Half Life 2 Benchmarks @ HardOCP.com

[spyderz]

What's the point of this benchmarking when the NVIDIA products are tested OVERCLOCKED.. but the "matching" video card from ati is left stock?

so to be fair, why we just underclock to x800's to the same clock speeds as the nvidia's cards?

 

[CrimandEvil]

Good grief, some of the ATi owners in here are whinging so bad you'd think it was NVIDIA who had the fastest HL2 card!! It's all waaa-waaa-waaa no +r_fastzreject, waaa-waaa-waaa overclocked NV cards, waaa-waaa-waaa no 16x aniso, waaa-waaa-waaa no reflect all......

Totally agree, maybe because they're looking to burn NV owers like they where with D3? It's pretty lame since both cards run it ubber great.

World's smallest violin guys.

 

[fallguy]

I dont think the +r_fastzreject should be in the benches. But it would be nice if Brent could run it and post in the forums, just for curiosities sake. I dont have a problem with BFG cards being used. To silence all critics a stock speed card would be ideal. Would people have a problem with ATi enabling Overdrive by default? I think a lot would. 16x Aniso should be used in the "highest settings", just as reflect all should. Reflect all would be more demanding on the GPU, and I cant see any reason for it not to be used.

Just an oversight Brent, or am I missing a reason for it to not be used?

 

[spyderz]

Totally agree, maybe because they're looking to burn NV owers like they where with D3? It's pretty lame since both cards run it ubber great.

World's smallest violin guys.

my guess is they where thinking more of a whopass (ala doom3) since they didn't get it they're pissed

 

[Brent_Justice]

How's it apples to apples if its enabled on one and not on the other? Sounds like laziness to me...

its the default game settings

out of the box vs. out of the box on each card

how do you know if there aren't other commands that are enabled vs. disabled on each card?

 

[CrimandEvil]

my guess is they where thinking more of a whopass (ala doom3) since they didn't get they're pissed

Never believe PR.

 

[Brent_Justice]

True to an extent, I forced it thru radlinker though using the "force all levels" in AF setting to improve the visuals even though it's a bit of a performance hit.

And the AA thru control panel looks better for two reasons I think;

1. It applies AA to all the layers rather than just those the developers intend (better visuals but usually a bit of a performance hit)

2. Temporal AA

I'm not trying to argue, just explaining my settings better.

I have no clue how I'll set it up on the 6800, I'll have to do a bit of research since I'm a bit rusty on all things nVidia right now.

When we use games in our testing we always use the in-game AA/AF selections if available (since that's how the developer intended the game to be seen) and if the AA/AF options aren't there then we go CP.

 

[CrimandEvil]

Reflect all would be more demanding on the GPU, and I cant see any reason for it not to be used.
?

I do agree to an extent. If the game chooses not use it then thats how it should be benched but we are talking about max playable settings though.

 

[Brent_Justice]

That's just the thing....I think I know a bit about what I'm talking about having played with both and there just ISN'T a "best card" right now, it comes down to a matter of preference.

I still prefer ATi's AA over nVidia's, and temporal AA & ATi's regular monthly driver releases along with my personal favoritism towards ATi (oh hell yeah I'm biased, I just try to not let that influence my opinions too much or get in the way of the facts as they come) has decided me this round on which is better....but if someone is a Linux user or prefers nVidia for some other reason I really can't say they're wrong for it.

Personally I prefer it this way, everybody is a winner.

I have to commend you DW, that was an AWESOME post.

 

[Brent_Justice]

It's not laziness to not know about an obscure setting that only the lunatics who bloody live on B3D's boards would happen to have caught on to, it wasn't exactly posted on the main page.

Besides, I'm surprised/impressed that Brent & Kyle took the time off from playing to benchmark like this...I still haven't brought myself to that point yet, every spare minute I got I'm playing.

It's way too soon in this to call fouls or make accusation, 'specially against Brent who I personally feel works a bit harder than most to stay on top of these kinds of things.

The more reviews/benchies that come out, the more we'll all learn, the betterer we'll all get at it.

Patience padawan, patience.

Isn't HL2 awesome, i'm having SO much fan playing it!

 

[Brent_Justice]

Look at his and others posting history over the last few months and its clear some have a sort of angenda


Brent, completly understand why you did not test with the +r_fastzreject set to 1 on the ATI cards. However since its supposed to be turned on soon, and it may impact ATI (both postive/negitivly) then when you get time can you try it? You should realize the the readers here and on your forums are much more adept at tweaking things and this is something that 99% of the people reading could do and if it does bost/hurt FPS then it could be interesting to see....

I don't really have time to do a set of seperate benchmarks testing it right now.

Just got an SLI system in

When/If a patch is released for HL2 we'll of course use the new patch.

 

[digitalwanderer]

Isn't HL2 awesome, i'm having SO much fan playing it!

Yeah, the most fun gaming I've had since Half-life. I finished it last night about 4:30am, I haven't spent a night up gaming that I shouldn't in a long time and it felt just great!

I'm totally blown away by HL2, it's living up to the hype for me. I'm hoping to get over my ga-ga phase soon and get busy on benching some stuff, but I'm probably gonna just enjoy it a bit more first.

