Not really.
-
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.
Who/why buys physics cards?
- Thread starter MumboJumbo
- Start date
Lets face it. We're all enthusiasts here - well most of us which is why we shell out 300+ pounds on a graphics card.
I'm sure no one here would have a problem in that case in shelling out 100 pounds on a PPU but we don't? Hardly anyone here has one.
Why? You get nothing from them that is remotely worth anywhere near 100 pounds.
I'm sure no one here would have a problem in that case in shelling out 100 pounds on a PPU but we don't? Hardly anyone here has one.
Why? You get nothing from them that is remotely worth anywhere near 100 pounds.
Indeed, none of the games i'm playing support it so I know for a fact i'm not missing out on anything.
“a couple of tech demo level's while not being used in the game engine for any other parts of the game.”
Give me examples please? I cannot think of any PPU games that just use the PPU for 1 tech demo level and not any other part of the game. Unless I forgot about some games your makeing stuff up.
“again i said DIFFERENT physics. you do realise EVERYTHING on the screen is drawn by using physics to work out where and what to draw.”#
You do realise its not the GPU that does that dont you?
GPU do not work out what to draw and where. Well some work out what not to draw, but they don’t work out what to draw.
“as for calling me a liar, first, name a game that improves by 40% in performance using a ppu, then tell me what 1st and 2nd level game engine stuff it effects.”
Name a game? I have done that plenty of times. I am fed up posting benchmarks and naming games. Just search my old posts. I have been posting for ages now about the drivers. Not one person has posted benchmarks with bad performance with the drivers that had the performance bug fixed. All the benchmarks with performance problems are over a year old with drivers before that major bug got sorted. Just look at how well GRAW 2 runs with a PPU when you use the new drivers.
Before I talk about 1st, 2nd and 3rd level physics can you define what you mean? I want to make sure we are talking about the same thing.
“just because a core built for one code, can do ONLY THAT CODE faster. doesn't mean that the same outcome, can be done with different code/engine, on a different core better.”
The core isn’t built for 1 code it’s built for physics. GPU’s are not built for physics just look at their tech specs. Spec wise GPU’s are far worse then PPU’s at physics yet you say otherwise. Prove it.
Your saying the PPU Core built for physics with better specs for physics is slower then the GPU core that’s not built for physics. If your going say something like that you need to prove it and post evidence. Compared the power and specs that matter for physics.
Its not just about the code its about the hardware and GPU don’t have the right hardware to be great at physics. Just look at the poor internal bandwidth on GPU’s far to low.
Give me examples please? I cannot think of any PPU games that just use the PPU for 1 tech demo level and not any other part of the game. Unless I forgot about some games your makeing stuff up.
“again i said DIFFERENT physics. you do realise EVERYTHING on the screen is drawn by using physics to work out where and what to draw.”#
You do realise its not the GPU that does that dont you?
GPU do not work out what to draw and where. Well some work out what not to draw, but they don’t work out what to draw.
“as for calling me a liar, first, name a game that improves by 40% in performance using a ppu, then tell me what 1st and 2nd level game engine stuff it effects.”
Name a game? I have done that plenty of times. I am fed up posting benchmarks and naming games. Just search my old posts. I have been posting for ages now about the drivers. Not one person has posted benchmarks with bad performance with the drivers that had the performance bug fixed. All the benchmarks with performance problems are over a year old with drivers before that major bug got sorted. Just look at how well GRAW 2 runs with a PPU when you use the new drivers.
Before I talk about 1st, 2nd and 3rd level physics can you define what you mean? I want to make sure we are talking about the same thing.
“just because a core built for one code, can do ONLY THAT CODE faster. doesn't mean that the same outcome, can be done with different code/engine, on a different core better.”
The core isn’t built for 1 code it’s built for physics. GPU’s are not built for physics just look at their tech specs. Spec wise GPU’s are far worse then PPU’s at physics yet you say otherwise. Prove it.
Your saying the PPU Core built for physics with better specs for physics is slower then the GPU core that’s not built for physics. If your going say something like that you need to prove it and post evidence. Compared the power and specs that matter for physics.
