OneENG-
I hope so
oldfart-
What res do you run at, and what monitor do you have? The GF based boards seem to have a problem with Trinitron tubed monitors concerning 2D image quality(which includes many makes besides Sony). Up to 1024x768 even the "bad" boards seem to be nearly identical even when paired with a Trinitron tubed monitor, for my Herc DDR(GF1) I would say up to 1280x960 on my monitor I have a problem telling the difference between a Matrox board.
If you are running a Trinitron, or running at very high resolutions then the V3
may have an edge depending on the particular board. I've seen several people comment on excellent 2D on the CL boards while others don't care for it at all, though that could be attributed to what resolution and monitor they have.
Soccerman-
<<
is it really that hard to make an engine (with say the graphics of Quake 3) run faster on slower systems? I mean, the Voodoo 3 is still pretty powerful when you think about it, (except it lacks some features), there's still lots you could do on a card of equal power as this. I don't see game programers trying to streamline games as much as I'd like. >>
It's easier then many would think, but there is a rather serious tradeoff. Only ~25%-35% of CPU time is devoted to graphics, you can change that ratio and increase visual quality quite a bit and give up the quality of game code. Gameplay vs graphics.
<<
does Quake 3 support 3DNow!? does UT? what about SSE? I mean, the developers don't seem to give a crap about getting more speed out of the game until they release patches, or if they get handed options to do so on a silver platter (ala 3DNow!, or SSE). >>
DirectX and OpenGL support both 3DNow! and SSE, and particular games do add their own support(Quake3 supports both, UT at least supports 3Dnow!, and the Athlon specific SIMD instructions). The speed issue tends to be problems with specific platforms outside of the development platform. For instance DeusEX on nVidia based boards. I think we will continue to see this as publishers are far too worried about shipping on time instead of shipping a completed title.
<<
if we were alot more restricted in terms of hardware speed, we'd actually see some creative programing to get as much speed as possible out of each mhz and each mpixel/second you have.. >>
This is possible, but you would need to focus on one graphics card, one processor. You could make some incredible graphics on say a GeForce, just look at the tech demos, but when you start to worry about that same code running on different boards you need to lessen the specific optimizations. An example of this can be seen by looking to the original Unreal. The graphics at the time were incredible, but it would only run properly on a 3dfx board. I forget what the high end rigs were running at the time, but until the patch released a couple of months ago my Athlon 550 GF DDR combo had trouble running it smoothly. Those type of breakthrough graphics can be done again, but they will more then likely force you to run one particular board to enjoy them.