 

[Brent_Justice]

Right..
and If I sell an ati card with a 40% overclock out of the core out of my store, that makes for a fair comparison?

Just because BFG overclocks their card from THEIR factory (i.e. higher than what's on the nvidia actual reference board) doesn't make it not a "benchmark cheating" card.

I understand that the BFG card was overclocked. What my point was, is that, if the BFG card is overclocked, say. 7% above the "official" nvidia spec for the processor and core, then the fair comparison is an ati card at 7% over it's official spec. (the 7% is an arbitrary # of course)

If I wrote a comparison of an ati card that's 40% overclocked versus the BFG card that's 7% overclocked and the ati card did 30% better.. wouldn't that be "un-fair"?

I think it's an error in methodology, and the funny thing is, without overclocking "direct from factory" the ati card seems within what.. 4% ?

I mean.. a water block cooled, $600 video card versus a stock settings, stock fan/hs cooled $500 card and that nvidia can barely beat it?

Seems a bit telling to me.

I'm betting that if the bfg card was not OVERCLOCKED it'd actually SHOW a difference

we tested with RETAIL video cards using RETAIL clocks

I did not install coolbits and OC, i am using default clocks that the cards ship at, period

 

[Brent_Justice]

Point is with OCing... you can buy the card that way, it's the #1 selling card with that nVidia chip. If Brent OCed his card, it's not giving you the "stock" settings... yes, BFG cards are OCed, but the x800XT and x800XT PE are the same core... PE is overclocked. Hmm.

the PE card was NOT OC'd

it was running at default clocks as well

 

[digitalwanderer]

I don't really have time to do a set of seperate benchmarks testing it right now.

Just got an SLI system in

YOU LUCKY DOG YOU!!!!!

That's something I'd love to play with but I got the weirdest feeling it ain't gonna happen anytime soon...

I'm kind of curious if there is anything an SLI rig can do that a 6800u/X800xt couldn't right now, both cards just seem so CPU limited right now that two of 'em seems like total overkill.

Not that I wouldn't mind a little total overkill...

Thanks for the comments on the post, it just happens to be how I feel on the matter.

 

[Brent_Justice]

confused as to why reflect world instead of reflect all was used. try as i might i cant think of 1 logical explanation for this.

it was the default setting for each card that the game suggested and had selected

 

[Brent_Justice]

I agree 100%.

They're benchmarking graphic cards performance on Half-Life2. And why would they not push graphic cards the hardest they could with MAXIMUM visual quality settings ?

default out of box experience

 

[Netrat33]

it was the default setting for each card that the game suggested and had selected

Not that I don't like the review...BUT...doom3 default recommending setting was 800x600 for games too. Farcry had medium settings autodecting doesn't mean it's the best pick

 

[Brent_Justice]

I dont think the +r_fastzreject should be in the benches. But it would be nice if Brent could run it and post in the forums, just for curiosities sake. I dont have a problem with BFG cards being used. To silence all critics a stock speed card would be ideal. Would people have a problem with ATi enabling Overdrive by default? I think a lot would. 16x Aniso should be used in the "highest settings", just as reflect all should. Reflect all would be more demanding on the GPU, and I cant see any reason for it not to be used.

Just an oversight Brent, or am I missing a reason for it to not be used?

I used the default advanced in-game settings

the game was as default and out of box as it gets, save for resolution/AA/AF

 

[Brent_Justice]

I do agree to an extent. If the game chooses not use it then thats how it should be benched but we are talking about max playable settings though.

Ok, I'll give you that, for the two new graphs I added showing the Maximum AA/AF settings I could have selected Reflect All.

But those graphs were really meant to just max out AA/AF.

 

[Brent_Justice]

Not that I don't like the review...BUT...doom3 default recommending setting was 800x600 for games too. Farcry had medium settings autodecting doesn't mean it's the best pick

keep in mind this wasn't an indepth hardware guide like we did with DOOM 3

this was just a little quick sumpin sumpin

i felt default out of box was the best way to go that would appeal to most gamers

 

[Netrat33]

Ok, I'll give you that, for the two new graphs I added showing the Maximum AA/AF settings I could have selected Reflect All.

But those graphs were really meant to just max out AA/AF.

Oh well, can't change it now, it is what it is.

I don't THINK it would be that much of a difference anyway.

 

[coz]

Just got an SLI system in

Remember to add some SLi numbers to that HL2 performance review.

 

[Netrat33]

keep in mind this wasn't an indepth hardware guide like we did with DOOM 3

this was just a little quick sumpin sumpin

i felt default out of box was the best way to go that would appeal to most gamers

Try to get in the goods with the STALKER team for a full hardware guide
That's the first game I looked at (before playing...damn you farcry) and truly went...wow..I need to upgrade

 

[Brent_Justice]

Try to get in the goods with the STALKER team for a full hardware guide
That's the first game I looked at (before playing...damn you farcry) and truly went...wow..I need to upgrade

Stalker is definitely on our list

 

[CrimandEvil]

Try to get in the goods with the STALKER team for a full hardware guide
That's the first game I looked at (before playing...damn you farcry) and truly went...wow..I need to upgrade

Me too. I would think it's possible, remember that Valve was fine with doing a Hardware Guide..... last year when ATi had a performance lead.