Its not just about the code its about the hardware and GPU don’t have the right hardware to be great at physics. Just look at the poor internal bandwidth on GPU’s far to low.
I can enable ultra PhsyX on GRAW2 without a PhysX card, and I get practically all the effects with no performance hit.
Me: 1
Ageia: 0
If I can enable PhysX on my system with virtually no performance hit and still get 99% of the effects, what's the point of buying the card. I mean, what does it actually do?
I just don't get it.
Me: 1
Ageia: 0
If I can enable PhysX on my system with virtually no performance hit and still get 99% of the effects, what's the point of buying the card. I mean, what does it actually do?
I just don't get it.
If you ask me it is way too early to buy a physics card. Hardly any games support them and it will be some time before they do or the one killer game comes out which helps sells them.
On a positive note, with nvidia launching motherboards with 3 pci express slots - 2 x 16x for SLI plus 1 x 8x for their upcoming physics card I am sure we will start and see more games using physics models especially if nvidia is backing them.
It will come down to some people will be prepared to pay the extra for the card for the effects and some won't.
It's the same with SLI. Some people will spend £800 on two 8800GTX for SLI but personally I don't see the point when the next gen single card might be as fast as two GTX in SLI now. However, I would never have a go at somebody if that is what they want to spend their money on.
Same with physics cards now. They are for sale and if somebody wants to buy one that's their choice. Me, I won't......well maybe next year
On a positive note, with nvidia launching motherboards with 3 pci express slots - 2 x 16x for SLI plus 1 x 8x for their upcoming physics card I am sure we will start and see more games using physics models especially if nvidia is backing them.
It will come down to some people will be prepared to pay the extra for the card for the effects and some won't.
It's the same with SLI. Some people will spend £800 on two 8800GTX for SLI but personally I don't see the point when the next gen single card might be as fast as two GTX in SLI now. However, I would never have a go at somebody if that is what they want to spend their money on.
Same with physics cards now. They are for sale and if somebody wants to buy one that's their choice. Me, I won't......well maybe next year
I think the below statement covers this topic nicely....
And it may not be Ageia's fault that their PhysX was implemented (i am not really sure i can truly call it implemented) so poorly in early games, but then again they allowed it to happen, so it is their fault. Lots of bad publicity & word of mouth, has casued most of the buying public to be against the company & their technology. Without a buying public to support them, there is no reason for game developers to do the extra work to include this technology within their games, & thus no reason for the public to buy their expensive PPU's which do so little. And so the downward spiral goes, despite however much Ageia may pay ( i suspect) companies to include their technology in their games / engines, they have lost the support of the public & it will be hard to get back.
And it may not be Ageia's fault that their PhysX was implemented (i am not really sure i can truly call it implemented) so poorly in early games, but then again they allowed it to happen, so it is their fault. Lots of bad publicity & word of mouth, has casued most of the buying public to be against the company & their technology. Without a buying public to support them, there is no reason for game developers to do the extra work to include this technology within their games, & thus no reason for the public to buy their expensive PPU's which do so little. And so the downward spiral goes, despite however much Ageia may pay ( i suspect) companies to include their technology in their games / engines, they have lost the support of the public & it will be hard to get back.
Last edited:
“I can enable ultra PhsyX on GRAW2 without a PhysX card, and I get practically all the effects with no performance hit.”
Then you must have some magic CPU that can do extra work without slowdown. Can you prove that? Show screenshots or better yet a movie on extreme physics and normal physics. I don’t believe you have no FPS drop.
Where did you record the data?
“If I can enable PhysX on my system with virtually no performance hit and still get 99% of the effects, what's the point of buying the card. I mean, what does it actually do?”
What does it do? Well it boost FPS while turning on those extra effects.
Some reports are up to 40% FPS in GRAW 2 while the CPU gets a performance drop.
Then you must have some magic CPU that can do extra work without slowdown. Can you prove that? Show screenshots or better yet a movie on extreme physics and normal physics. I don’t believe you have no FPS drop.
Where did you record the data?
“If I can enable PhysX on my system with virtually no performance hit and still get 99% of the effects, what's the point of buying the card. I mean, what does it actually do?”
What does it do? Well it boost FPS while turning on those extra effects.
