F2bnp wrote:

3DMark loves SSE so darn much...

Yeah the 3DMark synthetic CPU test does make 3DNow and SSE look pretty awesome. Unfortunately the calculations done there are probably just not that critical to game performance.

Superpi score and game performance are more closely related than 3dmark score and game performance* 😁

*not sciencetifically proven

Try Connectix Virtual Game Station. It supposedly runs full speed PSX games on the original iMac 233 Mhz. I had the PC version on a 633 Mhz Celeron PC, never noticed any slowdown, pretty sure it run fullspeed.

Finished testing the PIII 500. Tomorrow I will try the PII 266 and then it is K6 time. 😀
Tomorrow and the day after tomorrow however will be a bit slower as I have other things to attend to unfortunately. I will compile all the data on a nice Excel spreadsheet and also scan and upload my own notes.

I did not forget or give up on this. In fact, the Intel tests are pretty much done, I have tested the PII 400, PIII 500 and PII 266 running a V3 3000 PCI at 183MHz, effectively making it a 3500.
However, in the past I had experienced major performance drop when using my V3 PCI on my K6-3+. I don't know why, but only the AGP version runs properly (you would think it would be the other way around hahaha).

Unfotunately my V3 AGP will not clock higher than 170MHz without producing artifacts. I took out the heatsink, applied thermal paste, installed a fan and it just won't hit anything higher. I guess I'm just unlucky like that!
So, it turns out, I will have to redo all my tests using 166MHz. Gah... I'll probably skip a few games that didn't really change all that much anyway.

Better compare apples with apples, maybe the V3 PCIs performance is correct?

Nope. When I first got a K6-3+ system, the PCI was holding me back. I could see in reviews that for some reason it was underperforming. I think I even compared it to your K6-2+ 600 with V2 SLI and your cards were far faster.

Maybe you were seeing some sort of PCI performance problem on that Super 7 board.

Just a guess, no chipset drivers installed?

Or perhaps they are installed and thats the problem

The AGP drivers castrates Voodoo Banshee performance on my PC-Chips M577 MVP3

😀

GeorgeMan wrote:

FYI: Just set up a mini PC for my father to play 3D PSX games via ePSXe (e.g. Crash Bandicoot 3). Found a microATX QDI BrillianX […]
Show full quote

d1stortion wrote:

Another area of interest might be emulators. A PII 350 can already be enough to play ePSXe decently, while a PIII 500 might do fullspeed in many games. I would say SSE makes a big difference in such applications, and consequently K6 will be crushed mainly due to its weak floating point performance.

FYI:
Just set up a mini PC for my father to play 3D PSX games via ePSXe (e.g. Crash Bandicoot 3).
Found a microATX QDI BrillianX V (intel 440BX) mobo w/ Vibra sound chipset: http://www.motherboard.cz/mb/qdi/BrillianX%20V.htm
Combined it with Riva TNT2 M64 32MB AGP and 192MB RAM.

Came with PII 333 (FSB66). Not playable at all with DX6 plugin, major fps drops.
Switched to PIII 450 and downgraded to epsxe 1.7 from 1.9.
Playable full speed at "fast" settings, but really bad colors!

Then I lastly tried to overclock it to 500MHz (112x4.5) and downgraded to ePSXe 1.6.0.
It is playable at "Nice" settings, no lag at all! 😁
But I guess it's about ~95% CPU usage :p

But who cares, pretty happy now as I have 2 slot1 machines running on purpose @ 2013! 😁 😁

Here's what I wonder, why not just use a Mini ITX Atom board and run PCSX-R? It's a much more accurate emulator, and it would run better to boot.

Chipset drivers are installed. It must be just like swaaye says, something's going on with the PCI bus. You could take a look at this thread, it was kinda hectic and I'm pretty sure I tried everything (even another board of the same revision, GA-5AX rev 4.1).

Getting hold of a K6-2+/K6-3+

I'm finally done! I will slowly start making some charts using the numbers I got hehe

Finally done. I will update the first post to include this as well.
I sat down and wrote this on Word, as I couldn't get the graphs to show properly. If anyone is willing to do it, you've got the numbers, shouldn't be too hard. Then they can send it to me and I'll include the graphs and their username 😀.

Hope you enjoy, this took quite sometime to finish. I had some other games done as well, but a lot of the framerates had to be calculated without timedemos and that's not particularly accurate. I also had a lot of notes on them, so I just didn't bother to include them , I believe these are enough. I might upload my original notes for anyone who is interested 😀.

Nice!

It seems the K6-3@550 mhz performs like a P2/P3 between 350 and 450 mhz in most tests 😀

When run on the K6-III+, Magic Carpet recognizes the CPU as a Pentium. This may turn on certain optimizations, whic […]
Show full quote

When run on the K6-III+, Magic Carpet recognizes the CPU as a Pentium. This may turn on
certain optimizations, which I could not find unfortunately. Pentium II and III CPUs where not
recognized as Pentiums from the game. Still, I’m not entirely sure if there were any Pentium
optimizations, despite the message.

First off, very nice job!!

After working with Magic Carpet with the 3D Blaster version I know that the Pentium detection only puts certain settings to ON by default. Nothing more is done. Also I'm actually impressed with the 3D Blaster VLB version on a Pentium Overdrive 100mhz when seeing these results on the Pentium II.

Thanks 😀.
Apparently, that's exactly what it does so no worries. I have a Magic Carpet poster/ad on the wall next to my bed that advertises the Pentium. 😜 Also, I updated the downloads, they now contain a .pdf file, for easier viewing.
Vetz, where is the new 3D Accelerator Comparison episode!? I love these videos man 😁.

Btw, did you use FASTVID when testing Magic Carpet and the other DOS games on the Pentium II?

New episode.. Its in the works. Just alot of stuff that has been going on.

Oh crap. I knew I must have forgotten something. However, I did not use anything similar on the K6 either. It shouldn't be too hard to rerun those two tests. I'll see if I have the time to do them sometime next month!

F2bnp wrote:

Unfotunately my V3 AGP will not clock higher than 170MHz without producing artifacts. I took out the heatsink, applied thermal paste, installed a fan and it just won't hit anything higher. I guess I'm just unlucky like that!
So, it turns out, I will have to redo all my tests using 166MHz. Gah... I'll probably skip a few games that didn't really change all that much anyway.

I overclocked my voodoo 3 3000 agp to 188 mhz,but there is absolutly no increase in 3dmark 2001 on the k6-2+.