2900XT / 8800GTS Oblivion bench = On Par

willhub · 10 Jul 2007 at 19:18

Hi.

Well I just been benchmarking Oblivion and got Tom NBK to do the same benchmark at 1280x1024 16xAA/AF, reason why its 1280x1024 is becayse my old monitor only supports 1440x1050 max.

I just thought I would like to share the results, as you can see the HD2900XT and 8800GTS are on par I would say:

Tom's:
1280x1024
16AA/AF

Min Max Avg
35, 96, 55.559

Will's:
1280x1024
16AA/AF

Stock:
Min, Max, Avg
35, 114, 50.492

OC (650/2.2Ghz)
Min, Max, Avg
41, 126, 62.467

Now I'm just showing results, not trying to create another bad full on flame war Vs thread.

I know its one game but I can try to get some more benchmarks for other games.

Tom|Nbk · 10 Jul 2007 at 19:24

willhub said:
Hi.

Well I just been benchmarking Oblivion and got Tom NBK to do the same benchmark at 1280x1024 16xAA/AF, reason why its 1280x1024 is becayse my old monitor only supports 1440x1050 max.

I just thought I would like to share the results, as you can see the HD2900XT and 8800GTS are on par I would say:

Tom's:
1280x1024
16AA/AF

Min Max Avg
35, 96, 55.559

Will's:
1280x1024
16AA/AF

Stock:
Min, Max, Avg
35, 114, 50.492

OC (650/2.2Ghz)
Min, Max, Avg
41, 126, 62.467

Now I'm just showing results, not trying to create another bad full on flame war Vs thread.

I know its one game but I can try to get some more benchmarks for other games.

I did loads of benchmarks in the 2900 thread but no one was interested

Teaf · 10 Jul 2007 at 19:31

Yeah but 2900XT uses more power!!!11one

sorry, couldn't resist, think I may dig out my Oblivion disc if I can find it and do some benchies too

Cyber-Mav · 10 Jul 2007 at 19:33

would be interesting to see the results if you could have run them at 1600x1200.

Tom|Nbk · 10 Jul 2007 at 19:38

I did a little perf log

HD2900XT Oblivion Performance @ 1680x1050.

8xAA =

Min Max Avg
31 90 46.914

12XAA =

Min Max Avg
28 80 44.092

16xAA =

Min Max Avg
28 66 42.179

Rest of specs:

C2D @ 3.4Ghz
2GB RAM
ASUS P5B DELUXE

1280x1024
16AA/AF

Min Max Avg
35 96 55.559

HangTime · 10 Jul 2007 at 19:42

Not really a fair comparison though is it as you are using different rigs.
I was expecting the two cards to at least have been benched in the same system.

What we really need is a proper review site to churn out some numbers using the latest beta driver sets from both sides.

willhub · 10 Jul 2007 at 22:05

HangTime said:
Not really a fair comparison though is it as you are using different rigs.
I was expecting the two cards to at least have been benched in the same system.

What we really need is a proper review site to churn out some numbers using the latest beta driver sets from both sides.

Well It's something at least, it gives an idea, I dont think the 200Mhz difference in mine and tom's CPU will make much difference.

With dual monitors I think the best I can get is 1600x600 or 2560x1050 or something like that.

Hamoodii · 10 Jul 2007 at 22:37

but Tom runs a 3.4GHz C2duo and you run a 2.4GHz C2Duo, surely that would make the results vary a tad?

Hamoodii · 10 Jul 2007 at 22:38

oops already posted

Hamoodii · 10 Jul 2007 at 22:40

Can you run at high resolutions would be nice to see how they compare at something 1920x1200, also try and stick them on a graph would look nicer

willhub · 10 Jul 2007 at 22:42

Hamoodii said:
Can you run at high resolutions would be nice to see how they compare at something 1920x1200, also try and stick them on a graph would look nicer

My C2D is at 3.2Ghz and I cant get 1900x1200 results, if I had both card here I would do a propor benchmark with full detail and graphics but I dont, allthough I can buy one, dont wanna end up with 2 card and 1 PCI-E system though

Rroff · 10 Jul 2007 at 22:49

Hamoodii said:
but Tom runs a 3.4GHz C2duo and you run a 2.4GHz C2Duo, surely that would make the results vary a tad?

