Official HardOCP HL2 benchmarks

[Ruined]

http://www.hardocp.com/article.html?art=Njky

"This article will not cover te quality of the gameplay experience, as is our usual fare. Instead, today’s preliminary tests are to find out who has the fastest video cards for playing Half Life 2 in apples to apples benchmarks. This will dispel many of the myths being rumored about Half Life 2 performance as we cut to the chase and see who has the fastest Half Life 2 AGP video card."

HardOCP used custom timedemos to simulate normal gameplay.

Benchies:
6800U and x800xtpe trade blows depending on map (6800U does better on canals, x800xtpe does better on coast)
6800GT beats x800pro across the board

Conclusion:
"As for calling one video card a “winner” over the other, it just does not seem feasible and would likely be irresponsible to do so. ATI’s X800 series has not thrashed NVIDIA’s 6800 series silicon as many expected. Half Life 2 simply plays great on video cards from both NVIDIA and ATI. So let’s make this clear. The ATI Radeon X800 series of video cards will deliver a stellar Half Life 2 gaming experience. Also… The NVIDIA GeForce 6800 and 6600 series of video cards will deliver a stellar Half Life 2 gaming experience. There are simply no losers to be found if you have recently upgraded your video card. You have likely all made good gaming investments."

[Graphicmaniac]

then DriverHeaven can be called Lies-Heaven?

lol


it seems that Nvidia ha really won this nv40 vs r420 battle

same performance in H.L.2 and F.C. but hugely better in Doom3


i can finally say my uncle what to buy this Christmas :P

[NoWayDude]

Why do ppl take any notice on DH comparisons?Is it not plain to see that either they do not have a clue of what they are doing or they are just trying to put Nvidia cards in a bad light?
Besides, compare theyr findings with this litle article
www.theinquirer.net/?article=19707

Someone has coppied someone else i think.......
Now which way around , more dificult to know....

[zer0]

compare theyr findings with this litle article
www.theinquirer.net/?article=19707


that benchmark is a marketing propaganda of ATI...

that benchmark very is biased...

[Vapor Trail]

So what happened to the 30% lead for ATI in Hl2?????



:nanahump: :nanahump: :nanahump: :nanahump: :nanahump:

[NoWayDude]

by zeero1978
that benchmark is a marketing propaganda of ATI...

that benchmark very is biased...

I know,but check both findings in ATI and DH
Rings a bell?
Similarities aply?
http://img64.exs.cx/img64/6552/untitled713.jpg
http://www.theinquirer.net/images/articles/hal3.jpg

[Dr.Nick]

I can't find that on DH :retard: Anyone have a link or did they take it down?

[NoWayDude]

http://www.driverheaven.net/articles/Half%20life%202/

try to see if it works

[DSC]

Heaven of FUD and Lies? :lame: :lol2:

DH got :nanahump: :bash:

[Dr.Nick]

Thanks a bunch. It does seem they have taken it down and I'm guessing for a good reason. Best to read all the articles that come out on this

[bingo13]

The best HL2 benchmarks are here-

http://www.anandtech.com/video/showdoc.aspx?i=2278&p=1


Makes me wonder what Kyle was doing in his tests...... :rolleyes:

[fivefeet8]

The best HL2 benchmarks are here-

http://www.anandtech.com/video/showdoc.aspx?i=2278&p=1


Makes me wonder what Kyle was doing in his tests...... :rolleyes:

That's because HardCop used BFG OC 6800 cards which are overclocked out of the box. Anantech chose to use reference speed cards. Both perfectly fine IMO since that is what you can get at retail.

[holmes]

hehe.. nice try DH.
(mag)

[jbirney]

then DriverHeaven can be called Lies-Heaven?

lol


it seems that Nvidia ha really won this nv40 vs r420 battle

same performance in H.L.2 and F.C. but hugely better in Doom3


i can finally say my uncle what to buy this Christmas :P

No compared to other HL2 benchmark ATI is in the lead as we all expected them to be....

[fivefeet8]

No compared to other HL2 benchmark ATI is in the lead as we all expected them to be....

It really depends on which pricepoint your looking at. ;)

[Graphicmaniac]

6800GT still remain the best card of 2004

togheter with 6600GT :)

[NoWayDude]

Yes we would expect ATI to have a lead on this game.But is nothing compared to DH and
ATI themselves are trying to put out.
The X800PTE has at maximum a 20% to 25% lead (Anand dixit) with AA and AF cranked, that is easily explained by the memory bandwith advantage
The lower segment we see parity across the board (win some, lose some)
Is the X700 on Anand benchmark 128 or 256?And the 6600 is it the 128?
If so, ATI may have problems on that segment, because it looks like the Value cards, Nvidia is coming out with good products.

[jbirney]

Yes we would expect ATI to have a lead on this game.But is nothing compared to DH and
ATI themselves are trying to put out.