I doubt the Stalker team really cares about all that BS, atleast I haven't heard any of them bad mouthing either company yet.

 

[coz]

I'm kind of curious if there is anything an SLI rig can do that a 6800u/X800xt couldn't right now, both cards just seem so CPU limited right now that two of 'em seems like total overkill.

You could probably use 8xS anti-aliasing at 1600x1200 with a decent average frame-rate.

 

[digitalwanderer]

You could probably use 8xS anti-aliasing at 1600x1200 with a decent average frame-rate.

If it could make that smooth, THAT would be something!

 

[fallguy]

Ok, so reflect world was default. However reflect all is more demanding, and when trying to stress the cards, why not use it? I under stand the "out of the box" experience, but stressing the cards with a higher in game setting seems a logical choice. I imagine lots of users would change it, to reflect all for better graphics. Just my opinion.

 

[Visable-assassin]

mmmmm....Stalker...
On a side note hows that SLI setup brent? Initial thoughts? first 5-10 minutes of gaming on it?

 

[Moloch]

I dont think the +r_fastzreject should be in the benches. But it would be nice if Brent could run it and post in the forums, just for curiosities sake. I dont have a problem with BFG cards being used. To silence all critics a stock speed card would be ideal. Would people have a problem with ATi enabling Overdrive by default? I think a lot would. 16x Aniso should be used in the "highest settings", just as reflect all should. Reflect all would be more demanding on the GPU, and I cant see any reason for it not to be used.

Just an oversight Brent, or am I missing a reason for it to not be used?

Ati's override hardly overclocks the card, its only a tiny core boost.
Fastzreject should make a big diff based on forum posts, I have no problem with brent not using it, since he didn't know about it.
He said he'll do another test when the patch comes out, and that's good enough.

 

[NoGodForMe]

Great review. I'm enjoying HL2 at 1600x1200 with frame rates always near 100.
Anyone with an older card, upgrade, make that your early Christmas present or Black Friday gift for next week.
D3 was indoors and dark, so you couldn't see much. But HL2 is outside, and the views of the city are incredible.

I have one question. In the advanced display settings of the video card, do you have the slider set to High Performance or Quality? I assume Performance, because it can make a differance of 10 FPS. I have both boxes checked on application controlled.

 

[Jonsey]

Great review. I'm enjoying HL2 at 1600x1200 with frame rates always near 100.
Anyone with an older card, upgrade, make that your early Christmas present or Black Friday gift for next week.
D3 was indoors and dark, so you couldn't see much. But HL2 is outside, and the views of the city are incredible.

I have one question. In the advanced display settings of the video card, do you have the slider set to High Performance or Quality? I assume Performance, because it can make a differance of 10 FPS. I have both boxes checked on application controlled.

The textures look like ass in performance mode. Ever noticed the difference in the loading screen in 3dmark05 between performance and quality? Not worth the IQ tradeoff IMO.

 

[ShePearl]

default out of box experience

Out of the box experience? LOL!
It's retarded for any self-respecting benchmarker to publish a graphics card benchmark which is running CPU-limited tests.
Why haven't you written benchmarks based on "out of box experience" for Far Cry, DOOM3 etc ?

hello McFly?

Nice touch there.

 

[CATALYST MAKER]

Me too. I would think it's possible, remember that Valve was fine with doing a Hardware Guide..... last year when ATi had a performance lead.

*Had* a performance lead?

Ohh you must be talking about being tied in CPU bound cases....

 

[CrimandEvil]

*Had* a performance lead?

Ohh you must be talking about being tied in CPU bound cases....

LOL, I was beening cheeky with the Valve part. not the ATI lead part. I've got some issues with Valve and the way they've acted for the last year or two.

 

[CATALYST MAKER]

LOL, I was beening cheeky with the Valve part. not the ATI lead part. I've got some issues with Valve and the way they've acted for the last year or two.

Why do you feel that way?

Didnt they deliver the finest game ever to grace monitors in the history of PC gaming?

 

[tranCendenZ]

*Had* a performance lead?

Ohh you must be talking about being tied in CPU bound cases....

I'm actually quite surprised ATI doesn't have a significant performance lead while operating at 100mhz faster clock speed and at lower shader precision with less features on chip.

 

[Brent_Justice]

Ok, so reflect world was default. However reflect all is more demanding, and when trying to stress the cards, why not use it? I under stand the "out of the box" experience, but stressing the cards with a higher in game setting seems a logical choice. I imagine lots of users would change it, to reflect all for better graphics. Just my opinion.

Maybe they would change it, maybe they wouldn't, I dunno.

We tested with it the way it was because its the default setting, and we wanted an out of 'box' experience. That's the reason. We feel that the way we set it up lets gamers know how the cards compare in HL2 when installing it, and playing the game.

 

[Brent_Justice]

mmmmm....Stalker...
On a side note hows that SLI setup brent? Initial thoughts? first 5-10 minutes of gaming on it?

Can't Tell you Yet. Look forward to an article in the future.