-
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.
PPU more then 100% faster then CPU in UT3 benchmark here.
- Thread starter Pottsey
- Start date
Man of Honour
- Joined
- 17 Oct 2007
- Posts
- 3,871
Interesting Discussion.
Hey Pottsey, referring to the graph below:
It's quite obvious that a Non-Physx map is much better to play on than a Phsx Map even if you are using a Physics processing unit (PPU). Now are these Phsx Maps in games such as Unreal Tournament already encoded into the game in the sense off, you can't avoid them during the time you are playing the game?
Am I right in thinking that on a Phsx Maps, you get all of the physics effects regardless if you have a Physics processing unit or not, the PPU is primary there to boost the frames per second?
I am not sure how much of a difference a Phsx Map make's as regards to the Physic effects. Now i would say that most people who play's games are happy with the visuals and effects of the game as a whole, so in my opinion, surely it would be much better to do away with the whole Physics processing unit and Phsx Maps (At the moment) until the performance on a Phsx Map is much nearer to that of a Non-Physx map with or without a PPU?
How much does a Physics processing unit boost performance as regards to frames per second on a Non-Physx map?
It's quite clear that a PPU boosts performance on a Phsx Map from the graph above, however, to gain them extra frames per second you will have to obviously purchase a Physics processing unit for example the BFG Ageia PhysX Accelerator which costs £88.11, which is quite a lot of money for them extra frames per second. Surely a lot of people would rather get around the 50 frames per second mark with the loss of the Physic effects then to get around half of that with Physic effects. Now in that sense, they are now worth buying. (This obviously depends, how much of a difference the Physic effects make during the game play and concerning the loss of frames per second, is it still playable at an enjoyment rate. Now I can't comment on this side of things since I have never had first hand experience of this)
Hey Pottsey, referring to the graph below:
It's quite obvious that a Non-Physx map is much better to play on than a Phsx Map even if you are using a Physics processing unit (PPU). Now are these Phsx Maps in games such as Unreal Tournament already encoded into the game in the sense off, you can't avoid them during the time you are playing the game?
Am I right in thinking that on a Phsx Maps, you get all of the physics effects regardless if you have a Physics processing unit or not, the PPU is primary there to boost the frames per second?
I am not sure how much of a difference a Phsx Map make's as regards to the Physic effects. Now i would say that most people who play's games are happy with the visuals and effects of the game as a whole, so in my opinion, surely it would be much better to do away with the whole Physics processing unit and Phsx Maps (At the moment) until the performance on a Phsx Map is much nearer to that of a Non-Physx map with or without a PPU?
How much does a Physics processing unit boost performance as regards to frames per second on a Non-Physx map?
It's quite clear that a PPU boosts performance on a Phsx Map from the graph above, however, to gain them extra frames per second you will have to obviously purchase a Physics processing unit for example the BFG Ageia PhysX Accelerator which costs £88.11, which is quite a lot of money for them extra frames per second. Surely a lot of people would rather get around the 50 frames per second mark with the loss of the Physic effects then to get around half of that with Physic effects. Now in that sense, they are now worth buying. (This obviously depends, how much of a difference the Physic effects make during the game play and concerning the loss of frames per second, is it still playable at an enjoyment rate. Now I can't comment on this side of things since I have never had first hand experience of this)
Last edited:
Source?
If the PPU is meant to do all the physics then what is the CPU doing on PPU maps that it's not doing in non-PPU maps? You're showing some faulty logic here. Without a PPU those maps are obviously CPU limited, and with one they're obviously PPU limited. Neither apply to normal maps, at least not in a way that halves performance.
That's not what graphs on Anandtech show supposedly, nor does it happen with users on this forum who have posted their own task manager graphs. Who do you expect me to believe, Anandtech and users here or a site I've never heard of before today?
Last edited:
for the love of for **** sake, can we not just accept the PPU is pointless and move on, how many bleeding threads are there about how PPUs increase game performance, and in the end it turns out they actually slow it down, come on guys, seriously getting really old, really fast 
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
well, bit-tech say that the game changes for the worse on the ppu only maps, with as i stated, THEY said the explosions become clouds of particles which you can move under, despite realistically, if you shoot something it won't become a cloud of splinters that remain in the air that long. but thats just me. also some things become breakable, but in a arbitrary way. some walls take damage, others don't. which again backs up what i've been saying. it takes TIME to make things breakable. the ppu and ageia physics software doesn't just create the ability to design an entire map thats breakable instantly, they simply have to take time making every last bit breakable. according to them again, some bits of wall that look identical to others break, and some bits don't break.
