http://www.pcper.com/article.php?aid=464
Great article :)
Printable View
http://www.pcper.com/article.php?aid=464
Great article :)
all I want to see is PhysX in action.
this is the big game Ageias been waiting for.
so lets see if its actually worth it.
Good review. I love how they show not just average FPS, and also an FPS chart. It's obvious that the progression in performance is as follows: 8800GTS 640MB, 2900XT 512MB, and 8800GTX. The lower resolutions don't count due to the 60FPS cap, so I'm mostly looking at the more insane reses.
RV670 should be perfect for this game, especially if 8800GT only has the 92 shaders instead of the rumored 112.
Also, it seems to me that R600 architecture is very well suited for whatever the UT3 engine is doing. Both Rainbow 6, Bioshock, and UT3, all games that use the UT3 engine, seem to favor R600. Since most of the upcoming blockbuster games use the UT3 engine, that's another huge win for ATI/AMD.
Looks like another win for Nvidia. But from what i see, Is there a limit on FPS in UT3?
Also, The Rumored 112 shaders was for the New GTSQuote:
Originally Posted by Shadowmage
PS. I seems either the Beta Demo hates me or something, because its taking forever to install.
How I wish 8800GT and 2950 Pro/XT was included in that review...
I wonder if you can remove the FPS cap, should be possible in one way or another.
Indeed. But we also need to realize that this a "Beta Demo", Demo and beta in the same title made me rofl abit.:rofl:Quote:
Originally Posted by RPGWiZaRD
Based on price, this is a win for ATI, since the 2900XT competes with the GTS. Anyways, it doesn't matter that there's a FPS cap. Max FPS is not interesting. Who cares if you get 20000 FPS in certain areas, but only 6FPS when something explodes? Is average FPS an interesting figure in such cases? Obviously not. The key FPS figure should be min FPS, and the average FPS number needs to be biased to make min FPS count more than max FPS.Quote:
Originally Posted by Warboy
There was another rumor about 112 shaders for G92. That's what I was referring to.
The 8800GTS and the 2800 are almost the same price. So I wouldn't mark it as a win for ATI.Quote:
Originally Posted by Shadowmage
QFT. Min FPS is most important. I love reviews that actually include it.Quote:
Originally Posted by Shadowmage
That review should be titled "HD 2x00 Redeemed" :p
Perkam
Thanks for heads up! AA working for NV?
Quote:
Originally Posted by Warboy
The 2900 performs close to the GTX so the win does go to ATI, but I can tell you will draw your own conclusion anyways.
Guys...for the record...
They aren't using a supported driver for the NVidia side. 163.71 is the minimum driver as per epics statement.
http://forums.epicgames.com/showpost...79&postcount=4
That statement is from epic themselves.
Not saying it'd turn the tables, just pointing this out. What made it catch my eye is because on a much lower system I can't get the framerate to drop below 62 on my GTX, yet they did.
Wait till AA is turned on.
ATI will certainly fall back... ;)
good article
:up:
Indeed ^_^Quote:
Originally Posted by JoeBar
Quote:
Originally Posted by JoeBar
dont know why , with hd2900xt @ 1680x1050 with all max and AA ON i have almost all time 62FPS loked >>> http://www.xtremesystems.org/forums/...2&postcount=47
:up:
I thought the Unreal 3 engine didnt have AA?
I run it maxed at 2560 x 1600 just fine! 40-50 average fps, never below 35
Is it just me or this thing looks like crap? IQ is on max, 1600x1050 but it looks so 2005...:shrug:
I blame that on Gears of War using a Pre-Build of UT3 Eng.Quote:
Originally Posted by Caparroz
Oh so now next-gen games look like crap? Some people need to have their head examined.Quote:
Originally Posted by Caparroz
I guess some kids just don't appreciate games anymore and how much work goes into them. If you can't enjoy a game for what it is, and are fixated on the graphics and scrutinizing them against the other offerings on the market, then you will never be happy.
Gameplay > graphics
If graphics are WOW but gameplay totally sucks then what's the point?
