Source engine benchmarks

[FluxCap]

http://www.vr-zone.com/?i=1181

Looks like a close race.

 

[flyboy84]

close indeed, but HL2 will have some more demanding scenes, as they said in the article's conclusion

 

[Actaeon]

Cool benchmarks. The only real difference was at 1600x1200 @ 4xAA/8xAF, (the rest all appear to be engine/cpu limited) and even then (aside from the 6800), the difference was pretty small, and all are still above playable framerates.

ATi seems to have the upper hand with the Source Engine, but the difference was miniscule between the top cards. No where near as the significant difference between the D3 benchmarks.

Though, as the reviewer said, HL2 will be more demanding, perhaps we'll see different results when HL2 is offically benchmarked.

 

[StrangerGuy]

Why they didn't benchmark the R300 cards?

 

[gururu]

wow, the 6800s really choke at high resolution and AA/AF...

 

[jiffylube1024]

Originally posted by: gururu
wow, the 6800s really choke at high resolution and AA/AF...

Yeah, looks like the slower core and memory really catches up with the 6800's there.

 

[imported_oneiros]

Please don't start an Nvidia vs. ATI flame war over these benchmarks especially when all the cards performed quite well.

 

[Tab]

Those benchmarks dont look right at all...

 

[Schadenfroh]

this does not look like the 30% performance differance that was said by gabe to come with ATI

 

[Todd33]

Lame, they didn't bench any of the older cards. The 9800XT 4X AA shot looks bad too, something is wrong.

 

[Robor]

Wow, those non-AA shots look terrible - all those jaggies. Hope I can find a good deal on a X800XT or 6800GT/Ultra before HL2 comes out.

 

[nitromullet]

Originally posted by: Tabb
Those benchmarks dont look right at all...

Why do you say that..?

 

[Genx87]

Looks like they are pretty damned close to me. Certainly nothing like the disparity between the cards in Doom3.

Why do you say that..?

My guess is ATI didnt wipe the floor with Nvidia in a game everybody expects ATI to wipe the floor with Nvidia with.

 

[reever]

My guess is ATI didnt wipe the floor with Nvidia in a game everybody expects ATI to wipe the floor with Nvidia with.

Their fault for mis-reading/memorizing a quote from gabe which talks of HL2, not CS:Srouce

 

[Rage187]

Man, Nv is getting creamed....oh wait.


New and improved gabe, spouting off 30% more BS.

 

[nitromullet]

Well, I'm not going to put words in Tabb's mouth. Just curious is all...

I'm actually pretty impressed overall. All the cards tested minus the 6800NU are more than playable at 1600x1200 4xAA/8xAF. It's too bad they didin't give us a few screenies to look at.

 

[clicknext]

Big surprise, every card performs the same. HEEEY! Just like in every other game!! -_-'

 

[Childs]

Originally posted by: nitromullet
Well, I'm not going to put words in Tabb's mouth. Just curious is all...

I'm actually pretty impressed overall. All the cards tested minus the 6800NU are more than playable at 1600x1200 4xAA/8xAF. It's too bad they didin't give us a few screenies to look at.

Fileshack has some movies.

 

[Insomniak]

Originally posted by: Genx87
Looks like they are pretty damned close to me. Certainly nothing like the disparity between the cards in Doom3.

Why do you say that..?

My guess is ATI didnt wipe the floor with Nvidia in a game everybody expects ATI to wipe the floor with Nvidia with.

I'd agree. I think some of the ATi fannies are a little disappointed that they may not get their Doom 3 Revenge...

To be fair though, Half-Life 2 isn't out, and this isn't a free for all benchmark...just early, limited benchies of different content on the same engine. We'll see what happens with HL2 when (if) HL2 comes out. It may very well be different from these benches.

For the record though, I hope it's not. I do not like the prospect of having certain games perform much better on certain video chipsets. Buying one graphics card is expensive enough....

 

[kylebisme]

this is just plain funny, average framerate numbers from a capped benchmark is no absolutely no way to compare things. now if they would have made graphs of the fraps logs that would be telling, but as it stands their benchmarking was a complete waste of time.