if it all doesn't break, or it all breaks theres some continuity, when its arbitrary, it feels like someones not had time to make a whole level destructable.
no, in every thread i've ask for benchmarkS you haven't posted anything, you've repeatedly stated you wouldn't as you have before. if i leave a thread you posted benchmarks after you said i wouldn't, not exactly my fault. the benchmark isn't very indicative tbh. i play COH/COV now and then, been playing it again recently. frankly, top physics and particle counts add nothing useful to the game, and even now on very good hardware it doesn't run all that well. but yes, you've shown us one game that runs faster with physx, any others, a whole review showing how good it is across multiple games?
you seemed to be indicating that the ppu enabled ut3 to increase framerate in the non ppu maps because across this thread people have said higher fps in non ppu maps is preferable to crappy fps in the ppu maps even with a ppu. you responded several times to this by stating it was faster in non ppu maps. i looked around for a review to question that, couldn't find anything, you said it several times, and now finally put up a benchmark for a different game. which is fine, i wasn't specific, but i thought the generality of this thread was talking about UT3, and i thought my question implied that i wanted to see something as such in UT3 benchmarking.
as for the tornado, why does a tornado in a map that runs awfully , and damages only certain parts of the map represent better physics than a much more evenly destructable maps in crysis. its a tornado, its a flying circle of wind, it might look pretty, but, where exactly is the proof that its impressive physics? it also can't run the tornado in great framerate.
stop telling me what a tech demo is, a tech demo is something that shows off a few things, it can be playable, theres no standard to say exactly what a tech demo is. but when 99.9% of people regard something as a pretty way to show off a few features of a piece of hardware, without much real gameplay, fun, or any real lastability, yes, i DO call that a tech demo. 3dmark is a tech demo or sorts, if 3dmark were playable, it would still be a tech demo, even if people still played it now and then 5 years later, i'd still call it a tech demo.
if it all doesn't break, or it all breaks theres some continuity, when its arbitrary, it feels like someones not had time to make a whole level destructable.
no, in every thread i've ask for benchmarkS you haven't posted anything, you've repeatedly stated you wouldn't as you have before. if i leave a thread you posted benchmarks after you said i wouldn't, not exactly my fault. the benchmark isn't very indicative tbh. i play COH/COV now and then, been playing it again recently. frankly, top physics and particle counts add nothing useful to the game, and even now on very good hardware it doesn't run all that well. but yes, you've shown us one game that runs faster with physx, any others, a whole review showing how good it is across multiple games?
you seemed to be indicating that the ppu enabled ut3 to increase framerate in the non ppu maps because across this thread people have said higher fps in non ppu maps is preferable to crappy fps in the ppu maps even with a ppu. you responded several times to this by stating it was faster in non ppu maps. i looked around for a review to question that, couldn't find anything, you said it several times, and now finally put up a benchmark for a different game. which is fine, i wasn't specific, but i thought the generality of this thread was talking about UT3, and i thought my question implied that i wanted to see something as such in UT3 benchmarking.
as for the tornado, why does a tornado in a map that runs awfully , and damages only certain parts of the map represent better physics than a much more evenly destructable maps in crysis. its a tornado, its a flying circle of wind, it might look pretty, but, where exactly is the proof that its impressive physics? it also can't run the tornado in great framerate.
stop telling me what a tech demo is, a tech demo is something that shows off a few things, it can be playable, theres no standard to say exactly what a tech demo is. but when 99.9% of people regard something as a pretty way to show off a few features of a piece of hardware, without much real gameplay, fun, or any real lastability, yes, i DO call that a tech demo. 3dmark is a tech demo or sorts, if 3dmark were playable, it would still be a tech demo, even if people still played it now and then 5 years later, i'd still call it a tech demo.
for the love of for **** sake, can we not just accept the PPU is pointless and move on, how many bleeding threads are there about how PPUs increase game performance, and in the end it turns out they actually slow it down, come on guys, seriously getting really old, really fast
If people want to talk about them, which it seems in this thread, let them
Because it doesn't increase FPS by any amount that makes it useful. Most people already get an acceptable FPS in the game anyway, so adding a PPU will make no difference. Why pay money to get more FPS that you won't even notice? To top it all off you don't even get any extra effects.
I don't understand how you can't see what we are all trying to say. The PPU is wasted. Can't you see it is totally pointless in almost all situations? There is no reason to have one. It serves no purpose. It makes no difference. Therefore it is a waste of time. Can't you see why we are saying what we are?
