I am looking at the picture now and that chip is the one that I gathered the information on, labeled as a "Clock Gen" chip in the manual and located directly below the DIMM sockets in the picture.
I noticed that the board in the picture only has 512KB of L2 cache while mine has a full 1MB, which makes me feel like I have bragging rights. 😏

I dug this unit out to start testing the floppy machine ideas in it, but first decided to test the overclocking in it (as I have had little experience with super 7 overclocking). I received some interesting numbers that I think you guys might understand better than me. I ran two benchmarks on it with CPUZ and Quake III timedemo.
@ 570 MHz
RAM (3) latency (Fast)
52.6 CPUZ
26.8 FPS Quake III

@570 MHz
RAM (2) latency (Fast)
45.2 CPUZ
27 FPS Quake III

@570 MHZ
RAM (2) latency (Turbo)
52.4 CPUZ
27.1 FPS Quake III

@600 MHZ
(2) RAM (Turbo)
48.2 CPUZ
27.4 FPS Quake III

The only things that I could think of at the time was that I had changed the AWE 64 out for a CT4870 Live! and there was still some software conflict with it. But I really doubt it would have impacted the results across the board with that. It's kinda funny how the CPU benchmark registered lower while the quake benchmark registered higher every time... I even ran the tests several times incase it was just a fluke. Any thoughts are appreciated. 😀

EDIT: @616 MHZ (2) RAM (Turbo)
52.8 CPUZ and 28.1 Quake (btw the quake settings are maxed.)

If it is not the additional PCI card reducing performance, it may be from the PLL frequency multiplier circuit. In otherwords, for some FSB, your PCI frequency may be higher at lower clock rates, depending on what auto multiplier the PLL is using. You can check the spec sheet.

However, in such a case, if the PCI bus was dropping, I would expect the Quake score to drop as well. What CPU did you usein these tests?

feipoa wrote:

If it is not the additional PCI card reducing performance, it may be from the PLL frequency multiplier circuit. In otherwords, for some FSB, your PCI frequency may be higher at lower clock rates, depending on what auto multiplier the PLL is using. You can check the spec sheet.

However, in such a case, if the PCI bus was dropping, I would expect the Quake score to drop as well. What CPU did you usein these tests?

This might explain the symptoms mentioned at k6plus forums when people found their system was slower with a 6x multi then with a 5.5x multi?

I don't think this effect comes into play when changing just the multiplier, but only the FSB jumper. Perhaps when these guys changed their multiplier, they also changed their FSB jumper?

No idea, but your post about the PCI bus and the PLL chip is something I never even knew.

Thank you for the input. It's starting to make a little sense to me now. If the CPU is suffering from the 6x multi, and the frontside bus is still benefitting from it, then that would might explain why the cpu only benchmark is showing up higher than the cpu/3d benchmark. Or am I just so wrong that it's laughable?
EDIT: The CPU used was the on in the OP, AMD K6-2E+ 570.
Also I added fans and so far I'm only pushing the voltage to 2.1 (stock is 2.0). And I can't yet get the 124MHZ setting to boot past the Video bios screen.

nemesis wrote:

EDIT: The CPU used was the on in the OP, AMD K6-2E+ 570.
Also I added fans and so far I'm only pushing the voltage to 2.1 (stock is 2.0). And I can't yet get the 124MHZ setting to boot past the Video bios screen.

How's your PCI bus doing at 124Mhz? It's unfortunate only SiS-530 supports 1/4 PCI divider (but usually it's L2 cache can't take 133Mhz).

Not very well so far. I still can't get past the Video BIOS without it going blank on me. Now I'm running into issues with the explorer in 98se thinking that everything that happens is "illiegal" and shutting all my windows down. I guess I let the conflict go too far and now my registry is messed up... something I don't usually bother dealing with due to bad experiences. Guess I'll reinstall Win98se again 🙁 . As a side note; I installed a ZIP 750 IDE drive and as soon as I reinstall windows, I'll be able to install the drivers for it hopefully 😖 .

I finally got around to reinstalling the OS and I'm getting the drivers into place. Also, I'm planning on getting either an LS-120 or LS-240 drive into this machine to finish it as my "Floppy Project Machine A:".

Uploaded with ImageShack.us
If you look really really really close, you can tell that the 5.25" floppy drive is a few shades lighter than the rest of the drives... almost a gray color. The ZIP 750 is also slightly gray.
(Sorry yet again for the poor quality, I'm running out of excuses.)

I really like the dark colored ZIP drive!
I only got the white ones, and only some 100MB and 250MB ones.

I think it looks pretty neat! 😉

Is that a Mitsubishi 5.25" floppy disk drive?

Tetrium wrote:

I really like the dark colored ZIP drive!
I only got the white ones, and only some 100MB and 250MB ones.

I think it looks pretty neat! 😉

Thanks 😀 I originally had a white drive in it but it died (probably just needs some TLC), so I replaced it with this beauty for now. It was described as "black" but even with the poor quality of the photo you can tell that it's a very dark blue.

Is that a Mitsubishi 5.25" floppy disk drive?

Why, yes it is. 😀 Good eyes to pick that up through all that fuzz. I've had almost as good luck with Mitsubishi as I have had with TEAC. By the way, the model number is: MF504C-318MG, in case that information interests you.

I've got an identical drive, with the only difference between mine and yours being the light. On mine it's red. That drive is the only part I've managed to keep from my first PC, a 286 my parents bought me back in 1991. 😀

Right about now I'm facepalming myself. I posted that the results were CPUZ in the benchmarks when they infact were CPU Mark 99.
I originally had planned to use CPUZ, but then changed my mind at the last minute and forgot to change the text. 😦
Anyway, I tested the system with different RAM amounts and timings... this led to a slew of discoveries. Sadly most of which were discoveries of my own ignorance and carelessness. For example, I had originally set the RAM timings correctly with the OP ram at CL3, that would explain why I had problems with it set to CL2 I guess. I also discovered that I had set the CPU to 522.5MHz (for underclocking tests that I never posted because I'm too ashamed to admit that I'm not always obsessed with over-the-top systems) and never put it back up at 570. Ok, I'm getting carried away here so I'll get back to what I started to post about in the first place.
I tested the system with 512MB of crucial CL2 ram (timings set correctly this time too), and disabled the onboard cache just to demonstrate the difference between efficient memory and excessive memory. CPU Mark 99 posted a whopping 31.5 running only on the CPU L2 cache, and a 28.9 while using the onboard cache (this wasn't quite what I had expected but not far off). As a side note, I also disabled the swap file.
P.S. I'll share more thoughts and results after I get some rest.

@nemesis:
Would you be so kind as to format your text so it looks a little less like a monowall?
Because a big chunk of text has "DONT READZ ME!!!1" written all over it 😉

I did read it though (it's early morning and the chunk isn't too big to swallow 😜 ), underclocking is nothing to be ashamed about. And everyone makes mistakes all the time, but here on Vogons it's just a matter of writing "Oops!, I did blah blah" and all is good 😉

If you want to read about some REAL mistakes, we have a thread somewhere burried away here in Vogons about how we accidentally killed hardware by doing stupid things 😜
Hint #1: CRT monitors and white whine do NOT mix well!
Hint #2: If you want your room to smell funny, try running an Athlon XP Palomino at full speed without any thermal paste 😜

I like to share my mistakes for the main purpose of a laugh and also for the lesser important purpose of (hopefully) preventing anyone else from making the same mistake.

Tetrium wrote:

nemesis wrote:

EDIT: The CPU used was the on in the OP, AMD K6-2E+ 570.
Also I added fans and so far I'm only pushing the voltage to 2.1 (stock is 2.0). And I can't yet get the 124MHZ setting to boot past the Video bios screen.

How's your PCI bus doing at 124Mhz? It's unfortunate only SiS-530 supports 1/4 PCI divider (but usually it's L2 cache can't take 133Mhz).

Ever since you posted this, I've been tempted to launce it @ 124MHZ... I finally got it but forgot to do any benchmarking.

2x 128MB CL3 RAM
K6-2E+ @ 620 (124 x 5) and voltage set to 2.0

Got into windows once but it crashed the 2nd try... maybe more voltage would fix this but I don't want to push my luck too much today. At least now I can say for sure that the machine boots @ 124MHZ bus.

Do not push it that much. It is not going to get significantly faster but you risk frying your cards. Especially if you still use an ISA card - but PCI cards may get damaged as well at higher PCI clocks. And your system is not going to be stable on the long run. Stay with 5.5x100 or 6x100 Mhz; those are stable settings.

Thanks for the advice, looks like you have a similar system but with better chipset (probably anyway... I've never messed with that one) and sound hardware. In light of that I would like to add that I gave up on the Sound Blaster Live! card and switched to a Yamaha XG PCI (took out all ISA cards too).