Some reports are up to 40% FPS in GRAW 2 while the CPU gets a performance drop.
Well that's a first because 2 others managed to get the PPU only island map to work without the PhysXcard but it became a slideshow with the PhysX set to high.
Or are you talking about the normal maps by mistake.
Why is it hard to believe? I have an E6850 running at 3.8 Ghz which, I would think, should be more than capable of rendering a bunch of black squares around the screen, which is pretty much all enabling PhysX does, whether you have the card or not. The effects are really not that impressive in my opinion. Even these so call "wind" effects. They just don't impress me. Even the Ageia tech demos are pretty pants, but that's besides the point.
As for my results; I basically sat on the very first level with Fraps open, then checked what my framerate was at with and without high/extreme PhsyX enabled. The performance difference was little to none. Maybe later on in the game the difference would be more noticable, but I didn't get that far because the game sucks ass.
I find it hard to believe having a PhysX card would boost performance in GRAW by 40%, that's basically saying, thoeretically, if I am getting 100 FPS with and without PhysX enabled without the card itself, I would be getting 140 FPS if I had the card? How would having the card magically improve my peak performance since the performance difference is non-existant when I don't even have a PhysX card?
There is nothing staggering that the PhysX card can do that games implementing there own (Havoc?) physics engines can't already do. Perhaps the cloth effects are nice, but is that really reason to blow nearly £100? I think not. Even effects such as wind have already been done in STALKER.
Just the first level on the single player campaign, maybe that isn't enough to be conclusive about the PhysX card, but the game is so damn awful I wouldn't consider buying the game itself let alone a PhysX card for it. I guess that's pretty conclusive in itself, though.
As for my results; I basically sat on the very first level with Fraps open, then checked what my framerate was at with and without high/extreme PhsyX enabled. The performance difference was little to none. Maybe later on in the game the difference would be more noticable, but I didn't get that far because the game sucks ass.
I find it hard to believe having a PhysX card would boost performance in GRAW by 40%, that's basically saying, thoeretically, if I am getting 100 FPS with and without PhysX enabled without the card itself, I would be getting 140 FPS if I had the card? How would having the card magically improve my peak performance since the performance difference is non-existant when I don't even have a PhysX card?
There is nothing staggering that the PhysX card can do that games implementing there own (Havoc?) physics engines can't already do. Perhaps the cloth effects are nice, but is that really reason to blow nearly £100? I think not. Even effects such as wind have already been done in STALKER.
Just the first level on the single player campaign, maybe that isn't enough to be conclusive about the PhysX card, but the game is so damn awful I wouldn't consider buying the game itself let alone a PhysX card for it. I guess that's pretty conclusive in itself, though.
Last edited:
Yep i thought it was only the normal maps.
Get the real PhysX map loaded then show us what your cpu can do.
“As for my results; I basically sat on the very first level with Fraps open, then checked what my framerate was at with and without high/extreme PhsyX enabled.”
You sat on the first level? So you basically took your FPS measurements standing still!!!!!! I wonder why you got no performance difference.
You also need to quite and reload the game to activate high/extreme PhsyX.
Results and your conclusion based on those results are invalid and wrong if all you did was stand still or just walk around the starting area.
“How would having the card magically improve my peak performance since the performance difference is non-existant when I don't even have a PhysX card?”
Apart from the fact it looks like you didn’t use high/extreme PhysX so a FPS difference wouldn’t have happened, it doesn’t matter.
Your CPU is still doing work say 20% of the core is doing physics. With a PPU that 20% can be put onto none physic work. 20% is just an example the real number change’s constantly. Anyway if you CPU is not doing phsyics work then other things can be done faster. So even if the performance difference is non-existant between low and high physics the PPU can still boost FPS over the CPU.
“I find it hard to believe having a PhysX card would boost performance in GRAW by 40%
It’s not a constant 40% all the time, only when high amounts of physics are going off. Standing still for example would be pretty much 0% perhaps 1% max.
“I have an E6850 running at 3.8 Ghz which, I would think, should be more than capable of rendering a bunch of black squares around the screen,”
Well the PPU does far more then just black squares around the screen.