Last edited:
“no, in every thread i've ask for benchmarkS you haven't posted anything, you've repeatedly stated you wouldn't as you have before. if i leave a thread you posted benchmarks after you said i wouldn't, not exactly my fault.”
More lies. Lots of people can back me up on that.
http://forums.overclockers.co.uk/showpost.php?p=9079066&postcount=46
http://forums.overclockers.co.uk/showpost.php?p=9081872&postcount=53
You asked me for a benchmark. I posted. Then you came back into more threads after saying I never post benchmarks. I post more, you say I don’t The process keeps going.
There was one point about 2 or 3 months back that I got fed up and refused to post any more as you ignored all the others.
“which is fine, i wasn't specific, but i thought the generality of this thread was talking about UT3, and i thought my question implied that i wanted to see something as such in UT3 benchmarking.”
As I said many times I don’t know how to do UT3 benchmarking. If someone explains how, I will run tests. Been searching on yahoo but not found anything. I also said before I want to see some decent UT 3 becnmarks on none PPU maps.
“Because it doesn't increase FPS by any amount that makes it useful.”
How do you know this, please prove it. Or tell me how to run a benchmark in UT3. I would love to see what’s true about the FPS score. The only data we have is an amount thats usefull. But I want more persice data.
“It makes no difference. Therefore it is a waste of time. Can't you see why we are saying what we are?”
You keep telling me this but in my games I did benchmark its boosting FPS, so it makes a difference and isn’t a waste of time. Why can you not see this?
“How much does a Physics processing unit boost performance as regards to frames per second on a Non-Physx map?”
No idea precisely. Been trying to find out how to run a fair timedemo benchmark. But no one can tell me. According to other PPU owners its about 15fps faster. But I would prefer a benchmark over screenshot numbers.
“If the PPU is meant to do all the physics then what is the CPU doing on PPU maps that it's not doing in non-PPU maps?”
Processing extra data to send to the GPU. It’s a lot harder to render a roof being turn into lots of flying bits along with all the other rubble blown about then it is one big flat roof. There is a lot of extra none physics work for the CPU on the PPU maps.
You're showing some faulty logic here. Without a PPU those maps are obviously CPU limited, and with one they're obviously PPU limited. Neither apply to normal maps, at least not in a way that halves performance.”
It’s not faulty logic. How are they obviously PPU limited? With the CPU at 100% it’s clear its CPU limited and the PPU boosts FPS by over 100%. A 100% boost doesn’t sound PPU limited. It looks like the CPU cannot provide enough data to the GPU to render all the extra bits. Everything points to being CPU limited which is holding back the GPU which means low FPS.
“That's not what graphs on Anandtech show supposedly, nor does it happen with users on this forum who have posted their own task manager graphs.”
Can I have a link please not seen it my self, not seen any of the PPU maps on Anandtech. What post number on here says different? I must have missed it.
“for the love of for **** sake, can we not just accept the PPU is pointless and move on, how many bleeding threads are there about how PPUs increase game performance, and in the end it turns out they actually slow it down”
What is getting old are people like you coming out with BS. It’s not useless, its boosts FPS and there isn’t a single recent game showing the PPU slowing down the game.
“well, bit-tech say that the game changes for the worse on the ppu only maps, with as i stated, THEY said the explosions become clouds of particles which you can move under, despite realistically, if you shoot something it won't become a cloud of splinters that remain in the air that long.”
Well they lied and/or are wrong like the other findings. It’s the same with or without the PPU. I see no difference playing the game and I don’t recall anyone else saying there is a difference. None of other reviews say its diffrent do they?
“stop telling me what a tech demo is, a tech demo is something that shows off a few things, it can be playable, theres no standard to say exactly what a tech demo is.”
There is a standard. “A technology demo is a prototype, rough example or an otherwise incomplete version of a product, put together with the primary purpose of showcasing the idea,”
A complete version of something cannot be a tech demo. Using your strange definition Crysis is a tech demo. So is Half Life EP 1 and Portal and many others. tech demos are just that demos. A full game is not a demo just because you dont like it and play it once.
More lies. Lots of people can back me up on that.
http://forums.overclockers.co.uk/showpost.php?p=9079066&postcount=46
http://forums.overclockers.co.uk/showpost.php?p=9081872&postcount=53
You asked me for a benchmark. I posted. Then you came back into more threads after saying I never post benchmarks. I post more, you say I don’t The process keeps going.
