PPU more then 100% faster then CPU in UT3 benchmark here.

[Pottsey]

“Pottsey just agreed with me.”
I did not agree the game is PPU limited. Read it again.
Twice you said I am contradicting my self. I don’t see eithe,. Please quote me or give me post numbers

 

[oweneades]

All the maps are optimised for the card. Did you lot not bother to look at the hardware used? They didn’t use high-end hardware.

But I agree some none PPU and PPU scores on normal maps would be nice. With the same graph over time they used.

Did you not bother reading the link you actually posted. The system used is running a Core 2 Extreme X6800 and a 8800GT with 2gb of ram. How is this not high end hardware? The only thing its missing is a quad core and given that this is meant to be showing off ppu physics a quad core shouldn't be needed.

Under 30fps in UT3 is unplayable regardless of the performance increase with a PPU or not (simply given because the gameplay is so quick in comparison to something like crysis).

EDIT: Also please use the quote button when quoting people. Its extermely difficult to see who you are quoting if you just bold the test.

 

[Combat squirrel]

“Er thats total rubbish, my agiea physx PPU makes ~2fps difference to UT3”
I missed it and it appears your guessing. Did you run a timedemo and record the results? How did you benchmark? Did you use the UT 3 PPU drivers?

if you call fraps 'guessing' then go right ahead, makes nearly no difference runs in the high ~60 ish fps area on normal maps, crawls on physx maps, someone says its poss to run these maps with physx disabled, ill try it....................

SEEMS im one of the rare people with a physx card that tells people the truth about how it effects games.

Im using the latest drivers for every bit of hardware in my pc


POTTSEY PLEASE ANSWER THESE QUESTIONS:

1)Why did you not reply to that thread (my UT3 findings thread)then pottsey ?

2)And have you personally played these maps on your computer pottsey ?

3)Please list your systems specs for me

EVERYONE READING THIS THREAD, I HAVE A Physx Card in my PC, in UT 3 IT MAKES NO DIFFERENCE WHAT SO EVER, on the PHYSX ONLY MAPS ITS UNPLAYABLE, 1ST HAND EXPERIENCE, SIMPLE AS THAT - ANY QUESTIONS ?

 

[Marc Fraser]

I consider fanboyism very childish, which is why I love it when I see you and Pottsey get together in a PPU thread. Always a good read.

Please quote me once where I have said people should buy a PPU. I don't even own a PPU so how can I be a fanboy?

The reason I post in these threads is because it annoys me when people post factually incorrect information to degrade a product.

Pottsey just agreed with me.

And what Combat squirrel just posted also seems to indicate that the PPU is not up to the task.

Show me evidence that shows the PPU running at 100% load and bottlenecking the rest of the system.?

The only evidence I have seen is the website Pottsey linked to that showed the CPU at 100% on the PPU maps with a PPU in. If the PPU was the bottleneck it is unlikley that the PPU would be running at 100%.

 

[metalmackey]

Crysis seems to cope with the extra AI fallen trees and destructable walls bring pretty well seeing it has no PPU to off-load to.

 

[Combat squirrel]

OH FOR F SAKE IM GOING TO RUDDY FILM MY PC PLAYING UT3 ON PHYSX WITH FRAPS ON !!! *goes to hunt for something to film it on, be about 30 ish mins

 

[Pottsey]

EDITED
"The only thing its missing is a quad core and given that this is meant to be showing off ppu physics a quad core shouldn't be needed."
The map is CPU limited a QUAD core is needed.




“How is this not high end hardware?”
I consider quad core high end and dual core mid end. As for the GPU I made a mistakes and though the GT was under GTS speed. Didn’t realise it was near GTX.




“if you call fraps 'guessing' then go right ahead,”
How do you fraps the same map?
The 2nd run though all the bots are doing different things which mean different physics which mean different FPS. You’re also doing different things. Unless you know how to do a timedemo, in which case please share.




“1)Why did you not reply to that thread”
I missed it and only post if I have something to add. I find most PPU threads but I do miss some.





“2)And have you personally played these maps on your computer pottsey ?”
Yes and said before I don’t like the bonus maps. To many problems with them.




“3)Please list your systems specs for me”
P4 D 3.2 Ghz (the old ones) , 8800GTS, 4gig, Vista what else do you want?

 

[Lightnix]

Okay, I'm gonna argue one point. Just one.

£88 for a quite frankly pitiful boost in framerate? No thank you.

 

[willhub]

“3)Please list your systems specs for me”
P4 D 3.2 Ghz (the old ones) , 8800GTS, 4gig, Vista what else do you want?

You're CPU is a bottleneck to that GTS.

 

[Marc Fraser]

Crysis seems to cope with the extra AI fallen trees and destructable walls bring pretty well seeing it has no PPU to off-load to.

I am not saying it is definitley down to what I posted I was just giving some examples of what could be causing the 100% CPU usage when using a PPU.

I haven't played crysis yet just seen the videos. but out of interest how much of the terrain is destructable?

When the trees fall do they stay on the ground and become an obstacle or do they disappear?

Do you have to jump over them if they all fall on top of each other or do you push them out of the way if you try to run through them?

 

[Pottsey]

“*goes to hunt for something to film it on, be about 30 ish mins”
Don’t bother unless you’re running a time demo.

Unless you have the same bot path, same player path, same phsyics per run though its not a fair or accurate benchmark. Thats what I want to see as that will settle this one and for all. I want to see a timedemo with a slow, and fast CPU with and without the PPU on both.

 

[Marc Fraser]

if you call fraps 'guessing' then go right ahead, makes nearly no difference runs in the high ~60 ish fps area on normal maps, crawls on physx maps, someone says its poss to run these maps with physx disabled, ill try it....................

SEEMS im one of the rare people with a physx card that tells people the truth about how it effects games.

Im using the latest drivers for every bit of hardware in my pc


POTTSEY PLEASE ANSWER THESE QUESTIONS:

1)Why did you not reply to that thread (my UT3 findings thread)then pottsey ?

2)And have you personally played these maps on your computer pottsey ?

3)Please list your systems specs for me

EVERYONE READING THIS THREAD, I HAVE A Physx Card in my PC, in UT 3 IT MAKES NO DIFFERENCE WHAT SO EVER, on the PHYSX ONLY MAPS ITS UNPLAYABLE, 1ST HAND EXPERIENCE, SIMPLE AS THAT - ANY QUESTIONS ?

When you say unplayable what kind of fps are you getting? Is it similar to the article?

Could you let us know what the CPU usage is like on a PPU map with a PPU please?

 

[Pottsey]

“You're CPU is a bottleneck to that GTS.”
Yep pretty much, which is why I like my PPU. The less physics the CPU is doing the more CPU power goes to my GTS. The PPU means I can play normal UT maps without needing to upgrade my CPU/motherboard. The bonus maps are unplayable.

 

[Combat squirrel]

“if you call fraps 'guessing' then go right ahead,”
How do you fraps the same map?
The 2nd run though all the bots are doing different things which mean different physics which mean different FPS. You’re also doing different things. Unless you know how to do a timedemo, in which case please share.

What are you goin on about, fraps shows the current FPS on screen............like I SAID IM GOING TO FILM IT NOW !

 

[Combat squirrel]

“*goes to hunt for something to film it on, be about 30 ish mins”
Don’t bother unless you’re running a time demo.

Unless you have the same bot path, same player path, same phsyics per run though its not a fair or accurate benchmark. Thats what I want to see as that will settle this one and for all. I want to see a timedemo with a slow, and fast CPU with and without the PPU on both.

You dont need a time demo pottsey to see it runs crud constantly and when the amount of physx stuff going on increases, so dont talk utter rubbish, people will see for real in a real game its useless

 

[Marc Fraser]

What are you goin on about, fraps shows the current FPS on screen............like I SAID IM GOING TO FILM IT NOW !

What Pottsey means is that if you are planning on comparing with and without a ppu using fraps it will be impossible to get the same load on the CPU each time and it would make it an unfair comparison.

If however you are going to fraps a PPU level then please do as it would be good to see what results you are getting. If you are doing this please could you monitor the CPU usage at the same time. If you track the tasks using perfmon we will be able to see how much cpu is taken up by fraps as well.

 

[Pottsey]

“What are you goin on about, fraps shows the current FPS on screen............like I SAID IM GOING TO FILM IT NOW !”
Yes FRAPS shows the current FPS. But unless you can run the same test twice you cannot compare with and without the PPU.

What if the 2nd time you run though with the PPU there are twice as much phsyics going off due to bot actively. What if there is half as much physics going off due to bots? Its an unfair test so there is no point in filming.

You cannot run two different tests with different amount of physics and compare the results to work out the fastest



EDIT:
“You dont need a time demo pottsey to see it runs crud constantly and when the amount of physx stuff going on increases, so dont talk utter rubbish, people will see for real in a real game its useless“
So your saying the normal maps run crud constantly with the PPU as physics go off? if its the bonus maps well we all know they run bad. I get sub 20fps.

 

[dbmzk1]

“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.

Do you have a PPU? If you do try it for yourself and see. You don't need a timedemo to see that you are not getting much of an FPS boost. The amount going on is not going to be THAT much different. Sure it won't be totally accurate, but it will give you a rough guide as to what difference it will make.

I have a rather old 7800GT and single fore A64 @ 2.45GHz yet even then the PPU does not give me an FPS boost big enough to warrant its £80 price tag. If it doesn't help me, it will help people with high-end rigs even less. It is a pointless waste of money.

“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?

Just because it makes a difference in FPS does not mean it is worth it. If it was a 50% boost then everyone and their dog would have one. But you are lucky if you get just a few FPS, which at anything already over ~40FPS is pretty much pointless.

Also, see how easy it is to use the quote button and how much easier it makes it for everyone else?

 

[Pottsey]

“Do you have a PPU? If you do try it for yourself and see. You don't need a timedemo to see that you are not getting much of an FPS boost.”
I appear to be getting a decent FPS boost about 15fps. But I want a time demo to be sure. Like I said before bots could be messing up my score. Its not a fair bencmark and I might be wrong about 15fps.




“The amount going on is not going to be THAT much different. Sure it won't be totally accurate, but it will give you a rough guide as to what difference it will make.”
It is. Without a PPU, record results on the same map. The average FPS jumps around a lot based on the bots and your play style. If you spend lots of time with the sniper one map, then close range weapons next time on the same map get a very different FPS results. Hanging around one part of the level more then the other part can change results a lot as well.





“But you are lucky if you get just a few FPS, which at anything already over ~40FPS is pretty much pointless.”
100 is better then 60 and 65 is better then 55. More FPS is always better. Anything over 40 isn’t useless. Your just guessing at it only being a few FPS. There is no evidence to say that. We don’t have any evidence about the precise number yet. You could be right, then again you could be wrong we need a good benchmark.




“A64 @ 2.45GHz yet even then the PPU does not give me an FPS boost big enough to warrant its £80 price tag.”
So you have a PPU or is that another guess?

 

[Lightnix]

and 65 is better then 55

Aye, but is it £80 better?