STALKER being benchmarked at X-bit
- Thread starter micron
- Start date
I have been....I've also been cruising their forums for a bit. I even found Hellbinder there spreading the ATi word to the non-believersrubank said:
http://www.gsc-game.com/index.php?t=community&s=forums&s_game_type=xr
this the Xbit article in the first post of the thread? does it ever specify if it's a 128MB or a 256MB 9800 Pro? the extra RAM would probably account for it...geo said:
The Baron said:
Yeah, the Xbit article at the top. If it says whether it is 128mb or 256mb on the 9800 Pro, then I couldn't find it. Still, Wavey's XT review would lead one to look for something in the 7.5-9.5% range on performance difference, and quite a few of these are signifcantly higher even at lower res and no af/fsaa, not just the one eye-popper I cited.
Concerning STALKER the graphics look superb, while gameplay is something that I can judge only in real time. Apart from that I'm pretty happy that even mainstream cards, don't seem to have any trouble playing the game in 1280*960 with AA/AF enabled.
What the scores concerns the differences with AA/AF enabled in 1600*1200 with different cards, suggest that the framebuffer size is playing a significant role (case example 5900U vs 5900) and the 5700 seems to do extremely well probably due it carrying the same multi-array VS as it's bigger brothers.
Finally IMHO a game doesn't have to be shader limited at any price to be considered great.
What the scores concerns the differences with AA/AF enabled in 1600*1200 with different cards, suggest that the framebuffer size is playing a significant role (case example 5900U vs 5900) and the 5700 seems to do extremely well probably due it carrying the same multi-array VS as it's bigger brothers.
Finally IMHO a game doesn't have to be shader limited at any price to be considered great.
The result on 5700 sounds strange to me. Ok, 5700U got the same Vertex Shading power as the 5900 Ultra but the resultat are like that :DaveBaumann said:
1024*768 -> 1280*1024 -> 1600*1200
9600P : 138.6 -> 106.7 -> 72.7
5700U : 203.7 -> 149.8 -> 115.8
5900U : 207.7 -> 151.3 -> 117.3
If it's a Vertex Shading limitation the performance ratio between each card will change in favor of 9600P/5900U when we use a better resolution, but it's not really the case. Any idea ?
Fits some ideas pretty easily:
Limited shader capability usage, which goes with the DX 8 utilization discussion for at least some distance.
Undiscovered or undisclosed shader scaling for card capability by the "benchmark" being presented as equivalent, which goes as well with a lack of regard for professionalism as does a policy of using "warez" alphas without regret in a way that profits your site.
Limited shader capability usage, which goes with the DX 8 utilization discussion for at least some distance.
Undiscovered or undisclosed shader scaling for card capability by the "benchmark" being presented as equivalent, which goes as well with a lack of regard for professionalism as does a policy of using "warez" alphas without regret in a way that profits your site.
StealthHawk
Regular
BRiT said:The Baron said:
I keep asking radargs to supply his data, but he hasn't so far.
Mercury Research numbers had NVIDIA with a big share of the DX9 marketshare. Thanks to the lowend FX5200.
http://www.xbitlabs.com/news/video/display/20031029073519.html
The 5900 and 5700U performance comparisons are interesting. You usually see the 5900 tanking in comparison to the 5950 when the 9800 Pro dips to the 9800 XT (though not nearly as much)--usually at high resolutions, which leads me to believe we're seeing a bigger 126-vs-256MB RAM differential than we're used to. However, at the same time the 5900 has the tendency to dip right down to or even UNDER the 5700U. If RAM is really what's getting leaned on, would the 5900's 256-bit pathway make a difference in comparison to the 5700U's 128? The RAM itself is clocked lower, but why wouldn't the bit differential come into play and keep its performance dip closer in ratio to the 9800 Pro-to-XT, and well above the 5700U? 
Very curious...
It WOULD be good for them to have specifically mentioned what cards were being used, though, as it ads a bit of "assumption" to an equation which could easily be eliminated.
Very curious...It WOULD be good for them to have specifically mentioned what cards were being used, though, as it ads a bit of "assumption" to an equation which could easily be eliminated.