There was one point about 2 or 3 months back that I got fed up and refused to post any more as you ignored all the others.
“which is fine, i wasn't specific, but i thought the generality of this thread was talking about UT3, and i thought my question implied that i wanted to see something as such in UT3 benchmarking.”
As I said many times I don’t know how to do UT3 benchmarking. If someone explains how, I will run tests. Been searching on yahoo but not found anything. I also said before I want to see some decent UT 3 becnmarks on none PPU maps.
“Because it doesn't increase FPS by any amount that makes it useful.”
How do you know this, please prove it. Or tell me how to run a benchmark in UT3. I would love to see what’s true about the FPS score. The only data we have is an amount thats usefull. But I want more persice data.
“It makes no difference. Therefore it is a waste of time. Can't you see why we are saying what we are?”
You keep telling me this but in my games I did benchmark its boosting FPS, so it makes a difference and isn’t a waste of time. Why can you not see this?
“How much does a Physics processing unit boost performance as regards to frames per second on a Non-Physx map?”
No idea precisely. Been trying to find out how to run a fair timedemo benchmark. But no one can tell me. According to other PPU owners its about 15fps faster. But I would prefer a benchmark over screenshot numbers.
“If the PPU is meant to do all the physics then what is the CPU doing on PPU maps that it's not doing in non-PPU maps?”
Processing extra data to send to the GPU. It’s a lot harder to render a roof being turn into lots of flying bits along with all the other rubble blown about then it is one big flat roof. There is a lot of extra none physics work for the CPU on the PPU maps.
You're showing some faulty logic here. Without a PPU those maps are obviously CPU limited, and with one they're obviously PPU limited. Neither apply to normal maps, at least not in a way that halves performance.”
It’s not faulty logic. How are they obviously PPU limited? With the CPU at 100% it’s clear its CPU limited and the PPU boosts FPS by over 100%. A 100% boost doesn’t sound PPU limited. It looks like the CPU cannot provide enough data to the GPU to render all the extra bits. Everything points to being CPU limited which is holding back the GPU which means low FPS.
“That's not what graphs on Anandtech show supposedly, nor does it happen with users on this forum who have posted their own task manager graphs.”
Can I have a link please not seen it my self, not seen any of the PPU maps on Anandtech. What post number on here says different? I must have missed it.
“for the love of for **** sake, can we not just accept the PPU is pointless and move on, how many bleeding threads are there about how PPUs increase game performance, and in the end it turns out they actually slow it down”
What is getting old are people like you coming out with BS. It’s not useless, its boosts FPS and there isn’t a single recent game showing the PPU slowing down the game.
“well, bit-tech say that the game changes for the worse on the ppu only maps, with as i stated, THEY said the explosions become clouds of particles which you can move under, despite realistically, if you shoot something it won't become a cloud of splinters that remain in the air that long.”
Well they lied and/or are wrong like the other findings. It’s the same with or without the PPU. I see no difference playing the game and I don’t recall anyone else saying there is a difference. None of other reviews say its diffrent do they?
“stop telling me what a tech demo is, a tech demo is something that shows off a few things, it can be playable, theres no standard to say exactly what a tech demo is.”
There is a standard. “A technology demo is a prototype, rough example or an otherwise incomplete version of a product, put together with the primary purpose of showcasing the idea,”
A complete version of something cannot be a tech demo. Using your strange definition Crysis is a tech demo. So is Half Life EP 1 and Portal and many others. tech demos are just that demos. A full game is not a demo just because you dont like it and play it once.
Associate
- Joined
- 6 Nov 2005
- Posts
- 173
How can you say they are obviously limited by the PPU when you have the CPU running at 100% and have no idea what the usage of the PPU is?
I have given you some examples above of the potential additional load on the cpu when you start adding more objects and effects to a scene. This could explain why the framerate is not as high as none ppu maps and explain the 100% load on the CPU.
It may well be limited by the PPU but you cannot categorically say that as we do not have enough evidence.
Have you looked at the graphs on Anandtech?
They show the performance of the game improving when you add a faster CPU or a CPU with more cores which indicates some form of CPU limitation, however they also show an increase when you add a faster graphics card which shows it must be GPU limited at some points as well. I cannot see whether the framerates shown are the max min or average. Assuming they are the average you would expect a better GPU and CPU to both improve performance as it is very unlikley that the game is either 100% CPU limited or 100% GPU limited. An interesting point is that the article does say that the nvidia cards appear to be cpu limited at a resolution below 1600x1200.
