The Discussion- Do Numbers Lie ?

[Zurvan]

Ok, I have been around this forum (and in general the tech community) for quite some while, and from what I have seen, this is NUMBER thing. You have benchmarks to tell whats fast, fps to tell which is better -- but often I come across instances when people argue numbers dont justify.

Its happened with processors too, but with reference to gaming (where experience is more perceivable) GPUs are the main argument of this discussion.

Lets take an example from the latest gen-: the 4870 (the red card) and the GTX 260 Core 216(the green card), as thats what inspired me for this thread.

First thing --i own neither. I AM going by NUMBERS and info that OTHERS have provided.

The red performs almost equivalently according to most review sites. But its often bashed by members saying the green card has perceivably more smoothness in gameplay. And I have read that many times. Is it possible to justify this statement ?
(Here (http://www.anandtech.com/video/showdoc.aspx?i=3523&p=4), AT shows the red card winning 5 out of 7 vs the green at 16x10 :confused:)

On the other hand, review sites show the red card at an advantage when AA is enabled, and it scales very well with an increase in AA. Does it really give better eye candy at more smoothness of gameplay ?
The cold reality here is that, for most intents and purposes, current GeForces and Radeons are more or less functionally equivalent, with very similar image quality and capability, in spite of their sheer complexity and rich feature sets. I would gladly trade any of Nvidia's so-called "graphics plus" features for a substantial edge in graphics image quality or performance. The GTS 250 comes perilously close to losing out on this front due to the Radeon HD 4850's superior performance with 8X multisampled antialiasing. The saving grace here is my affection for Nvidia's coverage sampled AA modes, which offer similar image quality and performance[

The red is said to run hotter, make noise ,

I would easily tag this as fanboism but having come across so many instances its diffcult to believe that. Does buying a particular card really make us ignore/criticize/jealous of the other team ?

Whom to trust ?

Please pour in your thoughts.

Note - I start this thread with an intention to get a fruitful discussion and not FLAME WARS. Please dont BASH either side.

Also, I am not asking for buying advice. Discussing price points is irrelevant here.

Thank You.

[CGR]

So what are you asking then? Some will say one side is better, others will disagree.

I can say this. For the most part I tend to lean towards review sites and their numbers. The reason is, the better sites always use fresh installations and same hardware for their comparisons. This gives you the benefit of seeing how cards fare against each other with the rest of the system being identical.

When people give their opinions, they cannot always be trusted and ill explain why.

Lets say someone likes Green over Red because his system runs better for him with Green. Thats fine and all, but perhaps the person didnt install Red properly, or did a fresh OS install with Green and not Red.

There are too many variables when people post their opinions on performance, whether its GPU or CPU due to everyone having different hardware configurations and loaded down OS installs.

Now there are a few people on here, who actually take the time to test cards and will follow the review sites methods when testing, i.e they use fresh OS installs and dont install competitor drivers over each other, those are people and numbers I trust.

[Rinne]

The thing about most review sites (for graphics cards) is that they mostly show average frame rates only, but that's just not the whole story.

What really matters is the minimum frame rate (very few review sites test that), and how long a card stays in a low framerate area (I don't know of any review site that measures this).

Also the measurement of micro stutter would be important (defining how "smooth" the gameplay is even at high fps), but I have only seen one such measurement in my entire life and that was at xtremesystems forums.

As a general rule of thumb I always read review sites and then wait for personal opinions of users to arise. If they are generally good you can't go wrong with either team.

[ShaggyTDawg]

There are too many variables when people post their opinions on performance, whether its GPU or CPU due to everyone having different hardware configurations and loaded down OS installs.

I agree with this. There are too many variables that can affect performance. Unless two cards are compared in identical conditions, using straight numerical metrics (Some kid telling you his 3DMark score and some FPS numbers) can't be used as reliable definitive information.

As a general rule of thumb I always read review sites and then wait for personal opinions of users to arise. If they are generally good you can't go wrong with either team.

I also agree with this and tend to practice a similar method. You could also apply the individual information provided by people who match CGR's quote and add that to your overall "public/in the wild opinion" of a card. Combining numerical metrics from reviews sites with well founded comparison methods with the general consensus of a card is a good way to get a general feel for how the card will do.

The numerical metrics generally won't account for things like how much one chipset will stutter when AA is turned on vs. another. They also generally won't point out commonly occurring bug/issues with compatibility as they'll usually only test it on one motherboard/setup.

I generally look for a configuration that many have done (I have an EVGA 780i w/ a Q6600 and an EVGA 8800GTS (G92)... wonder how many thousand people have or have had that setup) and stick with that since I know there are many others who believe the performance of that configuration is ideal(at that point in time).