Consider this:

http://www.firingsquad.com/hardware/half-life_2_performance/page6.asp

Here they uses their own demo (which happens to be the same map as one of the ATI demos). And take a look at the 1600x1200 scores. 6800u is getting 43 FPS, the x800 PE is DOUBLE that. This is with out any AA/AF so we can not say its bandwidth, aa, optimizations (trylinear, ect). Which more or less backs up what ATIs PR crap said. You have to realize that every HL2 benchmark review is going to be different. They will use different maps, different things will happen, ect. You can not compare one with the other and say yeap, so and so is off. Also did you not think ATI was going to pick the best case and run with it? All PR is this way. The game plays just fine on all the cards so why bother complaining?

[Razor1]

Consider this:

http://www.firingsquad.com/hardware/half-life_2_performance/page6.asp

Here they uses their own demo (which happens to be the same map as one of the ATI demos). And take a look at the 1600x1200 scores. 6800u is getting 43 FPS, the x800 PE is DOUBLE that. This is with out any AA/AF so we can not say its bandwidth, aa, optimizations (trylinear, ect). Which more or less backs up what ATIs PR crap said. You have to realize that every HL2 benchmark review is going to be different. They will use different maps, different things will happen, ect. You can not compare one with the other and say yeap, so and so is off. Also did you not think ATI was going to pick the best case and run with it? All PR is this way. The game plays just fine on all the cards so why bother complaining?


Check driver versions

[dpagan]

Consider this:

http://www.firingsquad.com/hardware/half-life_2_performance/page6.asp

Here they uses their own demo (which happens to be the same map as one of the ATI demos). And take a look at the 1600x1200 scores. 6800u is getting 43 FPS, the x800 PE is DOUBLE that. This is with out any AA/AF so we can not say its bandwidth, aa, optimizations (trylinear, ect). Which more or less backs up what ATIs PR crap said. You have to realize that every HL2 benchmark review is going to be different. They will use different maps, different things will happen, ect. You can not compare one with the other and say yeap, so and so is off. Also did you not think ATI was going to pick the best case and run with it? All PR is this way. The game plays just fine on all the cards so why bother complaining?

they are using ati's beta drivers which where made for hl2 but they didn't use nvidia's hl2 beta driver 67.02, which IMO is not a fair comparison at all

copied from firingsquad

ATI RADEON X800 XT Platinum Edition
ATI RADEON X800 XT AGP
ATI RADEON X800 PRO
Driver version CAT 4.12 beta

NVIDIA GeForce 6800 Ultra
NVIDIA GeForce 6800 GT reference card
Driver version 66.93

by the way Anand's benchmark article also claims valve only gave nvidia two days with the game but gave ati a whole week and that can make a big difference in driver development for HL2

[jbirney]

Check driver versions

I realize that but no driver is going to add back 40 fps at high detial in a shader limited part of the game....

[jbirney]

they are using ati's beta drivers which where made for hl2 but they didn't use nvidia's hl2 beta driver 67.02, which IMO is not a fair comparison at all ....


by the way Anand's benchmark article also claims valve only gave nvidia two days with the game but gave ati a whole week and that can make a big difference in driver development for HL2

Anand bechies showed ATI was still "much" faster. Not the 40 fps faster. But enough. I agree not using the most current version was not the best apples vrs apples. But HardOCP used OC version of the NV cards. Which is fine but you have to realize not every 68xx sold is a OC card. Finally remember those Doom3 benchmarks that NV sponsered? ATI never even knew about the test. NV not only knew about it but had their own demos. However very few people here seemed to care back then about being fair.


Look the point is depending on the level and the time demo you are going to see varing scores. And so far ATI seems to be faster.

[holmes]

My question is does it really matter? If all top end card can pretty much do 1600x1200@4xAA/8xAF (provided you have the CPU)? And even if you don't, 1600x1200 with some AF pretty much seems to be a given. I know in my case, I'm so CPU limited anyway that it probably wouldn't matter what settings I was runing at, unless I was running at super high AA/AF @1600x1200 or greater. But, in my case, I don't even have a monitor that can do 1600x1200, so I could care less. As it is, this 6800GT can't fuly stretch its legs on my system... a card clocked 170MHz (core) faster would probably perform exactly the same.

just my thoughts...
:)

[NoWayDude]

Finally remember those Doom3 benchmarks that NV sponsered? ATI never even knew about the test. NV not only knew about it but had their own demos. However very few people here seemed to care back then about being fair.




This is wrong.ATI and Nvidia were both present when Carmack himself did the demo run.And ATI had a new driver then, and he did not make an ounce of diference.It still does not do it, even with Humus tweak (albeit it's much closer though)
In relation to HL2, we see discrepancies from diferent review sites.
Bottom line, the game runs great in any new gen hardware.It will improve with new driver releases from both vendors.

edit: link for [H] regarding Doom3 benchies as per Carmack himself
http://www.hardocp.com/article.html?art=NjQy

[Razor1]

I realize that but no driver is going to add back 40 fps at high detial in a shader limited part of the game....


so you're saying ATI seems to perform better with shader less intenstive parts of HL2? damn that goes against eveything you said before :) about pixel shader performance. Ok you can believe Firing squads benches if you want, but they seem to have always used maps and drivers slanted towards ATi, just like the Far Cry benchmarks where they only use outdoor scenes!