STALKER being benchmarked at X-bit

[Hanners]

they used the alpha leak(like they did with the hl2 beta) which doesn't have any ps2.0 shaders. There is no advanced feature in there hence look at the numbers. No dynamic lighting and no shadows(well blob) , no hdr , no bumb-mapping, no physics.

How do you know that the version of Stalker X-bit is using, is missing all those features?......if this is the case, they might as well be benchmarking the original UT

For what it's worth, I'd back tEd up on that - From what I have seen of this 'leak' there certainly doesn't seem to be any shader usage to speak of, as well of being bereft of bump-mapping, physics and so on. I really don't think it is a good indicator of performance in the final release at all, in fact it's probably very misleading.

 

[K.I.L.E.R]

Why in the living hell are people benching the alpha in the first place?

 

[Hanners]

Why in the living hell are people benching the alpha in the first place?

Because benchmarking with stolen property gets lots of hits?

They did it before with Half-Life 2, and no doubt they'll be using the Painkiller and World of Warcraft leaks next.

 

[micron]

Heh....I was reading your guy's thread over at EB about the Stalker leak, I understand about the lack of features in the benchmark they made....

 

[MistaPi]

Looking at the non AA/AF benchmarks though the geometry limitation also stuck out to me - the fact that the 9600XT was consistently behind the 5700 but the 5700 stuck with the 5900 highlighted this quite well. Interesting.

Does not that explain why nVidia is doing so well in this game, because its geometry limited rather than pixel shader limited?
Why is nVidia losing so much performance with AA and AF in comparison with ATi?

 

[Doomtrooper]

http://www.3dcenter.de/artikel/ati_nvidia_aa_performance/index_e.php

 

[rubank]

Looking at the non AA/AF benchmarks though the geometry limitation also stuck out to me - the fact that the 9600XT was consistently behind the 5700 but the 5700 stuck with the 5900 highlighted this quite well. Interesting.

Does not that explain why nVidia is doing so well in this game, because its geometry limited rather than pixel shader limited?
Why is nVidia losing so much performance with AA and AF in comparison with ATi?

Yes, the FX chips still include the "old time" HW T&L don´t they?

Graphics:

* Support for all second generation D3D compatible accelerators (TNT/Voodo2/etc), optimized for Geforce2 and up
* Visualization optimized for hardware TnL (both FF and shading)

 

[volt]

X-bit Lab's benchmark scores are not from DX9 version of STALKER.

 

[Neeyik]

Hmm, I seem to have been quoted a few times in this thread and misunderstood at the same time.

Neeyik was saying over at Futuremark that the devs "may have" chosen to avoid using PS2.0 routines, like Remedy did with Max Payne 2, to ensure the least amount of hassle and the best overall performance across the hardware base. Is this really so? I thought this was to be a game that utilized ps2.0?

What I meant by that was Remedy had chosen not to include any DX8 shaders in the first Max Payne for reasons of testing and development time, and to ensure that you get a good overall performance level across the existing hardware base. The same thing was obvious in MP2 even before Markus said anything in an interview. I had merely thought it might be the case with Stalker since the website mentioned nothing (that I could see) about DX9 shaders, but plenty about shaders in general. Let's face it - current DX9 cards are pretty good at DX8 shaders (with exceptions of course) but they're not working to the same level.

...according to Markus in our interview is was due to development time and testing in relation to the tired and tested engine they already had.

I don't think that Markus will like your description of their engine Dave! Sorry...I won't do that again....