3DMark 11 out on 30th November

[FM_Jarnis]

hey jarnis, you guys should add a line to the contracts with big prtners that they can change the launch dates BUT have to announce it a certain time before the launch, to prevent last minute delays like this... cause announcing a launch and then delaying it last minute makes you look unprofessional imo... no offense, im just saying...
ati and nvidia etc need to know more than a month ahead whether the next 4 weeks are enough for them to tweak drivers or not, and if they first think it is and then its not, well too bad for them...

and yes, bringing back the invidiual test run like in 2k1 would be awesome!

Partners or BDP members did not change the launch date. Futuremark did. BDP members did help uncover some bugs and that left us with the choice - fix or ship unfixed and patch later. Simple as that.

In any case, the problems are sorted and we hopefully have news on the launch date tomorrow...

[FM_Jarnis]

actually im going to blame FM for that one. if the benchmark was out as of yesterday, then everyone in here defending it would be off using it, and someone would be arguing with an empty classroom.

Agreed. Sorry about that. Our bad. :p

[Leeghoofd]

+100 Massman

Can we focus now on the real subject of this thread plz...

And Jarnis : to make up for the failed delivery date you have to give us all free copies :p, better give on to Katz too...

[Monstru]

In any case, the problems are sorted and we hopefully have news on the launch date tomorrow...

Now we're talking, looking forward to play with the benchmark. As I allways say, tes first, discuss later

[Katzenschleuder]

No, it has been my point all along.

- For initial performance indications, 3DMark gives quite good accuracy

The previous 3DMark or a fillrate tester combined with a trivial fragment program algorithm that only takes ten seconds for measurements delivers the same level of accuracy on the test subjects.

- For detailed performance indications, you need to test real games.

Which is also what Futuremark said in an interview: http://hwbot.org/article/newsflash/9...mark_president.

The point of dispute is NOT(!) that testing 20 games is more useful. In fact, I would say the data based on multiple test scenarios provide more meaningful results. The point of dispute is that you are trying to prove 3DMark is inherently biased/incorrect/producing false results, whereas truth is that it's quite capable of identifying the performance of a certain product.

Could you please quote me where I have allegedly said that? Why are you continuously putting words in my mouth?!

I am saying and always said the following things:
That 3DMark delivers an incorrect performance estimation because of the natural inability to mimic real games.
The other side of the criticism is that Futuremark is taking money for letting others have an influence on the substance of the benchmark. I am not the only one who critisizes this business model as scandalous in regard of respectability and harmful for the interest of the PC gamer because of the wasted resources which instead would have been spent on driver development.

I think you've learned a valuable lesson about using big statements in 'public'. Perhaps in the future you'll be able to provide your opinion in a more balanced manner, which ultimately will make you lean closer to the true nature of things.

Of course master.

[Manicdan]

That 3DMark delivers an incorrect performance estimation because of the natural inability to mimic real games.

what would it take to mimic games?
and who says game developers are doing things right being console ports derived from dx9 crap with dx10 and 11 added on after?

[s1nykuL]

what would it take to mimic games?
and who says game developers are doing things right being console ports derived from dx9 crap with dx10 and 11 added on after?

+1 yup.

<satire>
FM should re-code 3D Mark 11. It should be a DX9 benchmark with some DX11 features tagged on as an after thought. Only then will it accurately benchmark the performance of the latest and future PC games.
</satire>

[massman]

Could you please quote me where I have allegedly said that?

Here you go.

You say that 3DMark isn't useless? The point of 3DMark Vantage has been to predict performance of Direct3D 10 games, right? Well look for yourself: 6 out of 12 misplaced performance predictions - a perfect random mean! But what else can you expect from a benchmark that creates scenes that are totally atypical for games?!

What the heck?! Most GPUs are still rated incorrectly: 5870/470; 6850/460; 5830/465; 4870/5770; 260/5770; 450/5750; 4870/260 ... They are all rated incorrectly.

I never said that it is realistic to think that one can create a benchmark that mimics the workload of upcoming games, early in the GPU generation.

The whole point here is that Futuremark promises to deliver this and fails for good reason.

WTF?! The majority of the GPUs tested have an incorrect performance rating succession when comparing 3DMark with real games! All those combinations that I have listed are rated incorrectly by 3DMark.

The average of games is not "slightly more accurate" - it is (or at least should be) the reason and goal for the performance evaluation.

...

So 3DMark is nothing but an obstacle for what GPUs are actualy made for.

You can pick comparisons with many games and you will see the same obvious discrepancy when comparing to 3DMark results

That there are significant differences between games is the NATURE OF REALITY and exactly my argument why 3DMark will never be a useful tool to evaluate performance! The only goal of 3DMark can be to get close to the avarage, which it obviously (and naturally) doesn't achieve.

That 3DMark delivers an incorrect performance estimation because of the natural inability to mimic real games.

Please, just leave this thread and this forum (well, perhaps all forums for that matter). You're quite incapable of analyzing data, providing a well-balanced opinion and remembering what your initial point of view was. This is quite typical for internet discussions; now you're trying to explain that your original point of view was not the one you defended for the last 5 pages and others have to prove you did.

