[H] NV40 Review Sucks

[mhayenga]

Well, the title kind of says it all. I was really looking forward to the hardocp review, but overall... I feel it is a big letdown. They're non-consistent use of settings makes it hard to see how large the performance increases are. And in both UT2004 and Farcry they claim the 6800 offers little increase over a 9800xt... But according to anand's review, at most settings it offers near 2x performance in both games.

Anand Farcry
Anand UT2004

Something is strange/fishy.... I was insanely eager to see the [ H ] review, but now I feel kinda let down, and it seems the vast majority of other reviewers are disagreeing with [ H] .

 

[Genx87]

Kyle is a child. You cant expect children to have a clue.

I have heard about this review and wont even bother reading it.

 

[Draco]

I agree. I found it very skewed. Then again, they're ATI lovers.

Tom's and FS I felt were the best.

 

[YBS1]

You bet it sucks. I read it first as Anand's wasn't up yet, and I was left with a "This is what all the hype has been about? This thing is a steaming pile for a next gen 500 dollar card" impression. I like that Anand's included the 9700 pro, gives me a better idea of the increases I should expect. Looks like I'll be getting something very soon, now let's see what the R420 brings to the table.

 

[Fausto]

Wha? Did you see the benchmarks on Firingsquad's review of the 6800? It takes the 9800 back behind the woodshed at higher resolutions and/or with AA.

 

[Dman877]

They've been doing that "best playable settings" thing in their reviews for a while. I'm kinda so-so about it. On the one hand, you can see what's playable and what's not. It makes sense because no one here cares about fps over 100, you max out your graphic quality and try to keep minimum fps in the 30 to 40 range. On the other hand, how do you determine which settings make for the best graphics? Sometimes they used high res and low aa/af sometimes high af/aa but low res, sometimes high af but low aa, sometimes high aa but low af.

IMO, they should do both, test cards based on the maximum playable settings like hardocp did, and test at the same settings to get a real comparison. They also need comparative screenshots for every test if they continue to do it this way. Then they could say if your minimum desired frame-rate is xx frames, with card A this is what the game looks like, with card B, this is what the game looks like. Then the screenshots would tell the story by showing the 6800 running 12x10 w/aa/af which, in image quality, would blow the 5950/9800 away running 10x7 w/o aa/af.

 

[mhayenga]

Originally posted by: Fausto
Wha? Did you see the benchmarks on Firingsquad's review of the 6800? It takes the 9800 back behind the woodshed at higher resolutions and/or with AA.

No one here is disagreeing with that... (people referring to suckage have been complaining about hardocp, not the nv40).

Just clearing it up

 

[YBS1]

Originally posted by: Fausto
Wha? Did you see the benchmarks on Firingsquad's review of the 6800? It takes the 9800 back behind the woodshed at higher resolutions and/or with AA.

I was responding to the thread title..."[ H ] NV40 Review Sucks". Not the 6800.

 

[YBS1]

I'm not a fan of the way they are doing the videocard reviews at all. I understand what they are trying to do and it is a valid intention. However, it's not very real world. I (and would think most people) start with very lofty video settings and "tune-down" to what I consider an acceptable eye candy/framerate compromise. The way they are doing it is too "black and white" to be worthwhile for me, only messing with resolution, AA and AF. I'll turn down other things less noticable to me before reducing an AA or resolution setting I'd like to use (shadow detail, particle effects, or maybe something else).

 

[WicKeD]

Ya I agree. I am a fan of the HardOCP, but this review was total crap. I can decide what the "playable" settings are when I play the game. I want to see a straight comparison.

On the other hand, there may be people who wanted to see a review like the HardOCP. Thank god there is more than one review to choose from.

 

[Lonyo]

The whole point is that EVERY OTHER SITE HAS THE SAME REVIEW STYLE.

If you want a "normal review, you have a hell of a lot to choose from! Which b!tch about 1 sitre doing something different, there are more than 10 doing it the way you prefer, look at one of those.

HardOCP has an alternative, you don't have to read their review, nor will you be without info if you don't, the info is on many other sites, as HardOCP stste themselves in the review!
Why bother doing 2 lots of tests? Other sites have done the other tests for them!

HardOCP shows the 6800 beating both the other cards, how is that "disagreeing"?
All the sites have different margins, and different system setups.

The whole point is to look at all the reviews and see on average which is best.
Anand seems to give a lower margin in many tests to the 6800 than other sites.

 

[YBS1]

Lonyo, like I said...

I understand what they are trying to do and it is a valid intention.

I just don't think the way they are doing it works all that well, and honestly I think it would be hard to do it the "right" way.

Their review left me with an initial bad impression of the 6800. However, when I look at the other reviews that show some of the large gains that have been made in certain areas at various resolutions, and with varying levels of AA/AF I can automatically deduce roughly the exact same thing they are trying to show with their apples-to-oranges comparison. I will commmend them for showing us min/max/avg framerates though, that should be the norm everywhere.