Who has posted graphs of their cpu usage? I remember seeing metalmackeys percentage which showed it was not cpu limited on a quad core. The point is the graphs Pottsey posted showed the results of a dual core. If you take metalmackeys percentages of 60, 60 , 80 and 18 you can see that if you put the same work on a dual core you will have both cores running at about 100% most of the time which backs up the article.
Associate
- Joined
- 6 Nov 2005
- Posts
- 173
That's true but if you look at the chart of the CPU usage on a non physx map in the article you linked to it is also pretty high and hitting 100% on a lot of occasions. This indicates that it is CPU limited at times in the non PPU maps when only using a dual core.
I have also seen a few posts on other forums from people running AMD dual cores where they are saying it is CPU limited for them.
So you're saying the PPU effects put unnecessary strain on the CPU even with a PPU in there?
That would definitely explain the horrible frame rates. GG Ageia API.
See above. P.S. I love how you skipped my source, I still want it.
So you too are saying that PPU effects put unnecessary stress on the CPU?
Last edited:
Associate
- Joined
- 28 Jun 2005
- Posts
- 2,174
- Location
- Behind you
@marc I aint no god like programmer or anything but surely the AI woundnt increase the cpu by that much. I could understand if the fps were about the same, but half? I doubt there is much more AI but even if there is the CPU should cope seeing it doesnt have to worry about collision detection with the extra objects. I aint really used the API that much yet but couldnt the PPU do some of the AI using things like triggers?
Associate
- Joined
- 6 Nov 2005
- Posts
- 173
Where did I say it was unnecessary stress?
By offloading work to the PPU you are freeing up CPU cycles which means you should in theory be able to have more complicated AI routines and have the CPU power to make additional draw calls and render more objects on screen. That is assuming the GPU can handle the additional work. The problem is if you try to add too much to the scene you may put too much load on either the CPU or GPU. It is a balancing act just as it is when you only have a CPU and GPU.
It appears to me that in the PPU levels that have been designed so far they have got this balancing act wrong. At least it looks that way when you only have a dual core cpu. It would be interesting to see the results if you paired a quad core with a PPU.
You didn't, I did. And it is if you're losing half your framerate for effects that make a difference to gameplay.
I thought a PPU was supposed to take all the physics and stuff off the CPU, Ageia need to change their advertising because they've seemingly left us all grossly misinformed.
Last edited:
“So you're saying the PPU effects put unnecessary strain on the CPU even with a PPU in there?”
Yes sometimes.
There are 3 ways to use a PPU.
* Don’t add new effects get a FPS boost (UT normal maps, PPU takes strain off CPU)
* Add a few new effects so your FPS stays the same. (COH style PPU adds effects and takes strain off CPU)
* Add lots of new effects so you lose FPS. (UT bonus maps, PPU puts strain on CPU)
If you add lots of effects you take physics strain off the CPU and put it on the PPU. But you then add graphic strain onto the CPU and GPU.
There isn’t a single source it something I said and came up with from the facts we have. So I guess the source is me.
Increase physics past default and CPU usage goes up. During increase physics useage FPS drops down to sub 1fps when lots of physics happen at once, like half a forest is knocked down. Only conclusion you can get from that is the CPU cannot handle the extra phsyics so your CPU limited. If your back is turned away from the physics there is zero works for the GPU so its not being GPU limited.
http://www.youtube.com/watch?v=mxOSOJSmess&feature=related video evidence. You get the same 1fps without recoding video. So the video recoding isnt to blame.
http://www.crymod.com/thread.php?threadid=9226&threadview=0&hilight=nuke&hilightuser=0
“TIP - Keep the Physics radius fairly low, as this drags the performance down trying to calculate the destruction of too many trees.”
“Based on the performance drop, I'd say Crytek have disabled those effects because computers aren't able to handle them just yet.”
“Okay, I did this and it completely, and I mean COMPLETELY, rapes your PC.”
There forum is full of people going on about physics on a mass scale don’t work with today’s CPU’s.
“I thought a PPU was supposed to take all the physics and stuff off the CPU, Ageia need to change their advertising because they've seemingly left us all grossly misinformed.”
It does take all physics off the CPU. Why do they need to change there advertising?
Ageia aways said there are 3 ways to use a PPU. Not all boost speed.
Yes sometimes.
There are 3 ways to use a PPU.
* Don’t add new effects get a FPS boost (UT normal maps, PPU takes strain off CPU)
* Add a few new effects so your FPS stays the same. (COH style PPU adds effects and takes strain off CPU)
* Add lots of new effects so you lose FPS. (UT bonus maps, PPU puts strain on CPU)
If you add lots of effects you take physics strain off the CPU and put it on the PPU. But you then add graphic strain onto the CPU and GPU.
