3DMark Vantage Soon

>.< You are correct.

 

This is what I don't understand. The whole point of faster hardware and advanced rendering techniques is to make stuff look better. That's ultimately the end goal. All the stuff that happens in the background means nothing at all. How could they produce something so visually unimpressive and with such low performance? If GT1 is supposed to showcase cloth and water rendering they should have created feature tests for those. There is some nice reflection and stuff going on too but it all gets lost in the overall mediocrity of the assets and presentation.

It's easier to see where they were going with GT2 with the heavy instancing. But I ran the bench at performance and extreme settings and saw no discernable improvement in shader quality. What should we be looking for there?

In the end nobody is going to create a game using techniques that result in the output looking like ass. Would somebody from FM please explain what they're trying to demonstrate with Vantage? They would have been much better off producing a slide show that surpassed anything done in real-time today. At least then we would get something to look forward to....who cares if we can run GT1 at 60fps some day....it will still be just as ugly.

 

I was going to purchase the Advanced Edition. Now I'm not even going to download the free version.
Someone uploaded a video of the benchmark in action. Saved me money and time not downloading it.

I think Futuremark shot themselves in the foot with this version, hehe.

 

If Nvidia paying everybody good money because they don't have DX10.1 support, that is just wrong - I hope AMD would do something about it

 

http://www.theinquirer.net/gb/inquirer/news/2008/04/29/ubisoft-nixes-first-dx10

 

If it truly "was" a bug, wouldn't the rational response be to release a patch that fixes the bug, rather than removing the feature entirely? Or would the fix of the "bug" cause the DX10.1 rendering path to run even FASTER? I'm sure that would just make Nvidia's nightmare even worse.

It would be a shame if this were true that Nvidia is putting financial pressure or a threat to reduce devrel with UBIsoft in order to get them to remove a feature they don't support.

Just when the game was getting interesting. Throw money/weight around to make sure it goes away.

Either way, I can't see how UBI and/or Nvidia come out of this without some egg on their face. Then again, Nvidia marketing has done some wonderful Jedi Mind Tricks in the past... "These are the not the missing features you are looking for..."

Hopefully all of that is just paranoid conspiracy talk and nothing more.

Regards,
SB

 

unless your a IHV i cant see the point in buying a benchmark

 

I don't see anything wrong with your logic--good logic is never "paranoid," btw... No need to ever apologize for using logic.

Remember Eidos' Tomb Raider:AoD? Remember Eidos announcing that the AoD developer benchmark which originally shipped with the game, which demonstrated the performance and IQ differences in the game between the "DX8.1 path" and the "DX9 path" (respectively, between SM1.x and SM2.0), was unapologetically stripped from all future shipping versions of the game specifically because nVidia demanded it of Eidos? I do. Eidos unabashedly stated that it was removing the software in compliance with nVidia's wishes as Eidos did not want to be seen favoring one graphics IHV over another. Of course, nVidia didn't like its SM1.x product line at the time being shown up so clearly by ATi's SM2.0 products, especially since nVidia's PR for those products at the time indicated they had SM2.0 capability, too...

Really, the only difference here seems to lie in the fact that Eidos was upfront about why it was removing the software whereas UBIsoft clearly is being less than honest about what it is doing and why. As you say, the proper response to a "bug" is that you fix it. I mean, there seems to have been a strong reluctance on the part of Ubisoft to ever directly publicize the DX10.1 support in this game from the start, a glaring oversight that for me can have only one explanation...

 

What I want to know is why anyone hasn't questioned them about the details of this supposed graphical glitch it causes?

They could very easily justify themselves by simply posting a screenshot, or giving examples of instabilities that it causes.

All we have so far is vague references to some sort of "glitch" that the support brings with it. Stinks of foul play if you ask me.

Then again, why haven't we heard anything from AMD? Whats Dave's take on all this?

 

Whoops, sorry about that, I thought I was in the 3DMark Vantage thread

 

I bought the thing in its Advanced version. And kudos to FM for making that affordable. Having said that, I never had that "wow!" moment I've come to expect from a new 3DM over the years. I also thot that if I was Ronald Moore I'd want to have a little chat about their hommaging of Battle Star Galactica.

 

In addition to that do you remember all the 3DMark03 'optimizations' (read cheats) around the time of the NV30 launch? Absolutely insane. They tried (and arguably succeeded) in covering up the huge DX9 performance probs with NV30. NV is a very very unethical company when it comes to this sort of stuff.

 

I upgrade alot. So having a benchmark, (that used to look nice while running...) is something I like to have handy. No it's not needed for work. It's why I considered $19.95 vs $495.

 

yes but wouldnt buying a game be money better spent ?

 

Not really if what you are after as a test on your system changes.

Most games benchmarks are limited in what they can compare, or the engine is optimised in one direction only.

At least with FM's benches, each CPU/GPU test and the feature tests are stressing different parts of your system set-up.

 

Exactly. I know no other tool than 3DMark that can give us raw figures on how well a certain GPU is at at isolated tasks. Its fillrate tests are pretty much a 'worst case' for texture fetching, while the POM test is 'worst case' for unified shaders.

Such isolated tests give us a much better indication of what a certain GPU is good or bad at than just running some random games and getting average framerates. Sure, for gamers only the framerate matters... but developers would like to be able to see some synthetic tests so they can get a good feel of how far they can go with things like polycount and texturing, without having to bother writing their own tools to measure such things.
I don't see why 3DMark doesn't have a status much like SpecINT/SpecFP and such... essentially they do the same... one just does it for the GPU, the other for the CPU.

 

Yep, that's also why I bought it. It's a very useful canned benchmark for evaluating upgrades, overclocking etc. And as far as money well spent....$20 doesn't buy much where I live so blowing it on 3dmark isn't an issue.

 

OT:

The fact that this comes from an NV ("A few days ago, we received this statement from Andy Swanson of Ubisoft via NVIDIA" - in [H] write up on this) and not via Ubisoft makes me really wonder.

But it does not suprise me in the least that DX10.1 is not in his version of 3dmark....

 

But... did nobody notice that the boat/plane that Jane Nash escapes in, is in fact branded 'Sapphire', sporting the exact same font and logo as the wellknown ATi-only videocard manufacturer?

The plot thickens...

 

Sapphire is a boat brand, dunno if its at all related to the video card manufacturer.