If they are running at 1280x or beyond with FSAA and AF then no... I benched oblivion quite extensively and the fps difference between 2.4gig and 3.0gig was pretty much nothing... and even pushing it upto 3.6gig only gave it a couple of fps here and there over 2.4gig...

fornowagain · 10 Jul 2007 at 22:55

willhub said:
snip

What settings exactly? Multi, Coverage, Supersample AA, Which transparency? Optimizations? Mipmaps ? All in game settings maxed? Box, tent, wide filters?

I don't see the point of 16xAA. Its hardly a direct comparison after 4xMSAA.

http://www.techreport.com/reviews/2007q2/radeon-hd-2900xt/index.x?pg=6

willhub · 10 Jul 2007 at 22:58

fornowagain said:
What settings exactly? Multi, Coverage, Supersample AA, Which transparency? Optimizations? Mipmaps ? All in game settings maxed? Box, tent, wide filters?

I don't see the point of 16xAA. Its hardly a direct comparison after 4xMSAA.

http://www.techreport.com/reviews/2007q2/radeon-hd-2900xt/index.x?pg=6

I have gamma correction, transpaprency, SS.MS off, just 16xAA, I dont see what its got to do with this, do I look like i'm comparing it to that, I was comparing the GTS to the HD2900XT at the same res/AA/AF....

fornowagain · 10 Jul 2007 at 23:10

willhub said:
I have gamma correction, transpaprency, SS.MS off, just 16xAA, I dont see what its got to do with this, do I look like i'm comparing it to that, I was comparing the GTS to the HD2900XT at the same res/AA/AF....

Why don't you read it? The XT doesn't have 16xAA. It uses pixels from adjacent filters to build effective sample sizes.

HangTime · 10 Jul 2007 at 23:12

What he's getting as is, were you both using identical settings, mods etc etc?

There's too many factors outside of that anyway for a fair comparison, even if you ignore the 200mhz difference in cpu speed, factors like memory speed/latency, motherboard, OS install/background processes etc can all have an effect.

OK sure, you could say that they were very loosely comprable in performance terms, but I certainly wouldn't base any purchasing decisions based on these results.

willhub · 10 Jul 2007 at 23:16

Right, transparency, gamma correction turned off, 16xCSAA or 8xAA? at 1280x1024, I am just trying to get some damn close comparisons here, I'm even upping my cpu to 3.4 at 1.55V to make it even more fair.

Tom|Nbk · 10 Jul 2007 at 23:17

fornowagain said:
Why don't you read it? The XT doesn't have 16xAA. It uses pixels from adjacent filters to build effective sample sizes.

It dosn't have 16xAA as such, but has a sampling rate of 16x just as much as normal AA but uses 8xAA (Box) then by applying Wide tent it ups the sampling rate to 16x.

fornowagain · 10 Jul 2007 at 23:20

Tom|Nbk said:
It dosn't have 16xAA as such, but has a sampling rate of 16x just as much as normal AA but uses 8xAA (Box) then by applying Wide tent it ups the sampling rate to 16x.

Lol, thanks Tom. I know how it works and the two implementations are not directly comparable, other than graphically sort of. If you want to go frame to frame stick to 4xMSAA.

willhub · 10 Jul 2007 at 23:24

fornowagain said:
Lol, thanks Tom. I know how it works and the two implementations are not directly comparable, other than graphically sort of. If you want to go frame to frame stick to 4xMSAA.

As for as most people are concerned it is since if someone wants to use 16xAA on either card they aint gonna use 4xMSAA.

With my 8800 2xMSAA gives about the same performance hit as 16xAA