You sat on the first level? So you basically took your FPS measurements standing still!!!!!! I wonder why you got no performance difference.
You also need to quite and reload the game to activate high/extreme PhsyX.
Results and your conclusion based on those results are invalid and wrong if all you did was stand still or just walk around the starting area.
“How would having the card magically improve my peak performance since the performance difference is non-existant when I don't even have a PhysX card?”
Apart from the fact it looks like you didn’t use high/extreme PhysX so a FPS difference wouldn’t have happened, it doesn’t matter.
Your CPU is still doing work say 20% of the core is doing physics. With a PPU that 20% can be put onto none physic work. 20% is just an example the real number change’s constantly. Anyway if you CPU is not doing phsyics work then other things can be done faster. So even if the performance difference is non-existant between low and high physics the PPU can still boost FPS over the CPU.
“I find it hard to believe having a PhysX card would boost performance in GRAW by 40%
It’s not a constant 40% all the time, only when high amounts of physics are going off. Standing still for example would be pretty much 0% perhaps 1% max.
“I have an E6850 running at 3.8 Ghz which, I would think, should be more than capable of rendering a bunch of black squares around the screen,”
Well the PPU does far more then just black squares around the screen.
Last edited:
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
lol pottsey, for i dunno, 6 months in posts with people asking about ppu's i every time ask you what games have these increases and every time you say, i've said it all before, go look at old posts, i won't say it again.
graw 1/2 do almost nothing with physics, nothing that is embedded into the actual game engine, it adds a bunch of extra particles into say something exploding already. theres nothing to say that its realistic, its just different, its not better, just different. theres nothing thats massively prettier, or better. from what i've seen theres no 1st level (game dependant) physics intergration whatsoever. its all 2nd level "addition of extra effects", an explosion blowing up yellow after you hit it with 15 bits blown about gives the same level of realism as the same effect with the smoke ever so slightly more realistic and 30 bits blown away.
again for 2 years ageia spent the first umpteen months saying, woohoo graw 1, it will all be about graw 1. it came out, everyone realised it was poop. all of a massive sudden ageia backtracked and said it was just an early game, not much could be added at this late stage. wooo our tech demo game, i forget what its called. the thing where you jump about and can use jedi, opps, i mean magic "force" powers to push pull things and what not. that is a tech demo and NOT a game. graw 2 is it has a SINGLE PPU only level, can't name one? its their two, up till now most advertised and pimped as being PPU enabled games, yet you couldn't name one. deffo sounds to me like you're being unbiased and fair.
the post somewhere above summed it up perfectly. there are hundreds of people on these forums that went from a dual core to a quad core with ZERO gaming benefit(next to, COH maybe, flight sim or two and sup commander) for a couple hundred quid. plenty of people upgraded from a 8800gtx to a 8800ultra for significant cost for no real reason at all, yet none of them, bar 0.00000000001% are buying PPU's.
not a single review should great performance benefit OR massive improvements in game experience. the ONLY thing they've shown is a couple of games with extremely marginal visual gains for an absurdly high performance costs, and in a couple of those very limited situations the PPU shows a performance gain. big whoop.
its looking again, for absolutely all info available that UT3's PPU action will be limited to again, a single usable level, while the rest of the game does NOT use it.
again, you put words in my mouth, i didn't say gpu's do "exactly" what the ppu does better than the PPU. YOU said i said that. I said a GPU could get the same result, through different code optimised for that gpu just as well.
as for code, yes physics IS a code, its a completely non differing set of equations, if the ppu just does physics it just does one set of code.
where to draw stuff on the screen takes a massive amount of physic power. every single pixel, all millions of them every single frame, have physics calculations performed to determine what each one does.
every thread turns the same, you say its useful and you'll give info unbiased, then you randomly add in performance figures, someone asks you where from. you say you've told everyone a million times, despite no one being able to find a review showing anything whatsoever to show the PPU gives great performance or an increase in any real quality in any game, but you assure us they are out there.
