Splinter Cell - nvidia vs. ATI

[cvearl]

I was watching a friend play Splinter Cell on an almost identical system to mine. Same Mobo, processor and vid card. We used the FRAPS utility to monitor framerates. At 1024x768x32 with 2xAA and High detail settings (not highest), we were averaging between 25fps and 35fps (still seemed somwwhat smooth though) throughout the training mission. Worse yet was when you enable night vision. Then it dropped to around 15fps.

How does this game play on a G4Ti4200, 4600 or even 5800Ultra?

Anyone else able to test thier with FRAPS.EXE and play the training mission for a few minutes and give me an idea of thier performance?

Charles.

[digitalwanderer]

I was playing with it last night, but didn't have fraps on the first couple of times...but I also spaced and left quincux/2xAF on so it seemed a bit jerky on Bubbles for me.

I killed off AA & AF and fired up fraps and noticed that my low dips never really went below 29fps after that. (I think high details, medium shadows)

I was averaging mid-40s, 50s most of the time, it was just low-dips I was watching for. A VERY pretty game, and kind of fun too! :)

I did notice that when I played around with AA & AF later that when I tried night vision with AA & AF on it turned into a slideshow for me...but I really ain't got the CPU power to pull that off yet. I meant to see if I could get away with running quincux only on it, but haven't had a chance. (And doubt I'll get away with it with the lighting in the game, BEAUTIFUL lighting BTW. ;) )

[Rogozhin]

ONly have my 9700 non pro. So can't give any feedback using nvidia cards.

We should do a fraps txt log entry on a specific level for a certain amount of time to get standardized benchies.

I will try it tonight and post on what level i did it and with exact settings.

Rogo

[RobHague]

Using the ingame FPS monitor i was getting about 40-50fps i believe it was, and 30fps with night/heat vision. No AA or AF and all the detail levels set to max.

Thats on a 5800 Ultra.

[digitalwanderer]

What's the console command for the ingame FPS monitor?

[Rogozhin]

Here is my FRAPS txt log file.

This was on the prison level (same as 1st demo). STarted fraps log after it began and ran it for about 4.5 minutes. This IS a great way to benchmark, i suggest you all do the same.

Here is the log.

2003-04-25 13:47:57 - 1024x768 All options on highest, graphics and sound. NOAA and 8x performance AF.
Frames: 14052 - Time: 395929ms - Avg: 35.491 - Min: 14 - Max: 87


Rogo

[Rogozhin]

This was switching between night vision and regular, no Heat in real game at this level.

Rogo

[RobHague]

Originally posted by digitalwanderer
What's the console command for the ingame FPS monitor?

Press F5 to bring down the console, think its F5 anyway - may be F2. Just try both till it pops up at the bottem. Then type Stat FPS.

:)

[Rogozhin]

Fellas

that fraps log was with my 9700 running 300/300, i will do it again and repost my score.

Thanks

ROGO

[digitalwanderer]

Originally posted by RobHague
Press F5 to bring down the console, think its F5 anyway - may be F2. Just try both till it pops up at the bottem. Then type Stat FPS.

:)

It's F2, thank you. Anyone know the bind command? "set input x" ain't working for me. :(

[cvearl]

Originally posted by RobHague
Press F5 to bring down the console, think its F5 anyway - may be F2. Just try both till it pops up at the bottem. Then type Stat FPS.

:)

Stat fps.... hmmmm. That would suggest that this is based on the UT2003 engine. That's the same command in that game.

Charles.

[Rogozhin]

Anyone willing to try my fraps txt log bench procedure?

Rogo

[cvearl]

Originally posted by Rogozhin
Anyone willing to try my fraps txt log bench procedure?

Rogo

I will on any game I have installed. I am still waiting to "borrow" my firends copy of splinter cell. Then I would for sure if you can wait until then.

Charles.

[Onde Pik]

A list of reviews have used this game for benchmark. It really doesn't like the FX cards. Radeons seam to smoke the FXs in this benchmark even more so than in everything else.

[jAkUp]

i was getting an average of 35fps at 1600x1200 all options maxed on my nv30

[Skinner]

Yes it uses the Unrealengine, build 927 with a lot of modified stuf.

[vampireuk]

Link (www.vampireuk.pwp.blueyonder.co.uk/abe.wav)

Now does that sum up fanboys better than anything else?:p

[Rogozhin]

cvreal

let me know if you get the game and run a txt log.

thanks ;)

rogo

[cvearl]

Originally posted by jAkUp
i was getting an average of 35fps at 1600x1200 all options maxed on my nv30

I am wondering how you are beating my CPU score so much in 3DMARK03. ARe you overclocking anything? And what are ULTRA memory timings?

Charles.

[stncttr908]

Originally posted by cvearl
Stat fps.... hmmmm. That would suggest that this is based on the UT2003 engine. That's the same command in that game.

Charles.
That's because it is. It's one of those new Unreal engines...

[jAkUp]

Originally posted by cvearl
I am wondering how you are beating my CPU score so much in 3DMARK03. ARe you overclocking anything? And what are ULTRA memory timings?

Charles.

nothing is overclocked... my fsb and cpu are all at the default... my corsair memory is at 400mhz cas 2 ultra timings...... (in the bios there is timing settings safe/normal/turbo/and ultra) im thinking thats its probably the memory speeds. how fast is your memory?

[jAkUp]

im not using the patch yet... im about to install it

[cvearl]

Originally posted by jAkUp
nothing is overclocked... my fsb and cpu are all at the default... my corsair memory is at 400mhz cas 2 ultra timings...... (in the bios there is timing settings safe/normal/turbo/and ultra) im thinking thats its probably the memory speeds. how fast is your memory?

Ahhh. I see. My FSB and Vidcard clocks are all standard. No overclocking. As for memory I am running Micron 2.5 CAS PC2700 DDR333 at standard speeds. Unfortunately it wont go much faster. It's not like Corsair XMS3200 or somehting like that. But it appears as though your score is higher relative to your CPU score beating mine and I do know that memory bandwidth plays a large part in CPU performance in 3D.

My MOBO supports separate memory and FSB timings in which case I am thinking of going and getting PC3200 Corsairs and pushing the memory up a bit more aggressively. Essentially, that means DDR400??? Ok, so what are the timings for that? DDR333 is 166*2 right? So if I want to go DDR400 do I set it to 200*4?

If I am not "overclocking" the CPU leaving it at 18*133 (2.4Ghz), but I am separately pushing XMS3200 RAM to 200*4 (400Mhz), are there any side effects to that action? Stability? Heat? Timing with the processor and memory controller? Or do people do this all the time? I don't like overclocking but this really is not overclocking if the motherboard and RAM both support it right?

Also, does doing this give you near PC1066 RDRAM performance?

Just wondering.

Charles.

[saturnotaku]

I thought it was not beneficial for Athlon systems to run with their FSB and memory speeds out of sync, and in fact hurts performance to a degree. Am I wrong on this? :confused:

[jAkUp]

i had stability problems with my memory.... the only way to fix it was to up the voltage to 1.7v. i dunno why though... the memory is rated to run at 400mhz... but yes, it will run a little hotter at 400mhz, but now with my 1.7 voltage setting, i have no stability problems... windows xp is rock solid... but then again, i have massive amounts of cooling. my memory speed is set to 200mhz in the bios... for some reason it did not default to that though... my memory is corsair 3200 XMS (400mhz)