Skalabala wrote on 2021-05-10, 20:54:

lukas12p wrote on 2021-05-10, 10:29:

My GA-5AX with K6-3+ 400@580 and Voodoo3 3000 AGP (no OC)

That is a good score 😀

It has almost the same score @550 but 110 fsb

lukas12p wrote on 2021-05-10, 21:00:

Skalabala wrote on 2021-05-10, 20:54:

lukas12p wrote on 2021-05-10, 10:29:

My GA-5AX with K6-3+ 400@580 and Voodoo3 3000 AGP (no OC)

That is a good score 😀

It has almost the same score @550 but 110 fsb

Your CPU wont do 5.5x110?

Skalabala wrote on 2021-05-11, 09:09:

lukas12p wrote on 2021-05-10, 21:00:

Skalabala wrote on 2021-05-10, 20:54:

That is a good score 😀

It has almost the same score @550 but 110 fsb

Your CPU wont do 5.5x110?

It is not stable above 580 MHz in Win98. I have tested two 400 MHz and one 450 MHz versions with 2,3V on MVP3 and ALI boards.
Similar K6-2+, I have two 550 MHz versions, but under Win98 they are not stable above 600 MHz with 2,4V. Maybe I/O voltage is the problem?

galanopu wrote on 2021-02-12, 00:55:

Is this good enough?

That was close but you should hit 5000 to be in the club ))

Mentat-vvo wrote on 2021-07-02, 13:26:

galanopu wrote on 2021-02-12, 00:55:

Is this good enough?

That was close but you should hit 5000 to be in the club ))

T-T

To my understanding I have to use some other older version of windows.

Hi members 😀

I am slightly back at this 😁
I got super lucky to get my hands on an ASUS P5A 1MB cache motherboard 😁 😁
However unfortunately it is on the unstable side, 🙁 I will venture into it to replace the capacitors and do whatever I can to get it stable.
For now this is what I have, I am impressed with the speed of the motherboard though 😀 😀

The board 😁
https://drive.google.com/file/d/1hW4G5gruYqD3 … iew?usp=sharing

This was done with K6-3+ @550MHz Stock GF3 Ti200 and 128mb ram, this is my personal best at 550mhz, and I have only spent one night to get here 😁
https://drive.google.com/file/d/1WlqXrcgE0kYo … iew?usp=sharing

Finally I have a good record score 😀
One sad thing is I can not make the system run at 120MHz, Cache disabled or not, K6-2 and 3.
https://drive.google.com/file/d/1ZMRiDfRtesNX … iew?usp=sharing

Skalabala wrote on 2021-07-15, 19:26:

Finally I have a good record score 😀
One sad thing is I can not make the system run at 120MHz, Cache disabled or not, K6-2 and 3.
https://drive.google.com/file/d/1ZMRiDfRtesNX … iew?usp=sharing

That is insane score!
What is Your setup and BIOS/reg settings?

I have used ASUS P5A-B v1.04 + Geforce 2 GTS + K6-3+400@617 MHz
Playing in BIOS settings doesn't make real difference

I have noticed that Voodoo3 is about 30% slower on that board than on Gigabyte and score on Asus is the same with and without AGP drivers.

Yes that is a good score 😀 I feel confident to say I will go over my goal of 6000 😁

So the board is P5A 1.06 with 1024Kb cache
Graphics card is a chaintech Ti200 BIOS moded
Ram I am using PC150(single rank) and sometimes Twinmos(duel rank) 6ns 128mb
ALi driver 1.82E and ALi utility set to turbo.
Only reg settings are for Fast NAJ and RAS timing to 4T.
And then CTU tweak just to double check the write ordering.

I am still experimenting as I have only tried this one windows disk and ALi 1.82 and DX7.0A

Yes there is some sort of bug with Voodoo 3 😒

So this happened last night 😁

Skalabala wrote on 2021-07-24, 11:59:

So this happened last night 😁

Congrats! I saw you were able to reach 5700 on YouTube. Thanks for hires photo of your mobo. Maybe those jumpers are enough to switch the bios and chipset to 1MB cache. I also have turbo-tag sram IC that runs at 124M. Don’t have much time to mess with this stuff though))

No problem 😀 Any questions are welcome!
I was actually able to run at 133MHz but I think the hard drive was not happy, got some error of registry bla bla windows will restart and fix it.

Skalabala wrote on 2021-08-07, 06:35:

No problem 😀 Any questions are welcome!
I was actually able to run at 133MHz but I think the hard drive was not happy, got some error of registry bla bla windows will restart and fix it.

You might be able to fix that issue with an external PCI controller (one that tolerates running at an out of spec PCI frequency). Then again, it could also be caused by the CPU/northbridge.

bloodem wrote on 2021-08-07, 06:53:

Skalabala wrote on 2021-08-07, 06:35:

No problem 😀 Any questions are welcome!
I was actually able to run at 133MHz but I think the hard drive was not happy, got some error of registry bla bla windows will restart and fix it.

You might be able to fix that issue with an external PCI controller (one that tolerates running at an out of spec PCI frequency). Then again, it could also be caused by the CPU/northbridge.

Yes this is the idea I had, you also saying it gives me hope. The end goal is 5.5X133 for 733MHz 😁

Skalabala wrote on 2021-08-07, 11:02:

Yes this is the idea I had, you also saying it gives me hope. The end goal is 5.5X133 for 733MHz 😁

This looks like a nice challenge. Have a bunch of SS7 boards pending benchmarking, including AOpen AX59PRO (possibly a 2MB version as well), Asus P5A 1.04 and P5A-B, Gigabyte ga-5ax and (soon, hopefully) Freeway 5vgf+/ULTRA to test with my K6-III+ 550. Will give it a try when time allows 😀

vmr_ wrote on 2021-08-23, 14:56:

This looks like a nice challenge. Have a bunch of SS7 boards pending benchmarking, including AOpen AX59PRO (possibly a 2MB version as well), Asus P5A 1.04 and P5A-B, Gigabyte ga-5ax and (soon, hopefully) Freeway 5vgf+/ULTRA to test with my K6-III+ 550. Will give it a try when time allows 😀

Give me your CPU.... 🤣

Recently got a FIC PA-2013 version 2.1 motherboard in perfect condition for shipping cost.
Gave it a quick spin. Noticed something really weird. Check these two screenshots:

All parameters constant, only CPU clock multiplier changes.
First time seeing a system behaving in such weird way. Is there a plausible explanation to it ?
Or am i missing something, considering my first time exposure to this k6/SS7 business ?
Thanks.

Instability can do that.

The rig is very stable at 112×6. Only at 124x6 the cpu needs +0.2v, but that's another matter.
Can other parameters change implicitly based on cpu speed ?

The on-die L2 cache may change timings I guess but I don't know of any way to test this of other than seeing what cachechk says.

It could still be something to do with instability. The on-die cache could be catching errors and correcting them. That will lead to less performance.

It is kinda the same thing when overclocking video cards. Sometimes the card will seem completely stable at a higher frequency but a slightly lower frequency will produce higher performance.