkenmaster when are you going understand that the PPU can do the same level/detail of physics as the CPU only far faster? The PPU isn’t doing 100% completely accurate physics calculations. The PPU is there to increase the amount of physics done at once. Instead of having 5 box’s flying, you can have 50. Instead of 5 boulders going down the avalanche you have 100’s. Its not about the exter bolders flying more accurately. Well it can be about accuracy if you want but thats not how people use it for the most part.
“cheaper for Nvidia to pay a little money now and let the whole idea die,”
They have already said that’s not what they are doing.
“The wrote a demo to show how they can do cloth, which is in their code for their hardware and then they say a cpu can't do it. Sure it can't its code written for and optimised for the PPU,”
First of all the demo was optimised for the CPU as well. Second we have yet to see a CPU do cloth at playable speeds in game. Its one thing putting 100% of two cores on cloth and getting 30fps or what ever. But it’s another to do that in game. You cannot spend 100% of a core doing just 1 bit of cloth in a game.
“To which i'll say, get over yourself, its a piece of cloth. I'm shooting the guy behind the cloth,”
Well if the guy is in a tent then you cannot shoot him. As the cloth is solid. One of many examples with full cloth is usefull. Its also usefull in none shooting based games.#
“Completely accurate physics calculations are 100% and completely not needed.”“on the CPU either through full calculations or via a estimating physics engine using 1/10th of the power to do it could be done.”
Not this rubbish again. You get proven wrong time and time again, you then vanish from the thread once proven wrong. Then weeks/months later come back with the same rubbish again Why do you do this?
“ever last piece of cloth, and wall, ceiling weapon, car, barrel, EVERYTHING is destructable, and everything blows up nicely and looks realistic, but that game would take 10 years and incredible man power to create and debug.”
Totally and completely wrong as proven last time. It’s hardly any harder to place a destructible wall then it is a solid wall. Warmonger is 90% destructible apart from some areas that are better none destructible. It didn’t take 10 years to make. Many people have made their own 100% destructible levels/maps those didn’t take years to make in both UT and Warmonger and other games.
Clearly you know nothing about game design as you have proven time and time again with your rubbish about 1st and 2nd level physics.
“Everything Pottsey has ever said would be good in games from a physics standpoint are in games already to a lesser degree,”
Thats not true as far as I can tell. Well it depends how you define lesser.
“we can't tell what is the right and wrong way for something to explode.”
As normal your missing the point. That’s not what PPU’s are about in games. PPU’s are not about 100% accurate physics for games.
“cheaper for Nvidia to pay a little money now and let the whole idea die,”
They have already said that’s not what they are doing.
“The wrote a demo to show how they can do cloth, which is in their code for their hardware and then they say a cpu can't do it. Sure it can't its code written for and optimised for the PPU,”
First of all the demo was optimised for the CPU as well. Second we have yet to see a CPU do cloth at playable speeds in game. Its one thing putting 100% of two cores on cloth and getting 30fps or what ever. But it’s another to do that in game. You cannot spend 100% of a core doing just 1 bit of cloth in a game.
“To which i'll say, get over yourself, its a piece of cloth. I'm shooting the guy behind the cloth,”
Well if the guy is in a tent then you cannot shoot him. As the cloth is solid. One of many examples with full cloth is usefull. Its also usefull in none shooting based games.#
“Completely accurate physics calculations are 100% and completely not needed.”“on the CPU either through full calculations or via a estimating physics engine using 1/10th of the power to do it could be done.”
Not this rubbish again. You get proven wrong time and time again, you then vanish from the thread once proven wrong. Then weeks/months later come back with the same rubbish again Why do you do this?
“ever last piece of cloth, and wall, ceiling weapon, car, barrel, EVERYTHING is destructable, and everything blows up nicely and looks realistic, but that game would take 10 years and incredible man power to create and debug.”
Totally and completely wrong as proven last time. It’s hardly any harder to place a destructible wall then it is a solid wall. Warmonger is 90% destructible apart from some areas that are better none destructible. It didn’t take 10 years to make. Many people have made their own 100% destructible levels/maps those didn’t take years to make in both UT and Warmonger and other games.
Clearly you know nothing about game design as you have proven time and time again with your rubbish about 1st and 2nd level physics.
“Everything Pottsey has ever said would be good in games from a physics standpoint are in games already to a lesser degree,”
Thats not true as far as I can tell. Well it depends how you define lesser.
“we can't tell what is the right and wrong way for something to explode.”
As normal your missing the point. That’s not what PPU’s are about in games. PPU’s are not about 100% accurate physics for games.
Last edited:
Soldato
- Joined
- 20 Oct 2002
- Posts
- 12,023
- Location
- Derby
Nvidia have plans to integrate the hardware into thier Quadro Range