It seems the questions are now 2 fold.

1) Is a Matrox G200 maxed out on a K6-III+ at 500 MHz+ with Quake II in OpenGL mode? Luckily, this question does not have any influence over the nature of this benchmark comparison. I am doing a CPU, not a GPU comparison. The results from Quake II in software mode are unaffected by GPU saturation. The Quake II OpenGL results are included only for curiosity.

2) How many more frames per second do you get in Quake II software mode with and without the 3DNow! patch? This is certainly a valid question for a CPU-based comparison. As much as I am a stickler for good quality data, I will not be able to re-run Quake II on the affected 3DNow! CPUs. If anyone wants to volunteer their time with this testing, that would be great. If possible, we want to be consistent with hardware/software, but use whatever you have abailable. This is just to get a rough idea of frame rate boost with 3DNow! enabled.

• PCI Matrox G200 w/16MB (if you have it)
• PCI Yamaha DX-XG Sound (if you have it)
• 64 MB PC100 SDRAM (1 stick only)
• Windows 98SE
• Matrox G200 driver version 6.28 (if you have a G200)
• Unofficial Service Pack 2.1a
• DirectX 6.1a
• Internet Explorer 6 installed (probably unimportant)
• Quake II v 3.20 at 640x480, no 8-bit, software mode. Default settings.

F2bnp has run one such example thus far, which indicated a 22% enhancement on a K6-III+ 550. A few more CPUs and frequencies would be good to form an idea for average boost.

• AMD K6-2 300
• AMD K6-2 500
• AMD K6-2/3+ 300
• AMD K6-2/3+ 500
• AMD Athlon 500/600
• IDT Winchip2 233

F2bnp wrote:

I think a GF2/4 MX should suffice. Will give it a try tomorrow 😁

Doesn't T&L already take some load off the CPU in Quake 2? That might skew the results. Is it possible to disable T&L in the driver? Or would something like Kyro2 be fast enough?

I can do K6-2 300, K6-2 475, K6-3+ 300 and K6-3+ 500 with my current setup. Unfortunately I don't have a G200 (or a DX-XG for that matter but it doesn't really matter), but I'll be sure to run at the same settings I ran before (perhaps I'll leave out Software 1024x768 since it will be way too slow).

F2bnp wrote:

I can do K6-2 300, K6-2 475, K6-3+ 300 and K6-3+ 500 with my current setup. Unfortunately I don't have a G200 (or a DX-XG for that matter but it doesn't really matter), but I'll be sure to run at the same settings I ran before (perhaps I'll leave out Software 1024x768 since it will be way too slow).

Ok, great! Much appreciated. That should have read "DS-XG", by the way.

The Ultimate 686 Benchmark Comparison has been published in a new thread,
Re: The Ultimate 686 Benchmark Comparison

To eliminate thread cross-talk, further discussion on this topic should be addressed in the new thread. Thanks!