If gameplay is WOW and graphics suck well it is still fun. That is why I still have that stupid little worms armageddon game on my PC, graphics are totally outdated (2D anyone?!) but because it is so stupidly fun I still play it from time to time ;)
I don't like FPS games. I really can't stand their dullness, with a very few exceptions. And if I got it right UT3 is, essentially, the same old UT gameplay with some new modes and weapons. But that's just this "kid" opinion...Quote:
Originally Posted by Soulburner
I grabbed this demo for the sole purpose of checking U3 Engine out. I didn't know that this was a childish thing... I'm sorry if I spoiled your favorite game.
Besides that, being a sim race junkie, gameplay is the most important thing to me by far. My favorite game is still Grand Prix Legends, a 10 year old game.
Don't need to start calling people stupid kids...:rolleyes:
AA doesn't matter...at all. Not only does it look great wo/ it, but you're moving @ UT99 speeds!
Nice to see the 2900XT performing well.
NV, right? I remember folks renaming Bioshock's .exe to Oblivion and then using "override" awhile back. I wonder if the rename's even necessary.Quote:
Originally Posted by Seraphiel
Also, I know override's used when there's no AA option and enhance's used when the setting's in the game, but I've noticed HL2 looks better w/ override than enhance and that game does have AA in game. I suppose there's no better than trying it yourself, but it'd be nice if it worked the way it was meant to be playe...er... :shakes:
I tried that, didn't work.Quote:
Originally Posted by Seraphiel
Yeah those are the drivers I'm using.
EDIT: Nevermind, it's working now.
Sorta blah'ish lookin levels in the demo as is Epic norm. I'm just worried that UT3 will be a modders nightmare compaired to UT2k3/2k4, kinda like Oblivion was in the face of Morrowind. The community bonus packs are what made ut2004 worth owning.
LOE, what?
Epic developed the engine, Epic developed the game...
At any rate, the graphics aren't by any means disappointing me. The only slight thing I am disappointed about is that the demo is whoring bloom all over the place.
Talking about PhysX for a bit...
The engine part isn't being pushed very much in the demo, not beyond debris and perhaps some waterfalls. Nor can we see whether or not using the card can gain one some fps, as the engine is capped at 62FPS right now, and I have been unable to unlock it yet, but I am working on it...
However, as with any UT game, the engine isn't pushed until the community gets it's hands on the editor, and has a go at modding around.
So, when the full game comes out, then I can assure you we'll see some proper physx usage.
C:\Documents and Settings\*windows account*\My Documents\My Games\Unreal Tournament 3 Demo\UTGame\Config\UTEngine.INIQuote:
Originally Posted by SafeFire
[Engine.GameEngine]
bSmoothFrameRate=TRUE
MinSmoothedFrameRate=22
MaxSmoothedFrameRate=200
Don't forget to disable Vsync for your test (altough I'd keep it on during normal gaming for smoother gameplay).
AND:
[SystemSettings]
*.............................*
MaxAnisotropy=16
MaxMultisamples=4 (not sure if this works)
*............................*
game looks to be well optimized.... a single gtx @ 2560 x 1600 avg 50fps
very nice :)
Roger, will test asap. Thanks
UPDATE
I did some testing, ran Suspense a few times (should be the map that strains physics simulation the most, having a lot of vehicles and destroyable architecture)
First up a few screenshots with Hardware Physics Acceleration OFF:
http://www.imagehosting.com/out.php/...oPhysXOFF1.jpg
http://www.imagehosting.com/out.php/...oPhysXOFF2.jpg
Now, a few shots with Hardware Physics Acceleration ON:
http://www.imagehosting.com/out.php/...moPhysXON1.jpg
http://www.imagehosting.com/out.php/...moPhysXON2.jpg
In terms of visual quality, I can see very little difference, this comes as no shock though, since it wouldn't make sense to add extra debris to the scene when a card is slotted in.
However, I am pleasantly able to announce to you that when playing with the physics card in, I was seeing a performance gain of up to 10 frames per second, something that I was not expecting, but was indeed hoping to see.
In terms of raw numbers, this spells out like this:
When physics are going on, that means debris flying all around, explosions etc. Judging from the screenshots, with the physX card off, I'm seeing 29-30 FPS
With the Card in, in similar situations, I'm seeing 39-40 FPS.
Thanx a lot for these tests SafeFire, very useful indeed.
Now, only one test remains: Quad Core vs Dual Core. Unreal Engine 3 promised to be seriously multithreaded, is it true?