There isn’t a single source it something I said and came up with from the facts we have. So I guess the source is me.
Increase physics past default and CPU usage goes up. During increase physics useage FPS drops down to sub 1fps when lots of physics happen at once, like half a forest is knocked down. Only conclusion you can get from that is the CPU cannot handle the extra phsyics so your CPU limited. If your back is turned away from the physics there is zero works for the GPU so its not being GPU limited.
http://www.youtube.com/watch?v=mxOSOJSmess&feature=related video evidence. You get the same 1fps without recoding video. So the video recoding isnt to blame.
http://www.crymod.com/thread.php?threadid=9226&threadview=0&hilight=nuke&hilightuser=0
“TIP - Keep the Physics radius fairly low, as this drags the performance down trying to calculate the destruction of too many trees.”
“Based on the performance drop, I'd say Crytek have disabled those effects because computers aren't able to handle them just yet.”
“Okay, I did this and it completely, and I mean COMPLETELY, rapes your PC.”
There forum is full of people going on about physics on a mass scale don’t work with today’s CPU’s.
“I thought a PPU was supposed to take all the physics and stuff off the CPU, Ageia need to change their advertising because they've seemingly left us all grossly misinformed.”
It does take all physics off the CPU. Why do they need to change there advertising?
Ageia aways said there are 3 ways to use a PPU. Not all boost speed.
That's all you had to say. It's an £88 framerate halving card, at least in UT3.
Associate
- Joined
- 6 Nov 2005
- Posts
- 173
It's not just additional AI. One of the biggest problems with DX9 is the large CPU overhead of making a draw call and changing the state of the DX9 device. These are things such as changing textures, shaders, rendertargets etc. One solution to this is instead of sending one object at a time to the GPU the vertices that make up these objects are batched together and sent as one draw call. Obviously you can only batch the vertices that use the same textures and shaders. Therefore if you add additional objects you are likely to increase the number of draw calls at least slightly. However even if don't increase the draw calls by many the CPU will still need to undertake additional work to batch the additional vertices.
On the subject of AI, I have only seen some brief videos of the PPU levels of UT3 so this may not all apply but assuming you have a destructable terrain which I believe you do you have the situation where some parts of the map that were inaccessable sundenly become accessable. If you have 20 computer characters that are moving around the map they must be able to have the ability to find there way around these areas. Assuming the parts of the wall you have knocked down remain in the environment they may or may not be able to be walked or jumped over depending on how they have come to rest. The computer characters are going to need to be able to calculate whether a path has now become available to them. All of this would lead to more work being placed on the CPU.
Whilst the CPU does not need to cope with the collision detection for the point of view of physics it may do for the computer characters pathfinding routines.
I also haven't used the API much so I am unaware whether this is possible. I also don't have a PPU. It would assume that unlike a GPU the PPU must be able to pass data back into main memory to be used by the CPU so it may be possible for the PPU to do some of this work if the API allows it to.
“See above. P.S. I love how you skipped my source, I still want it.”
I posted back in post 80 post
http://forums.overclockers.co.uk/showpost.php?p=10588985&postcount=80
http://www.pcper.com/images/reviews/491/cpuusage_physxmapwithphysx.gif
“That's all you had to say. It's an £88 framerate halving card, at least in UT3.”
But that’s not true
Plug in PPU and your FPS go up by over 100%. That does not = a framerate halving card.
On normal maps you framerate goes up not down. So again not a framerate halving card.
The card its self never lowers FPS in UT. You either have the card and get a FPS boost or you don’t have the card and get less FPS.
I posted back in post 80 post
http://forums.overclockers.co.uk/showpost.php?p=10588985&postcount=80
http://www.pcper.com/images/reviews/491/cpuusage_physxmapwithphysx.gif
“That's all you had to say. It's an £88 framerate halving card, at least in UT3.”
But that’s not true
Plug in PPU and your FPS go up by over 100%. That does not = a framerate halving card.
On normal maps you framerate goes up not down. So again not a framerate halving card.
The card its self never lowers FPS in UT. You either have the card and get a FPS boost or you don’t have the card and get less FPS.
That's not the source I meant. I meant the one for your claim about Crytek removing effects.
You use the PhysX API and its effects with a PPU in UT3 and you get half the framerate.
Doesn't exactly make me want to whip my wallet out.