 

[g3pro]

the problem with HardOCP is that their review design is beyond logic: they use 1 graph and use 2 variables. THAT IS NOT WHAT A F***ING GRAPH IS SUPPOSED TO BE LIKE!!! his reviewing style shows how amateur and infantile Kyle is. when i first saw the benches, i was like "WTF?", but as you look into the details, it just doesn't make any sense.

the reason why all review sites are similar is because they have to be! they do standard data analysis unlike HardOCP. compare it to a benchmark of a GF4MX440 and an ATi Rage 128, and an ATi 9800XT. put it on "playable settings" (kyle is so f***ing stupid...), and guess what??? the ATi RAGE 128 beats the 9800XT!!!!!! OMG!!! then again, the 128 is on 640x480x16, while the 9800xt is on 1024x768x24.

 

[Killrose]

I like HardOCP's testting method with the graph and giving you a min/max frame rate, ect., but they really need to keep it apples to apples with all settings. They should have just added more graphs at different settings, which would have kept it easier to read, and now they appear to be an Ati fanboy site.

 

[CaiNaM]

the problem with their methodology is that there are more settings than "min" and "max".. there are many ways to 'tweak' for the most acceptable combination of performance and IQ; not to mention their opinion of what's "playable" is at best subjective. we all have our own preferences. it also makes it difficult to put the performance differences into perspective.

on top of that, their reviews are often quite opinionated. reviews should present the info and let the reader make their own opinion.

i do like the fact they include min and max fps however; all reviews should do that.

 

[PrinceXizor]

Originally posted by: g3pro
the problem with HardOCP is that their review design is beyond logic: they use 1 graph and use 2 variables. THAT IS NOT WHAT A F***ING GRAPH IS SUPPOSED TO BE LIKE!!! his reviewing style shows how amateur and infantile Kyle is. when i first saw the benches, i was like "WTF?", but as you look into the details, it just doesn't make any sense.

the reason why all review sites are similar is because they have to be! they do standard data analysis unlike HardOCP. compare it to a benchmark of a GF4MX440 and an ATi Rage 128, and an ATi 9800XT. put it on "playable settings" (kyle is so f***ing stupid...), and guess what??? the ATi RAGE 128 beats the 9800XT!!!!!! OMG!!! then again, the 128 is on 640x480x16, while the 9800xt is on 1024x768x24.

Can you have an opinion without being a troll? I think its clear that many differ on their opinon of HardOCP's videocard reviews. But it is a differing viewpoint (one I personally disagree with), but I can state so without trolling. Can you?

I understand what Kyle is TRYING to do. But what he is trying to do, actually removes itself from the realm of "benchmarking". When you aren't comparing apples-to-apples, it no longer is benchmarking. What he is doing is more of a value analysis (money/performance vs. what will actually be used) as opposed to a statistical analysis (which is what benchmarking is). That's my two cents.

P-X

Edit: Either I missed this the first time, or its new. Either way, it spells out HardOCP's stand quite clearly. And after reading it, I have no qualms at all with the "review". The best thing about any analysis is clearly stating your methodology and objectives up front. Kyle has done that and I applaud him for it.

Warning!

Be aware that not all results represent "apples to apples" comparisons. We try to find the highest playable resolution, anti-aliasing and anisotropic filtering settings for each card in each game. Keep that in mind as you look at the graphs and find that each card may be set at a different resolution, AA and AF level. What we strive to illustrate to our readers is consistency in performance and which cards provide the highest level of gameplay performance. Should you need apples to apples benchmarks that focus on average framerates, there are plenty of other hardware sites to deliver that content to you.

HardOCP.com "video card reviews" have stopped being video card reviews. We are truly trying to measure the level of gaming experience the hardware can provide. Average frame rates only tell the real gaming experience story if every frame is "average," and we all know that is far from being the case.

That being said, it is HardOCP's goal to provide our readers with information that should help you understand what level of gameplay experience would be provided by the hardware in a similar computer of your own.

 

[Insomniak]

Be aware that not all results represent "apples to apples" comparisons.

Then, pardon my French, but what is the fucking point of posting the results then?

I mean, the SOLE PURPOSE of benchmark/game testing is to show the differences in hardware ability under identical conditions. Jesus Christ, I didn't think the guys over at [ H ] were so stupid.

 

[IHYLN]

hardocp is okay but kyle is a total clownshoe when it comes to reviewing video cards

 

[DAPUNISHER]

I'll bet dollars to donuts that [ H ]'s "testing methodology" or whatever it is does not catch on elsewhere It's nonsense plain and simple. It might help a complete neophyte find playable settings for games regardless of the card they use but for everyone who actually reads these sites frequently it's ludicrous :disgust: BTW, understanding their intent does nothing to ameliorate the feeling I have that they are trying to manipulate data to further a personal agenda.

 

[eastvillager]

I actually like what they're attempting to do, raw framerates are about as useful as penis measurement when it comes to measuring real world performance.

