Layte's super-duper 480SLI Benchmark-O-rama!

[layte]

Layte's super-duper 480SLI Benchmark-O-rama! - AMD shareholders cry salty tears

As some of us know, when ever 480SLI benchmarks are talked about somebody will pop up and state how much faster their ATI tri-fire setup is. I am interested to see if this is really the case, so to do that I have benchmarked a few different titles (more to come, give suggestions) at a resolution of 1920x1200 on my 480SLI setup.

I would appreciate it if any tri-fire owners could run their own tests with identical settings so that the two different setups can be compared.

System Specs:
[email protected]
SLI GTX480 900Mhz Core 4400Mhz Memory

So to start, here are the pretty pictures (will links to pretty pictures..)

Just Cause 2
Dark Tower
Concrete Jungle
Desert Sunrise

Crysis Warhead
Crysis Warhead

Stalker: Call of Pripyat
Ultra DX11 (click the rate button and run the DX11 test) The first test had a weird loading judder at the start, hence the low minimum. (I couldn't be bothered to re-run as it takes ages and the subsequent tests are more demanding anyway )

Street Fighter IV
Settings
Result

As requested by RavenXXX2

Resident Evil 5
RE5 1920x1200 16xAA (memory use never went above 700MB)
The test decided to go into window mode in test 4, hence the momentary dip.

Dirt 2
1920x1200 DX11 all settings maxed 8xAA

Crysis
Assault Harbour 1920x1200 8xAA

27/07/2010 20:07:32 - Vista 64
Beginning Run #1 on Map-harbor, Demo-Assault_Harbor
DX10 1900x1200, AA=8x, Vsync=Disabled, 32 bit test, FullScreen
Demo Loops=4, Time Of Day= 5
Global Game Quality: VeryHigh
==============================================================
TimeDemo Play Started , (Total Frames: 4100, Recorded Time: 132.23s)
!TimeDemo Run 0 Finished.
Play Time: 77.30s, Average FPS: 53.04
Min FPS: 35.22 at frame 3906, Max FPS: 89.88 at frame 985
Average Tri/Sec: -10573434, Tri/Frame: -199343
Recorded/Played Tris ratio: -0.07
!TimeDemo Run 1 Finished.
Play Time: 65.71s, Average FPS: 62.39
Min FPS: 35.22 at frame 3906, Max FPS: 89.88 at frame 985
Average Tri/Sec: -14277312, Tri/Frame: -228832
Recorded/Played Tris ratio: -0.06
!TimeDemo Run 2 Finished.
Play Time: 65.48s, Average FPS: 62.62
Min FPS: 35.22 at frame 3906, Max FPS: 89.88 at frame 985
Average Tri/Sec: -14586503, Tri/Frame: -232940
Recorded/Played Tris ratio: -0.06
!TimeDemo Run 3 Finished.
Play Time: 66.01s, Average FPS: 62.11
Min FPS: 35.22 at frame 3906, Max FPS: 89.88 at frame 985
Average Tri/Sec: -14470539, Tri/Frame: -232988
Recorded/Played Tris ratio: -0.06
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Completed All Tests

<><><><><><><><><><><><><>>--SUMMARY--<<><><><><><><><><><><><><>

27/07/2010 20:07:32 - Vista 64

Run #1- DX10 1900x1200 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 62.373333333333333333333333333

_______________________________________________

As requested, here are some single card numbers (same OC)

Just Cause 2

Dark Tower - Single Card (I accidently ran this test at stock and got 39.54fps)
Desert Sunrise - Single Card
Concrete Jungle - Single Card

Dirt 2

Dirt 2 - Single Card

Crysis

30/07/2010 20:15:15 - Vista 64
Beginning Run #1 on Map-harbor, Demo-Assault_Harbor
DX10 1900x1200, AA=8x, Vsync=Disabled, 32 bit test, FullScreen
Demo Loops=4, Time Of Day= 5
Global Game Quality: VeryHigh
==============================================================
TimeDemo Play Started , (Total Frames: 4100, Recorded Time: 132.23s)
!TimeDemo Run 0 Finished.
Play Time: 107.50s, Average FPS: 38.14
Min FPS: 29.77 at frame 2903, Max FPS: 54.18 at frame 981
Average Tri/Sec: -8984397, Tri/Frame: -235573
Recorded/Played Tris ratio: -0.06
!TimeDemo Run 1 Finished.
Play Time: 101.53s, Average FPS: 40.38
Min FPS: 29.77 at frame 2903, Max FPS: 54.18 at frame 981
Average Tri/Sec: -10843651, Tri/Frame: -268515
Recorded/Played Tris ratio: -0.05
!TimeDemo Run 2 Finished.
Play Time: 101.49s, Average FPS: 40.40
Min FPS: 29.77 at frame 2903, Max FPS: 54.18 at frame 981
Average Tri/Sec: -10864966, Tri/Frame: -268941
Recorded/Played Tris ratio: -0.05
!TimeDemo Run 3 Finished.
Play Time: 99.62s, Average FPS: 41.16
Min FPS: 29.77 at frame 2903, Max FPS: 54.18 at frame 981
Average Tri/Sec: -11251692, Tri/Frame: -273391
Recorded/Played Tris ratio: -0.05
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Completed All Tests