graw 1/2 do almost nothing with physics, nothing that is embedded into the actual game engine, it adds a bunch of extra particles into say something exploding already. theres nothing to say that its realistic, its just different, its not better, just different. theres nothing thats massively prettier, or better. from what i've seen theres no 1st level (game dependant) physics intergration whatsoever. its all 2nd level "addition of extra effects", an explosion blowing up yellow after you hit it with 15 bits blown about gives the same level of realism as the same effect with the smoke ever so slightly more realistic and 30 bits blown away.
again for 2 years ageia spent the first umpteen months saying, woohoo graw 1, it will all be about graw 1. it came out, everyone realised it was poop. all of a massive sudden ageia backtracked and said it was just an early game, not much could be added at this late stage. wooo our tech demo game, i forget what its called. the thing where you jump about and can use jedi, opps, i mean magic "force" powers to push pull things and what not. that is a tech demo and NOT a game. graw 2 is it has a SINGLE PPU only level, can't name one? its their two, up till now most advertised and pimped as being PPU enabled games, yet you couldn't name one. deffo sounds to me like you're being unbiased and fair.
the post somewhere above summed it up perfectly. there are hundreds of people on these forums that went from a dual core to a quad core with ZERO gaming benefit(next to, COH maybe, flight sim or two and sup commander) for a couple hundred quid. plenty of people upgraded from a 8800gtx to a 8800ultra for significant cost for no real reason at all, yet none of them, bar 0.00000000001% are buying PPU's.
not a single review should great performance benefit OR massive improvements in game experience. the ONLY thing they've shown is a couple of games with extremely marginal visual gains for an absurdly high performance costs, and in a couple of those very limited situations the PPU shows a performance gain. big whoop.
its looking again, for absolutely all info available that UT3's PPU action will be limited to again, a single usable level, while the rest of the game does NOT use it.
again, you put words in my mouth, i didn't say gpu's do "exactly" what the ppu does better than the PPU. YOU said i said that. I said a GPU could get the same result, through different code optimised for that gpu just as well.
as for code, yes physics IS a code, its a completely non differing set of equations, if the ppu just does physics it just does one set of code.
where to draw stuff on the screen takes a massive amount of physic power. every single pixel, all millions of them every single frame, have physics calculations performed to determine what each one does.
every thread turns the same, you say its useful and you'll give info unbiased, then you randomly add in performance figures, someone asks you where from. you say you've told everyone a million times, despite no one being able to find a review showing anything whatsoever to show the PPU gives great performance or an increase in any real quality in any game, but you assure us they are out there.
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
http://www.ageia.com/physx/ut3.html
this pretty much sums up what i've been saying. pretty but limited to a single level and all the cool bits are nowt to do with the PPU. the problem with making a full game like this is simply, design time. having a PPU doesn't suddenly mean that every building is destructable, the problem is 1st level physics(the game itself and how it reacts and just is) is simply down to designer time. every one of those buildings, and every wall can be made quickly as a single solid objects that don't react to anything properly. To make it destructable only requires that the designer make it 500 single small pieces instead of one massive long wall for instance. but that means 500 edges to design, render, make sure they pull apart properly and so on, which takes, have a guess, 500 times longer. its nothing to do with PPU, assign one massive team for an age and you get one small map that is really interactable, but make a game on the same level of detail and it will take 10 years to come out and be obsolete before it makes it out.
the "physics" in the demo were actually fairly dodgey. wall panels , when a rocket explodes next to them don't exactly react as you would expect at all. because again, its not the physics, its the design of the panels. its incredibly complex and hard to design. we're slowly getting more interactive games as design teams are getting larger for every new game, budgets are expanding massively and experience is easier to come by in the industry, added to new engines designed from the ground up with these things in mind.
this is the main and core problem with ppu's, they are designed to cope with lots and lots of particles moving realistically. the problem is, no one cares about particle count, we care about being able to shoot a rocket into a wall and see any kind of realistic interaction. estimation, or the usual prescripted outcome looks THE SAME as the supposedly "realistic" outcome. i don't want to see lots more particles from the few boards in levels that are prescripted to be destructable. we want to see lots more of the level destructable in the first place.