 

[DAPUNISHER]

Originally posted by: eastvillager
I actually like what they're attempting to do, raw framerates are about as useful as penis measurement when it comes to measuring real world performance.

WTF are you talking about? If the FPS charts using 10x7 through 16x12 both with and without AA/AF in real games that all the other sites are using doesn't indicate real world performance what does?!?

 

[nRollo]

Originally posted by: WicKeD
Ya I agree. I am a fan of the HardOCP, but this review was total crap. I can decide what the "playable" settings are when I play the game. I want to see a straight comparison.

On the other hand, there may be people who wanted to see a review like the HardOCP. Thank god there is more than one review to choose from.

Yeah, there are people who want to see different, hand picked settings that make the 6800 appear equal to the 9800 XT.

BFG would say I'm probably one of them for my statements that a 9700Pro and 5800U are the same at 10X7 and 12X10 4x8X.

Most of the rest are over at Rage 3d.

 

[PrinceXizor]

I think their idea has merit (in and of itself), but their implementation of it is not clear and precise. When you are moving into a relativistic comparison, the clarity of the approach and presentation is of utmost importance, else the article comes off as one big editorial piece.

In other words, they need to define what their criteria is for determing what a "playable" scenario is. That's where most of the controversy lies. (There would still be controversy, of course, because the whole concept is subjective, but you would still have a stable measuring line). For example, you could say that a "playable" scenario is one in which the Minimum frame rate is above 30 fps for the entire test. Then they could run their gamut of tests and configurations. Then you arrive at a conclusion (or more appropriately, several). What I don't see is a clearly defined measuring line. The means for arriving at said measuring line may be subjective in this type of "testing", but once it has been determined, the results become objective.

In this type of scenario, you could actually have a quite complex formula for determing your "measuring line". You could have a weighted scale of factors, such as minimum frame rate, maximum frame rate, average frame rate, total allowable downward spikes, etc.

To summarize a long post:

I think that the method has merit as a different way of approaching assesment of videocards, but, since the method is a new one, the approach and presentation of this method is not up to the levels it needs to be.

I just picked up my two cents and depostited a quarter

P-X

 

[tazdevl]

I think what they did was admirable.

They stepped out of the norm and it does show how impressive the card is. Jebus, the card can run AA/AF at a higher framerate than the NV38 and R360 at the same resolution without the eye candy. That's qualifies as noticeable in my book.

Looking at a raw number these days isn't terribly helpful... I could care less that a card can get 400 FPS in Quake 3 @ 1600 X 1200. But riddle me this... how does it do with the goodies turned on?

As to the whole bias issue mentioned, I think Brent does a solid job keeping things objective. Let us also not forget the blind pro-nVIDIA bent Anand has had despite the craptastic products they have put out over the last 2 years. I think Anand has done an excellent job being Brian Burke's unfiltered mouthpiece.

As for me, I'll be interested to see how the R420 stacks up and will make my purchase decision based on IQ, performance and cost when the cards hit the streets.

 

[Alkali]

Originally posted by: PrinceXizor
I think their idea has merit (in and of itself), but their implementation of it is not clear and precise. When you are moving into a relativistic comparison, the clarity of the approach and presentation is of utmost importance, else the article comes off as one big editorial piece.

In other words, they need to define what their criteria is for determing what a "playable" scenario is. That's where most of the controversy lies. (There would still be controversy, of course, because the whole concept is subjective, but you would still have a stable measuring line). For example, you could say that a "playable" scenario is one in which the Minimum frame rate is above 30 fps for the entire test. Then they could run their gamut of tests and configurations. Then you arrive at a conclusion (or more appropriately, several). What I don't see is a clearly defined measuring line. The means for arriving at said measuring line may be subjective in this type of "testing", but once it has been determined, the results become objective.

In this type of scenario, you could actually have a quite complex formula for determing your "measuring line". You could have a weighted scale of factors, such as minimum frame rate, maximum frame rate, average frame rate, total allowable downward spikes, etc.

To summarize a long post:

I think that the method has merit as a different way of approaching assesment of videocards, but, since the method is a new one, the approach and presentation of this method is not up to the levels it needs to be.

I just picked up my two cents and depostited a quarter

P-X

Amen

Nothing wrong with them doing something different, its just that they need a consistent level at which 'quality' is judged.

The problem is that in one graph it is 2xfsaa,4xaf, and in the very next review 4xfsaa,16xaf - the problem is simply changing the settings for each specific individual card at a whim, which ends up seeming kinda pointless to most people. Over many years looking at graphics card reviews, I have grown to understand the effect of fsaa and af, depending upon the architecture, on scores represented with 0xfsaa and 0xaf, so I myself also, prefer to view scores with no enhancements, or indeed 'apples to apples'.

As a side note: If I buy this 6800 Ultra, I will be running consistently at 1280x1024 with 4xfsaa and 16xaf in every game. This (according to the results) would give me between 50fps and 120fps in any game, and this is heaven for me because im a fsaa and af fanboy