<><><><><><><><><><><><><>>--SUMMARY--<<><><><><><><><><><><><><>

30/07/2010 20:15:15 - Vista 64

Run #1- DX10 1900x1200 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 40.646666666666666666666666667

Now some GTA:IV TfLC numbers

Ballad of Gay Tony

Statistics
Average FPS: 46.47
Duration: 33.21 sec
CPU Usage: 36%
System memory usage: 62%
Video memory usage: 91%

Graphics Settings
Video Mode: 1920 x 1200 (60 Hz)
Texture Quality: High
Shadow Quality: Very High
Reflection Resolution: Very High
Water Quality: Very High
Texture Filter Quality: Anisotropic x16
Night Shadows: Very High
View Distance: 100
Detail Distance: 100

Hardware
Microsoft Windows 7 Home Premium

Video Adapter: NVIDIA GeForce GTX 480
Video Driver version: 258.69
Audio Adapter: Speakers (High Definition Audio Device)
Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz

File ID: TBoGT-Benchmark.cli

Devil May Cry 4
Everything maxed - 16xAA

Far Cry 2
Settings page 1
Settings Page 2
Results
Average Results


To be continued.....

 

[layte]

Sweet, how about Dirt 2, RE5, and the original crysis but do the assault harbor bench and not the fly through.

Good suggestions, watch this space.

 

[layte]

Ooohh just found out FFXIV beta is running tonight, tomorrow for more benchies folks.

I dont have BC2, if somebody would like to give me a copy....

 

[layte]

Need to state your system specs and any overclocks you got on.

Speeds as shown here: http://homepages.nildram.co.uk/~layte/pics/vantage-p-new.PNG

 

[layte]

Sweet, how about Dirt 2, RE5, and the original crysis but do the assault harbor bench and not the fly through.

As requested.

 

[layte]

Yea, lets do the like for like for the time being. Then I can turn up the settings to make the 1GB cards cry in the corner.

 

[layte]

I dunno, a momentary minimum of 35fps in a very graphically intensive scene while running at 1920x1200 with 8xAA sound pretty darned good to me. We all know that the lolconsoles are going to struggle to get 30fps in the sequel while only having to hit 720p.

 

[layte]

I can do 2560x1600 tests. Also, I would like to eventually see some tri-fire scores.

 

[layte]

i too can benchmark at 2560x1600, 480 sli, if you tell me what benchmarks to run, i did run the final fantasy 14 official benchmark yesterday, maybe that could be a test everyone could run. it only runs at 1920 though.

also i am going to a lan tomorrow and my m8 has the exact same mb/ cpu / overclock as me but he has a 5970 instead of 480's, i could run a test on his too if needed.

The FFXIV test currently has no multi GPU support.

 

[layte]

I can try a few tests with SLI disabled later on.

 

[layte]

Layte, single 480 results if you please.

Tomorrow, had a busy evening.

Also; still waiting on some tri-fire results.

 

[layte]

GTA4 benchmarks would be nice

I'll do some tonight.

 

[layte]

Is it time to post "You're not singing anymore" for the tri-fire crowd?

 

[layte]

@ Razor Time / Layte

What PSU's have you got powering those monsters?

edit: he sees Razor Time's sig and immediately edits the post.....

I have a Be Quiet! DPP 1200W: http://www.overclockers.co.uk/showproduct.php?prodid=CA-040-BQ&groupid=701&catid=123&subcat=1083

 

[layte]

I'd just like to remind people that I would be happy to do any BFBC2 tests if somebody has a spare code lying around.

 

[layte]

First post updated with some single card numbers.

Also, given the lack of any tri-fire numbers. I am calling this a total whitewash for the GTX480 SLI setup.

 

[layte]

Layte's super-duper 480SLI Benchmark-O-rama!

I would change the title of the thread if it was meant to be a TRI 5870 CF v SLI 480.

And i would imagine many of the 5870 & 5970 users are all benched out seeing how long they have had them for.

Try looking back & find some TRI benches.

Well surely they could post some results then if they are all benched out, sounds like they would have results to spare. Perhaps the guy who liked to repeatedly mention how much faster his tri-fire setup was than any 480SLI system could put his money where his mouth is. I asked time and time again for tri-fire owners to post results, especially those who liked to go on about how much faster they were. I'd like the numbers to be posted, even if it shows my setup to be slower. At least we would have an answer then, instead of the constant sniping that occurs at present.

Until then I'm calling this a performance win for the green team.

Edit: Oh and I did look for tri-fire benchmarks but the only ones I found were Vantage, Haven and AvP. All of which you trashed. Hence this thread and an open call as to what tests to run.

 

[layte]

Hey.

Take a look at the first post and try a few of those. If you want to do some other tests please feel free. If I can run them as well then more the merrier. The more hard numbers we have then the better for anyone looking at buying a high end setup.

 

[layte]

i dunno.. im on 10.4beta at the moment
what settings do you use for the advanced option in stalker?

Hit the rate benchmark button and choose Ultra DX11. This sets a number of fixed options and gives a score at the end.

 

[layte]

First post updated with some GTA:IV TfLC, DMC4 and FC2 results.

It would be nice to see something more comprehensive than RE5 numbers from the Tri-Fire crew, does that one scale nicely or something?