PS, from what i can remember the tornado's in Alan wake looked better, and did the same stuff, yet doesn't need a PPU(last time i checked anyhow).
this pretty much sums up what i've been saying. pretty but limited to a single level and all the cool bits are nowt to do with the PPU. the problem with making a full game like this is simply, design time. having a PPU doesn't suddenly mean that every building is destructable, the problem is 1st level physics(the game itself and how it reacts and just is) is simply down to designer time. every one of those buildings, and every wall can be made quickly as a single solid objects that don't react to anything properly. To make it destructable only requires that the designer make it 500 single small pieces instead of one massive long wall for instance. but that means 500 edges to design, render, make sure they pull apart properly and so on, which takes, have a guess, 500 times longer. its nothing to do with PPU, assign one massive team for an age and you get one small map that is really interactable, but make a game on the same level of detail and it will take 10 years to come out and be obsolete before it makes it out.
the "physics" in the demo were actually fairly dodgey. wall panels , when a rocket explodes next to them don't exactly react as you would expect at all. because again, its not the physics, its the design of the panels. its incredibly complex and hard to design. we're slowly getting more interactive games as design teams are getting larger for every new game, budgets are expanding massively and experience is easier to come by in the industry, added to new engines designed from the ground up with these things in mind.
this is the main and core problem with ppu's, they are designed to cope with lots and lots of particles moving realistically. the problem is, no one cares about particle count, we care about being able to shoot a rocket into a wall and see any kind of realistic interaction. estimation, or the usual prescripted outcome looks THE SAME as the supposedly "realistic" outcome. i don't want to see lots more particles from the few boards in levels that are prescripted to be destructable. we want to see lots more of the level destructable in the first place.
PS, from what i can remember the tornado's in Alan wake looked better, and did the same stuff, yet doesn't need a PPU(last time i checked anyhow).
Pottersy you're flogging a dead horse here. I see what you're saying in that they are not completely worthless.
However, as technically accomplished as they may be (assuming they are) they are in no way worth their asking price. Being worthless is different from being useless. A £1,000,000 Veyron is not worthless but is useless if you're standed in the middle of the ocean.
Similarly a £100 physics card is useless if there are no tittles that REQUIRE it or gain (in the majority of the educated publics opinion) significantly from it.
As such to 99.99% of us it's useless.
However, as technically accomplished as they may be (assuming they are) they are in no way worth their asking price. Being worthless is different from being useless. A £1,000,000 Veyron is not worthless but is useless if you're standed in the middle of the ocean.
Similarly a £100 physics card is useless if there are no tittles that REQUIRE it or gain (in the majority of the educated publics opinion) significantly from it.
As such to 99.99% of us it's useless.
Soldato
- Joined
- 17 Aug 2003
- Posts
- 20,167
- Location
- Woburn Sand Dunes
Funny that you demand benchmarks when that is exactly what you are unable to post
pottsey posts everything he can in order for peoepl like you to take notice rather than come out with crap like that. evidently it doesnt work, does it?
Funny that you demand benchmarks when that is exactly what you are unable to post
Pottsey has posted more benchmarks/links than all of the naysayers have posted together.
Do your home work & use the now working again search feature as im sure Pottsey is tired of repeating links & benchmarks.
Wow, that's great, you took everything I said completely out of context for your own personal gain/enjoyment. You actually think I literally stood still when trying to establish an idea of the performance difference? You must be joking. I played the entire level with PhysX off and with them on. And I know you have to restart in order to enable them. Don't you think I would notice if I hadn't turned the effects on?
I didn't try the "Ageia level" or whatever was previously mentioned, and to be quite honest, why would I? Even if the "PhysX map" did kill my CPU, who cares? I don't buy cards for benchmarks or tech maps, the real single player campaign plays exactly the same on my system with high PhysX on as it does without, thus, if I wanted to play the game with PhysX, I don't need the add on card!
And you're right, the PhysX card does do more than black squares, it also does awkward cloth and water effects. (Yes, I have checked tech demos that come with the drivers)
I'm sure the PhysX card does help in a few obscure titles out there, and I'm sure they can provide pretty tech demos of what it can do, but even though I have plenty of money to throw around on whatever the hell I want for my PC, I still draw the line when it comes to wasting it on a card that really, in it's current state, does not benefit my gaming experience 98.999% of the time.
Last edited:
Well you have never used one so you wouldn't know.