THAT is one thing I am unable to test, as I am only in posession of a Quad core chip right now :P
If anyone feels like donating me one then I'll be happy to give it a shot (although I'm sure quite a few review sites are working on it).
When I hear some news on this, I shall post it :)
Start->run->msconfig->BOOT.INI->Advanced Options->NUMPROC= 2 (4 is default for quad core). I've tried running UT3 on 1 core by setting process affinity in Task manager, but this will screw up game performance even when changed back to 2 cores on my C2D. Multi threading in UT3 is very complex, so I recommend booting with the amount of cores you want to test the game with.Quote:
Originally Posted by SafeFire
Have fun :)
From what I have seen, Unreal 3 Engine supports dual cores, but I don't think it has been able to scale up to more than that yet. My CPU activity never seem to exceed 50% (more than 2 cores, 1 core = 25% each).
Looks like 7900/1900 series will be the minimum you want to get away with if you are running the widescreen resolutions.
damn it, i need another 8800GTX now :( i wont be able to run any game @ 1920x1200 with Max AA and Max AF if i dont
Ever notice how you really really really can't tell the difference between 4x AA and 8x AF versus 16x AA and 16x AF? ;)Quote:
Originally Posted by p8ntslinger676
I was just playing at 1920x1200 with 4xAA 16xAF and the lowest I hit was 40fps? :confused:Quote:
Originally Posted by p8ntslinger676
I forced AA in the nv panel wo/ renaming exe and it def worked, but halfed my framerate. Honestly, it's too fast paced to be concerned about jaggies. :D
Yeah, the unreal engine 3 is VERY good at masking jaggies anyway honestly. :DQuote:
Originally Posted by Jodiuh
the midrange card review is crap.
first, they use uxga as lowest resolution (:rofl: ) and they wonder that it will not run with 60fps at max details :rolleyes:
But as this is not crazy enough, the even RAISE the resolution.
Who would be a full dumbass and want to play a res higher then full-hd with a 100€ card ?
then they only compare 8600GTS and 2600xt, which makes no sense.
2600xt is much cheaper then 8600gts, where i live it's even cheaper then 8600gt.
Quote:
Originally Posted by SafeFire
The game looks almost nothing like that on my rig. No where near as much bloom when I play. :confused:
True...though the max a GT will clock to is what a GTS is at stock..which says quite a bit for a card that is cheaper than the GT, beating out at stock what the Nvidia card fully clocked can get to.Quote:
Originally Posted by grimREEFER
Again, this is all ATI`s late-to-the-game performance that we are seeing.
Perkam
what res and settings is safefire playing at? it worries me taht a pc of his calibur is dipping into the 20's and 30's
I'm actually going to tone down the graphics when I play. The pretty graphics distract me too much, can't concentrate on hitting my enemy ;)
There are no graphics options to be found in the demo.
Yes there are under advanced. Just some basic things like texture quality, etc.Quote:
Originally Posted by Soulburner
UT3 update patch
Quote:
If you have Windows XP 64-bit or Windows Server 2003 and you are receiving the "Modified executable code is not allowed" error at startup then you should replace the game's executable with the one from this patch.
Possibly, but we'd have to see it re-benched with a epic approved driver. It even says in the readme that the lowest driver you should be using is the 163.71 driver, and everything below that can cause hitching(which obviously drops the overall framerate). ;)Quote:
Originally Posted by perkam
So, essentially, we aren't seeing the full nvidia performance in the review. You can check the readme file, or check the link I posted earlier where epic themselves made the statement on their forum. :D
havent played since ut1 properly, ut2k3 was crap as a ut game but good as just a stand alone game, never tried 2k4...
this is make or break for me ;P
edit: pretty sure i got a constant 60 fps with textures and geom at 2 :\
some things though
vehicle ctf.. not unreal
having a deathmatch but then a poonsauce lightning walker thing pops up to own you all - not unreal
no ripper (i think ? ) - not unreal
overall though the gameplay is ALOT faster than i expected, its even faster paced than ut1, one thing i wish for is instagib and some kind of weapon skin+map pack that brought back the skins of original weapons + deck16][ mmmm i loved that map as well as iCTF eternal caves/face1 /lavagiant
good times... good good times
They need to bring back Assault mode. As one of the funnest gameplay modes of UT I cannot believe they have ignored it for this long.
PyroFire...don't know what you're on, but everything here is classic UT. They have taken the best of what has kept the franchise alive and made it better.
