3D Mark: Bias or no?

[Phoenix64d]

I've read a number of snippets about 3dmark favoring this processor more or that GPU more. I've also heard rebuttals countering this. You'd think it would perform better on the system it was developed on, but that might not be the case. Any hard evidence that it does or doesn't?

Also, do any find that other benchmarking programs return more stable, repeatable results? Everyone knows how easy it is for product to monopolize an industry just because of its brandname, regardless of its quality. I for one never knew that other companies made mp3 players until I got ready to buy an iPod.

Note: Before someone accuses me, I'm not trying to attack any one manufacturer. Just a discussion. ^^

[johneva916]

Na not from my experiance this is simpley excuses fan boys use when their fav manufacturer comes out bad on benchies.

Thing is though you should not only take notice of 3dmark, I personally find it more important to see benchmark results from games I use at the res I use, these are more important than any 3dmark result.

But you can normally get a rough idea on what is a more powerful setup from running 3dmark and it is a very well made benchmark which generally will be backed up results wise when game benchmark results are also done.

So no IMO not bias simpley limited in what you can take from the results of a single benchmark program, even though it is an exellent one.

[Zefram]

Futuremark, being a business establishment, have all the reason to be BIAS towards company that did not contribute to them.

However, so far, futuremark never shown signs of any bias, and most of the cheats and stuff originated from graphics drivers........

[johneva916]

Thing is if it were bias then what would be the point in the benchmark cos you could not rely on the results and nobody would use it.

I have just realized how nieave I can be at times, I did not even realise there were any cheats for 3dmark, I never really gave it a thought why would you want benchmark results that are fake? Simpley so you can go look my numbers on a screen are higher than your numbers on a screen, even if your computer is quicker than mine and you can play games I carnt lol, great numbers on a screen will get you far when it comes to wanting good frame rates in your new favorate game, or when your trying to get something like a video encoded quicker cos you have loads of them to do. Lame lame lame.

It is the same theory as people who use hacks in multi player games they are so annoying why bother? So they can go oooo look at me am so so great even though in reality they suck ass need hacks to be any good at all. It must be so boring aswell when your not really playing the game you hacks are doing all the work for you.

Rant over sorry about that but it just annoys me the things some people will do.

[Phoenix64d]

why would you want benchmark results that are fake? Simpley so you can go look my numbers on a screen are higher than your numbers on a screen,

I hear that man; people are petty. But think about the economic motives for forging results. For example, we all know that most would gamble their lives on the absoluteness of 3dmark results. So what happens if SuperProcessor Corp., knowing the popularity of Benchmark Inc's program, donates equipment and money to Benchmark Inc? Well, benchmark inc might return the favor by developing software that favors SuperProcessor Corp's CPU's a little more. Result?Superprocessor stocks up. Sales up. Competition beaten down. After all, we all make judgements about which card/CPU is better based on benchmarks!

Now let me intercept the guy who's about to call me a conspiracy theorist. We all see these stupid tags placed on our power supplies and memory that say "SLI Ready." Hmmm...I wonder if they're getting money from Nvidia? What about all of the PSU's that run SLI just fine and aren't "SLI ready"? These labels don't mean jack; they're designed to manipulate people and promote one brandname over the other.

Again, I'm not suggesting that 3dmark is biased towards one manufacturer -- I have no evidence to support that one way or the other. I'm just saying that it's not all that inconceivable that ppl would want to publish slanted benchmarking software. That makes it more important that we use a variety of programs and not stake everything on one or even two.

[Zefram]

We all see these stupid tags placed on our power supplies and memory that say "SLI Ready." Hmmm...I wonder if they're getting money from Nvidia? What about all of the PSU's that run SLI just fine and aren't "SLI ready"? These labels don't mean jack; they're designed to manipulate people and promote one brandname over the other.

Well, the fact that it is not nVidia who paid the PSU manufacturer to put them into the "SLi Ready" sticker and groups, but its the PSU manufacturer that pays nVidia so that nVidia would test and verify their PSU as "SLi Ready"

Its all business strategy. Lots of consumer are driven by 'brand power', thus a PSU with "SLi Ready" would put confident and effectively effect the purchase decisions


Again, I'm not suggesting that 3dmark is biased towards one manufacturer -- I have no evidence to support that one way or the other. I'm just saying that it's not all that inconceivable that ppl would want to publish slanted benchmarking software. That makes it more important that we use a variety of programs and not stake everything on one or even two.
The exact reason why we had such a large online community, comparing scores and many of us tweak around with the benchmark programs. This is also the exact reason on why some hacks were found, like the infamous nVidia 3DMark03 cheats (http://breakingwindows.com/2003/06/nvidia_accused_of_cheating_in.php). Apparently some ppl found that when the 3DMark .exe were changed, the score changed too.
nVidia admitted to the cheats, calling them 'optimizations' and due to that controversy, graphics manufacturer tried hard and best to keep any hacks invisible and under the 'public radar'

Its like an old saying, "Hope for the best, prepare for the worse". Cheat in 3DMark, and if get caught just admit that you'd done some optimization. Tis' that simple :P