One thing I think shouldn't be done as much is comparing two cards by two different vendors. I see posts often that say something like "Should i get the HD4870 or the GTX 260 Core 216!?!!". As long as the two cards in comparison are more or less the same generation and performance levels, there's no sense in trying to compare. The noticeable difference between the two will likely be nearly negligable, otherwise they should be compared to begin with. At that point it should just boil down to which company do you prefer or can you possibly get a better deal on one or the other.

[x30Jx]

"There are 3 types of lies: lies, damned lies, and statistics."

[wingman99]

Green is smoother then the red, I have both and when you move side to side like in a first person shooter or race car game you get micro stutter or stutter, so it's blurry on Red and the FPS is over 80 when it's stuttering. I spent years testing this phenomenom on new ATI cards and i think it has to do with ATI's memory controller.

You won't see that in any numbers review, what do I know I only play games all the time.:beer:

[Cuiiey]

well i wouldn't always trust the people who post here. They could be wrong or post what they have heard

[chevro1et]

When it comes to review sites, you have to take some of them with a grain of salt as well. Some of them may have some sort of affiliation with one side or the other, and can skew results in either direction. I have dealt with enough stats in my life to know that you can always manipulate raw data to support whatever you want.

[Zurvan]

Fanboism is the worst disease to hit the tech community ever.
And then we have famous review sites, now you cant even blame them for skewing the results, the companies are skewing the cards :bang head

Green is smoother then the red, I have both and when you move side to side like in a first person shooter or race car game you get micro stutter or stutter, so it's blurry on Red and the FPS is over 80 when it's stuttering. I spent years testing this phenomenom on new ATI cards and i think it has to do with ATI's memory controller.

You won't see that in any numbers review, what do I know I only play games all the time.:beer:

Yes i was looking for something like that as a proof. :thup:
Should you be trusted or the Anandtech :beer:, because read nowhere I have what you write. :santa:

[don'tknow]

When it comes to review sites, you have to take some of them with a grain of salt as well. Some of them may have some sort of affiliation with one side or the other, and can skew results in either direction. I have dealt with enough stats in my life to know that you can always manipulate raw data to support whatever you want.

Fanboism is the worst disease to hit the tech community ever.
And then we have famous review sites, now you cant even blame them for skewing the results, the companies are skewing the cards :bang head


Those are the reasons I don't trust most review sites, and there's also the factor of different games being optimized for different GPU's. Many nowadays are developed on and optimized for Nvidia. And rigged results are very easy (at least, for me) to notice on hardware review sites.

Fanboyism is a disease pretty much everywhere, since humanity is full of brainwashed 'tards. Whether it's a person, a piece of hardware, or a piece of paper, pretty much anything can be an idiot magnet. :/

[wingman99]

Green is smoother then the red, I have both and when you move side to side like in a first person shooter or race car game you get micro stutter or stutter, so it's blurry on Red and the FPS is over 80 when it's stuttering. I spent years testing this phenomenom on new ATI cards and i think it has to do with ATI's memory controller.

You won't see that in any numbers review, what do I know I only play games all the time.:beer:


Yes i was looking for something like that as a proof. :thup:
Should you be trusted or the Anandtech :beer:, because read nowhere I have what you write. :santa:It is true I wish all my competitors were using ATI, i'm not saying Nvidia is perfect when sweeping 90, 180, 360 degrees or around a turn, however it's a night a day difference between Red and Green.

You also mentioned AA red seems handle AA better and it looks better than green, however I don't car about AA that much, I play for the win.

When I read this post I could see you were doing your homework vary well, take it one step further and see for your self, you will appreciate the risk like I do trying both cards and having to return or sell the RED. I always try RED and GREEN when they change the architecture to see if i'm missing out.:beer:

[liftedcj7on44s]

I owned a HD4870. Was my first ATI card ever. In all my games the framerates were great but it was choppy. My old 7800gt(green guy) was always smooth. My 8800GTS(G92) was smooth as butter and my 2 gtx 295's even with a little bit of microstutter were still smoother than the ATI card. The problem does not lie in the cards. The problem lies in the drivers for ATI. If ATI can ever get there driver program straight then i have no doubt they would kick some butt.

[wingman99]

I owned a HD4870. Was my first ATI card ever. In all my games the framerates were great but it was choppy. My old 7800gt(green guy) was always smooth. My 8800GTS(G92) was smooth as butter and my 2 gtx 295's even with a little bit of microstutter were still smoother than the ATI card. The problem does not lie in the cards. The problem lies in the drivers for ATI. If ATI can ever get there driver program straight then i have no doubt they would kick some butt.If you know exactly what the problem is why don't you help them fix it, pleas tell me why it's the drivers i'm all ears.

Here some food for thought, who cares if it is the drivers or video card i've seen this problem since ATI's 9800 pro all they way up to ATI 4870 when they get it fixed I will let you know.

[liftedcj7on44s]

why is it every time i respond to a post on here someone gets defensive? I didnt say i could fix there problem, Nor do i really care to fix there problem. But i honestly believe there drivers to be at fault.