That's all I need right there. They had this in the UT2K3 demo. :(Quote:
Originally Posted by PyroFire
EDIT: Wait, this too...
Quote:
Originally Posted by Soulburner
, you forgot to mention that with new ati drivers the ati cards will in turn have greater performance , :up:Quote:
Originally Posted by DilTech
i have an e6600 @ 3.0ghz 333fsb and a single 7900gto not overclocked ran at 1280x1024 and max quality settings no aa or af i get this:
Frames=14813
Time (ms)=396034
min=15
max=57
avg=37.403
ill try to get some with af enabled soon..
*taken with fraps benchmark utility.
We can only hope.Quote:
Originally Posted by Jodiuh
guys its a BETA DEMO, what you see here is only the begining.
The 7.10 was an increase for cross-fire, or should I say it was for making X-Fire work.Quote:
Originally Posted by Papu
Also, they used the 7.10 beta, which there wasn't much of a change between 7.10beta and the full version.
However, with the NVidia driver, it's cited clearly in the readme NOT to use anything lower than 163.71 with unreal tournament 3's beta demo as it causes hitching.
I'll be interested to see SLI and crossfire scaling when the FPS cap is gone, like hopefully in the final build.
hmm, when one changes the FPS cap, ones framerate drops.
I was running @ 62fps constantly, then I set the fps cap to 200 my framerate was pulled down dramatically, perhaps the 62fps at all times was wrong, after all.
Might just be a bit buggy. I'm still going to wait untill final build before I make up my mind to get it.
Quad core vs dual anyone? Or at least open your task manager those with quad cores :)
That can happen if you are running VSync but cannot push 60fps. Smoothness is more important than framerate, if you can run 60 just cap it. There is no reason to need any more.Quote:
Originally Posted by SafeFire
Or if you have a higher refresh rate like 75, 85, or 100Hz cap your framerate to that but keep Vsync on.
I don't know, I used 162.18 just fine in the demo.Quote:
Originally Posted by DilTech
Diltech you know well that ATI`s comparable drivers come out at least a month after a game is released, when tweaks for the game show up in the driver logs.Quote:
Originally Posted by DilTech
If you want to compare ATI`s with 163.75`s, then wait for 7.11 or 7.12.
Perkam
Good that you bring this matter up, this will be one factor to take into consideration when picking between 8800GT/2900Pro/XT. I love trying out different drivers and NVIDIA spits new drivers out currently almost every week :D and they have a lot wider driver tweaking community and I like especially the XG drivers.Quote:
Originally Posted by perkam
core 1, 3 and 4 were sitting neatly at 60-70% usage, while core 2 was hovering around 3-40% usage:up:Quote:
Originally Posted by kiwi
Yeah, I got exactly the same thing. Sort of :)
core 1 80%
core 2 60%
core 3 40%
core 4 40%
Seeing as that is 220% all together, it's pretty well optimised!
This sucks. I tried to play World in Conflict w/ a bro last month and it just wasn't happening. (2900xt vs 8800gtx) He had cheetah spots and all sorts of issues that made it virtually unplayable. Then TF2 came out and he couldn't stop crashing. :( I suppose once we get past this "Q4 of 07 shotgun game release" we'll finally have some stability to play in 08. :DQuote:
Originally Posted by perkam
It's not driver tweaks I'm talking about. It says clearly in the readme for UT3 not to use lower than 163.71. Not for performance, but for an actual hitching/all out STOPPING issue.Quote:
Originally Posted by perkam
http://forums.epicgames.com/showpost...79&postcount=4
From epic's own mouth.
Like I said before, I'm not saying it'll turn the tables, but I am saying that epic themselves say not to use any driver lower than 163.71 with unreal tournament 3, and the performance review used lower than that. See what I'm saying? ;)
Quote:
Originally Posted by Epic
Is this "hitching" an issue with very short pauses in the framerate, sort of like .1 second blips as you run around?
If so, UT3 did not have that issue for me with an 8800GTS and 162.18. However, Call of Duty 4 did, and 163.76 fixed it.
Yep, something like that. I experienced it until I updated my driver, so I can confirm it definitely exists.Quote:
Originally Posted by Soulburner
Either way, I just find it odd that a site would review performance of a title using a driver that's not recommended by epic themselves. :confused: