That's what I have imagined. If this is true, then the MX 460 would be the top performing DX7 card. The only advantage with the […]
Show full quote
That's what I have imagined. If this is true, then the MX 460 would be the top performing DX7 card. The only advantage with the GeForce 2 Ultra would be if you needed a very specific driver that the MX 460 doesn't support. But I can't think of any example of this.
Another thing to consider is that the GeForce 2 Ultra supports the 23.11 driver, which is much faster than anything that came after it if you have a slower processor. Conversely, if you have a faster processor then later drivers, especially the 50 series, see a huge performance boost. My guess is that NVIDIA moved the hardware T&L to software around this time because modern processors with SSE and SSE2 could outpace the original hardware.
For comparison, on a 440BX with a P3 650-100 (Coppermine), my GeForce 2 Ultra scored the following on 3DMark 99/00/01:
99 - 7051
00 - 4939
01 - 2553
But on 56.64 it scored much less:
99 - 5229
00 - 4684
01 - 2729
When pairing the GeForce 2 Ultra with a fast P4 3Ghz it scored the following on 23.11:
99 - 7483
00 - 5439
01 - 4070
And on 56.64:
99 - 7531
00 - 10342
01 - 4855
The question I have is how the MX 460 performs by comparison. Even if it's faster, is the GeForce 2 Ultra still faster on a slower processor because it has the ability to use older drivers?
I also believe that this makes a lot of the modern "benchmarking" where old cards are paired with modern or much later processors and drivers to be an unrealistic comparison that can skew scores significantly. It also provides no frame of reference to how these cards perform on hardware from that era. A perfect example is when people try using the latest 81.98 drivers on Windows 9x and everything is broken. These are details that a lot of modern reviewers fail to take into consideration.
Some things worth considering. 😀