Democoder wrote:
First and foremost, a benchmark should show the potential of what a piece of hardware is capable of if used correctly. Shoddily coded games that don't scale are no reason for us to give up on buying higher performance video cards. They are a reason to stop buying games from that manufacturer. Consider for example, a game that runs at the same framerate on a V3+Celery as it does on a P3+GF3. Does this mean that high performance video cards are irrelevent? Should manufacturers just NOT BOTHER making new cards because crappy games fail to scale? Should we just keep our 400Mhz V3/TNT's and go back to 1998?
A benchmark should show the potential a piece of
software performs on a variety of
systems. If you want to be/are a "hardware guy", that's fine. Games that are coded extremely well can show no scaling when it comes to hardware simply for various reasons. If gamers want better AI or sound or physics or etc. it will cost but not in terms of
video hardware. I don't think I have seen a game that fails to perform better with a P3 compared to a Celeron. Really
good games don't rely too much on 3D horsepower. No, 3D hardware manufacturers should not stop pushing
their envelope just because games fail to scale - they should, but if you think pixel/vertex shaders is just "performance", that's simplistic to the point of being absurd. 3D hardware manufacturers have to strive for stuff like shaders, and then if their first attempts fail to impress in terms of performance due explicit to these shaders being used, they should try to improve on their performance in this aspect. Hope my point is clear... lots more to say but I really should get sleep.
Secondly, Q3 is a very popular engine and is used in a large number of games, so performance in Q3 dictates potential performance for all Q3 licensees (but not guaranteed of course, since you can always screw it up)
And Medal Of Honor could be an indication of how useful even such an esteemed engine that has scaled so well can come to nought. MOH is CPU limited almost 80% of the time in single player. It is a great game and this is primarily down to really good stuff that has nothing to do with a 3D hardware's capabilitilies (AI, sound).
As for your mention of Sweeney and Unreal - Unreal Tournament is played more than Q3. UT sucks for benchmarking 3D cards. If this is the way Sweeney continues, he will make more money. Which means more people are buying his games.
Smart developers will make a game that sells well (duh). It doesn't matter to him if his game(s) doesn't show how evidently more powerful 3D cards will or will not run his game better than a lousy 3D card. He'll probably be more concerned about image-quality features than raw performance. If this means nothing to 3D hardware manufacturers, too bad.
We are at stage in the 3D hardware scene where we cannot talk about "old" games if I am to talk at the same wavelength as you, Democoder. Pixel and vertex shading games will perform better on supporting hardware that are faster
and benchmarks will prove this to be so. Q3 scales so well because it is fillrate limited. Newer cards has higher fillrates or better memory opts that improves on fillrate, which helps to show how good an engine Q3 is. You want to be forward thinking about 3D hardware or 3D games, then talk about games that already knows how fast 3D hardware is when it comes to fillrate/bandwidth and wants faster performance while using shaders. That's not what Q3 is.
UT, MOH, various simulation games, etc are/has been the sign of things to come - they are CPU limited for a reason... and that reason has probably proven to be the thing that sells the game. Better CPU related stuff, not better 3D stuff.
Tell me which is a better and more profound question :
Where would the 3D hardware industry be if not for games?; or
Where would the games industry be if not for 3D hardware?
If I'm not mistaken, you're responding to a poster basically asking question #1. And you're wrong in your response when it comes to the overall picture but you're not wrong if you're talking strictly about what you're really interested in, that of 3D.[/b]
