Real HL2 benchmarks

[BenSkywalker]

I think everyone here is aware that firing squad and a couple of others got results similar to DH, but don't really like to admit they were wrong.

Or perhaps the sites showing ATi with a commanding lead decided to use ATi's beta HL2 optimized drivers but decided not to use nVidia's beta HL2 optimized drivers.

 

[Son of Thunder]

look at the picture i posted again, they've got the nvidia running at 4xsaa 16xaf, 6xsaa 16xaf, and 8xsaa 16xaf, all of them are about half of what the x800 was getting. also benskywalker is correct, driverheaven used the latest official release for nvidia and the beta 4.12's for the ati. they only did that because the beta 4.12's were officially released to the public but you can't get the latest beta nvidia's off from the nvidia website.

i personally believe that hardocp, while still a great reviewing site, screwed up by doing these benchmarks at lower quality settings. of course all of the cards will be equal if you don't really tax them! at 1024x768 i believe it's safe to say that the 9800pro might beat the x800pro by a fraction of a fps.

who knows, call me an ATi fanboy if you like, i've owned plenty of ATi cards and a lot more nvidia cards. i threw in my money with the x800xt this round, that's what i bought. when the doom3 benchmarks came out nvidia owners spit in ati owners faces and really flaunted their new cards. well, now a benchmark comes out that shows that at low quality settings the 6800 series and the x800 series are equal, and once again nvidia owners spit in ati owners faces. what you all are neglecting is that at high resolutions, the x800xtpe blows the nvidia flagship card out of the water, even with the newest nvidia drivers.

 

[Cat]

Normally 'SAA' means super-sampling, which is NOT multi-sampling, and is much slower.

 

[BenSkywalker]

look at the picture i posted again, they've got the nvidia running at 4xsaa 16xaf, 6xsaa 16xaf, and 8xsaa 16xaf, all of them are about half of what the x800 was getting.

ATi's PR department won't allow their cards to offer SSAA because of benchmark results like those posted @H. SSAA is vastly superior and significantly more intensive then any mode ATi can run on any of their current parts. The xS mode is offered for legacy titles like CounterStrike where nV utterly humiliates ATi's parts because of their lack of any super sampling modes(load up Italy with 6x AA on your X800 and check out the railings- aliased badly as they have no AA applied at all). Right now ATi does not have FSAA on any of their current parts- only MSAA. H's comparison should have explicitly spelled this out but they aren't terribly interested in reporting accurately there nor have they ever been.

 

[ronnn]

If I understand Ben, HL and 6800 owns quality graphics, but hl2 and x800 supplies slightly higher quality. So if you play legacy games buy a 6800 series and if you play newer titles such as hl2 buy x800series?

Fail to see the problem for ATI pr folks.

 

[BenSkywalker]

If I understand Ben, HL and 6800 owns quality graphics, but hl2 and x800 supplies slightly higher quality. So if you play legacy games buy a 6800 series

Not sure, haven't seen any high res HL2 benches yet so I don't know who ends up ahead in HL2 when you up the settings. If you are talking at the limit performance then nV wins across the board, they have nothing comparable to SLI so I guess I'm not sure what you are getting at here. ATi's 6x AA is the best method of reducing edge aliasing without taking too much of a performance hit as long as you are using a title that doesn't use alpha textures to simulate geometry? If that's what you are saying then I would agree.

Too bad sites stop testing at the middling 1600x1200, it would be nice to see a high end board running a setting fit for a high end display(2048x1536 would be ideal, but even 1920x1440 is better then 1600x1200). Of course, if you play legacy games you really shouldn't consider ATi anyway. When they have problems that break legacy titles they won't do anything about it unless it causes them negative PR, and that isn't too common(although their breaking HL1 engine a while back got the community in an uproar).

Fail to see the problem for ATI pr folks.

If it effects their PR then they will fix it. If it is limited to a couple hundred or couple thousand of their customers they won't even acknowledge you exist. Even if they know exactly what the problem is and how to fix it, they won't bother.

 

[Son of Thunder]

yah, the hl engine was a weird thing for ati, my now dead radeon 9000, my mobility 9700, and the mx440 i just sold all got 100 fps with the most recent drivers, but my old 9800pro, and x800xt get between 70 and 100. i'm still a firm believer that the mx440 and the ti4200 are the best cards for the original HL engine.

in case people were wondering (i am kinda new here) these are the different vid cards that i've owned at different times: geforce 2 32MB onboard, sis650 64MB onboard, radeon mobility 9700 64MB, nvidia vanta tnt2 8MB, mx440 128MB, sapphire radeon 9000 128MB, radeon 9550XT, radeon 9800pro 128MB and 256MB, geforce 6800 (on loan), geforce fx5600xt 256MB, radeon x800xt (currently involved in a legal issue, aka not in my hands)

out of all of those i still have the vanta, the two onboards, the r9000 (which died a month ago, but it was my first video card purchase, so i kept it) and the 9550XT aka 9600XT w/ lower proc. clock.

as far as my x800xt goes, well, never buy from xpcgear.com... it came in a ESD bag that had been torn and was sitting in a bunch of packaging peanuts, crashed a lot, got bad performance, had a glitched boot and lots of artifacts. i sent it back to xpcgear.com and they say it's working fine, but has a few "minor" problems. they won't give me my money back and won't send a replacement. and i won't accept my faulty card back, so they're under both a BBB investigation and my credit card company is looking into it.