Whats Next...?

http://www.beyond3d.com/forum/viewtopic.php?topic=4248&forum=9


Recieving Shipments from Nvidia is referenced here for cards.

 

Dave,
I understand your position and agree with your choice. Still I find it piss poor of nV to do that.

 

Oh well, good luck with your preview.

 

Saying that you should not test a game that has different code paths is really illogical, just think a little bit and you will realize

The point of testing a game is to see how fast a game will run on a given piece of hardware, in the case of Quake3 is also gave insight onto how many other Quake3 based games would run. Running them on the same path would be stupid if the end user would not experience that.

If you want things that are identical paths and not necessarily reflective of real life performance get a sythetic benchmark.

We know Quake4 and Doom3 will be on the same engine, and I assure you other games will also, even though this engine will be less widespread than Quake3 was in its application.

 

I'm glad this appears to have been "settled" but I find your comment needing a final intrusion from me.

It applies to me, both as a reader of Dave's articles, as a member of Beyond3D. The latter makes this all the more pertinent because I write for the site as well.

Of course, you may have missed my final sentence in that post of mine, you know, the one that reads "My opinion, of course". And as such your assertion of "What is fine by you doesn't make it fine by everyone." gets a big "Duh!" from me .

 

Actually no it is not settled, still very disappointed in the lack of DirectX 9 benchmarking with stencil shadows i.e Battle of Proxycon. Yet 3Dmark 2001 is being used with only Pixel Shader 1.1 and 1.4 support.

I will be overlooking this article due to the politics involved, I do commend Dave for bringing out the clause into the open though.

Interesting to see how companies work behind the scenes.

 

Why the heck would you overlook an article over politics? That sounds more like bitterness than anything else. The rest of the tests are totally valid and relevant. It's not as if "politics" changes the scores. . .

 

No need to point out the obvious ostol

 

*sighs* Well, maybe not to the rest of us. . . (BTW, you obviously spelled my name wrong! )

 

oh come on are you seriously saying you see no reason to investigate exactly what the visible effect of those so-called "optimizations" which dramatically improve the score are? i mean i imagine you are under certain nda agreements but i about fell out of my chair when i read that.

-kyleb

 

What's the big deal? So we wait another month or two to get 3DM03 benches from B3D. The important thing is the FX will get the B3D run-down on tests I thought we all agreed were more relevant--actual games. And Dave has more than enough tests to ferret out other architectural quirks.

Methinks Mr.Guest is making much ado about very little. Given that Dave gets the card for free, I don't see anything wrong with nV attaching preconditions. Given that we read the review for free, we should appreciate what we get (and we get more than most here, which is why we visit). I expect to see a no-holds-barred review when B3D gets a card from another manufacturer (or buys one itself--it'd be nice if they made enough off the site to afford to buy one of each new card ), but that one benchmark doesn't make or break a review for me.

If Mr.Guest is implying somehow that B3D is selling out or compromising their integrity by accepting a review card with preconditions (much like an NDA), then he must not be aware of the fact that B3D is a big reason why everyone is fussin' an' feudin' over NV30's supposed architectural limitations. I'm glad nV gave us a card at all, given they know to expect an incisive and impartial review.

 

I just assume you would be given the card without strings attached, yes the normal recomendations of benchmarks that the card has been tested on and timeline with the card if required to return.

Asking a website to agree to not running certain benchmarks, especially a DirectX 9 benchmark that this card was supposed to be designed for is simply shocking. What is next, games that include Pixel Shader 1.4 ??

I don't like the backdoor excuse or Preview as we all know the reference card is the card being to the public, and Nvidia has won this round with Futuremark. They have discredited the benchmark and have so much pull on reviewers that they are agreeing to recomendations from Nvidia in the fear they may get overlooked.
Of course most review sites will also follow this trend, not wanting to miss out on review samples so in the end Nvidia has won their conflict with Futuremark.

 

That is reasonable, because it is not a shipping product. The final product may have better (or worse) performance.

 

Dave dont take this the wrong way as I love your work and your knowlede. But this is much different. ATI had not setteled on the clock speeds at the time. Well that what they told us. Only NDA people I guess really know if that was true. But it only took a sec to calculate about where the peformace was at. Here nV is telling you what not to do and that is just bad Mojo. Your know what you want to test and you should have compelte and 100% control. Weather you did not plan to use a test does not matter. Again the reviewer should have the right to review what ever he/she wants to......

 

I may as well weigh in with my $0.02...

Dave is obviously between a rock and a hard place. On one hand, the B3D faithful are all clammoring for "the skinny" on the NV30. On the other hand, we want "the skinny" with absolutely no strings attached.

I don't envy Dave's position.

I do find it upsetting for a hands-on preview to come with strings attached. I think that tells us more about nvidia's faith in their own product than anything else. However, I also think Dave has chosen the lesser or two evils...either risk not getting one at all and have a bunch of unanswered questions remaining on the NV30....or get one and be able to answer most of the questions (and possibly all) about it.

It seems like today, there's only one way that "the community" will have an honest assesment of the architectural characteristics of the NV30 (or any new product), and that's for B3D to put it through its paces. If we didn't have "code writers" around here and all kinds of other synthetic tests that will allow Dave to get to the bottom of the NV30, then I would feel differently about not using 3DMark03 for a preview. I also think Dave has made that point clear. (That the absense of 3DMark for this preview case has not hindered anything in terms of fleshing out the architectural characteristics.)

As we know, there is in fact a slew of different synthetic tests out there that should tell us pretty much everything that 3DMark03 tests would have, except for the "actual performance on 3DMark03," of course. This is more work for Dave, but he's obviously willing to do it. So IMO, the choice of not using 3DMark03 at this time, is really only putting DAVE out a bit, as he has to go to a variety of other sources and programs to gather similar data. So in all honesty, I think we should be THANKING Dave because, despite the 3DMark limitation, he is still going to get "to the bottom of it." Now, I don't want to kiss Dave's ass TOO much...actual praise will be reserved pending the actual preview.

Having said all that, there is absolutely no reason to have ANY strings attached to a retail product review. So I (and I belive Dave also) would not tolerate and IHV putting restrictions on a review. On a related note, I would also like to see, in any actual product reveiw, 3DMark overall score, and individual test scores, detailed.

 

good points Joe DeFuria, one must admit though that with actual retail products reportedly already shipping to best buy; this preview seems to be running rather late and a review would be much more welcome at this point. also i do have to wonder considering the fact that nvidia contracted all the manufacturing; what is the difference between this 'preview' card and the currently shipping retail product that we have yet to see a single review of anyway

-kyleb

 

Hmm. . .

Wants to use 3dMark03 == NVidia doesn't give Beyond3D a GeforceFX
Agrees not to use 3dMark03 == NVidia gaves Beyond3D a GeforceFX

I think Dave made the good choice. It's not as if he bought the card and NVidia asked him not to use the benchmark. . .