What an outrageously meaningless illusion of significance.

[Manicdan]

+1 yup.

<satire>
FM should re-code 3D Mark 11. It should be a DX9 benchmark with some DX11 features tagged on as an after thought. Only then will it accurately benchmark the performance of the latest and future PC games.
</satire>

i think power consumption should be almost the best candidate for determining how well a game or benchmark works. if the unreal engine loads nvidia gpus to 90% of the furmark temps, and amd cards only to 60%, it would show that amd cards are heavily under utilized in it. but compared to other games ati cards might always be around 60% which just shows how the architecture works, but if another game shows 85% on the amd card, it could mean that engine is just not amd friendly. then a quick comparison on temp relative to framerates to see if the engine and architecture compatibility are showing expected results.

considering that 3dmark vantage is commonly used now as the main way to test load temps and consumption, its clear that people like how well it stresses cards.

[cowie]

I think I am.

What you are describing is a way to game the system with 3DMark 2001 as things like this were not taken into account during the development.

It ultimately boils down to the question - do you want benchmark to be;

- A program that spits out a number that depends on your hardware and on how you tweak and poke the settings to the maximum in order game the number to be bigger.

OR

- A program that spits out a number that gives an indication of the performance of your hardware. Period. Want a bigger number? Tweak the hardware.

We want both for a even bigger number.

[Lightman]

Finally some good news:
- we will know when new 3DMark launches tomorrow
- nice admin finally banned that annoying kid on crusade against FM

On several occasions I was going to reply to him but refrained myself thinking he's not worth it. Besides, for insulting quite a few of XS members referring to their IQ or comprehension capabilities in every second post is way below level we should maintain on this forum, therefore he should be banned a lot sooner .


Anyway I'm really looking forward to testing new 3DMark on my temporary HD5770 and then comparing results to Cayman once I get it.
Including proper DEMO in new benchmark is a great move! Because of that feature (and partly because I'm old school demo scene guy) I've ran all previous marks hundreds of times just for entertainment, but Vantage only once or twice after upgrading GPU. This made Vantage sort of bad investment for me, but I hope FM will redeem my loss soon

[TurboDiv]

So its coming out tomorrow ? I pre-ordered a while back and was looking for the download link in my e-mail November 30 but i could't find it the other day... lol

[FM_Jarnis]

So its coming out tomorrow ? I pre-ordered a while back and was looking for the download link in my e-mail November 30 but i could't find it the other day... lol

Sorry, no, but we hope to have news on the date later today.

[s1nykuL]

i think power consumption should be almost the best candidate for determining how well a game or benchmark works. if the unreal engine loads nvidia gpus to 90% of the furmark temps, and amd cards only to 60%, it would show that amd cards are heavily under utilized in it. but compared to other games ati cards might always be around 60% which just shows how the architecture works, but if another game shows 85% on the amd card, it could mean that engine is just not amd friendly. then a quick comparison on temp relative to framerates to see if the engine and architecture compatibility are showing expected results.

considering that 3dmark vantage is commonly used now as the main way to test load temps and consumption, its clear that people like how well it stresses cards.

I am not sure about using temps, a badly seated cooler will skew results. A more efficient cooling solution would also skew results, as would differing ambients. GPU load is something else this could be used to measure GPU efficiency providing the algorithm used to generate the load value is accurate. For me this is not an area of expertise.

[Manicdan]

I am not sure about using temps, a badly seated cooler will skew results. A more efficient cooling solution would also skew results, as would differing ambients. GPU load is something else this could be used to measure GPU efficiency providing the algorithm used to generate the load value is accurate. For me this is not an area of expertise.

well power consumption turns into temperature (indirectly based on what is cooled). and it wouldnt be such a simple flat number, but where this fall relative to their other examples using the same card. i know its not simple, but it would be the optimal way i think to figure out architecture efficiency in an engine.

[Kimofil]

3DMark 11 will be released Tuesday December 7th

[Oztopher]

Oh, how convenient.

[FM_Jarnis]

This announcement was just sent to everyone who preordered:
---

It took us a little longer than planned, but 3DMark 11 has now gone gold. Earlier today we announced that 3DMark 11 will be released at 14:00 GMT on Tuesday December 7, 2010.

From that time onwards you will be able to download 3DMark 11 from the Futuremark website and start submitting results to the new online service. You will need the upgrade code from your pre-order confirmation email to activate the Advanced Edition so make sure you have it to hand on Tuesday.

Till then, happy benchmarking!

[jmke]

the installer is smaller than 3DMark05! Excellent

[FM_Jarnis]

the installer is smaller than 3DMark05! Excellent

We were told that you guys hated big unwieldy installer blobs that take a long time to download - so that (too) was fixed

[Monstru]

Praise the lord

[Manicdan]

does that imply load times will also be quicker?

[cold2010]

Praise the lord

[Monstru]

does that imply load times will also be quicker?

[FM_Jarnis]

does that imply load times will also be quicker?

Let's see... less data to load... *ponder*