Reply 20 of 24, by Disruptor
- Rank
- Oldbie
CRG wrote on 2026-08-15, 20:40:I can do more testing changing the turbo switch later.
Please do that.
CRG wrote on 2026-08-15, 20:40:I can do more testing changing the turbo switch later.
Please do that.
Disruptor wrote on 2026-08-15, 20:41:CRG wrote on 2026-08-15, 20:40:I can do more testing changing the turbo switch later.
Please do that.
I'll get screenshots tomorrow but quickly ran speedsys on the original AMI BIOS at 133mhz. It's made no difference.
As I said earlier that ram speed is phenomenal 😀 and chipset definitely does support EDO. You are pretty much getting Pentium chipset levels of ram speed here.
penultimate picture is a double up of previous upload, you missed Award L2 EDO 40MHz, but no revelations there I imagine.
Now the weird bit, WTF is this chipset doing, is it breaking the laws of thermodynamics? in my house?!?! 😀 . Afaik 18 us/KB L2 is what you get with 133MHz CPU and L2 set to 2-1-1-1 yet you get that on EDO like the chipset datasheet promised (which I called out as a lie and have to eat my hat now!). HOW? Hmm looking at EDO datasheets the best they can do is https://www.parastream.com/downloads/datashee … 20Em614163A.pdf
Fast Access Time: 50ns/60ns
Fast EDO Page Cycle Time: 20ns/25ns
So that indeed is doable with 40MHz at 3-1-1-1 timings 😮 Well I learned something new indeed 😀 I kept remembering 5-2-2-2 as minimum becvause thats what all the Pentium Chipsets support, but Pentiums run at double the FSB of 486 duh, stupid rasz!
I dont remember, do you have digital scope? Can you put a probes on 1 L2 cache lowest address line and 2 CPU clock to see what is the real timing, and then repeat same measurement thing with EDO and probes on 1 CPU clock and 2 CAS. This will give you definitive answer, but the math checks out.
That weird cache speed with AMI is most likely connected to jumper setting. Something Award bios does or doesnt works ok with that jumper position, would have to trace that jumper to particular CPU pin to know for sure. Same with FDD not working, might be different ALI IO chip on Award bios and initialization doesnt work completely the same, easy to patch but ultimately not needed seeing how EDO is as fast as L2 on this board 😀
TLDR: you won with chipset capable to run EDO with 3-1-1-1 timings, no L2 will ever beat that. This is the fastest you can do.
https://github.com/raszpl/sigrok-disk FM/MFM/RLL decoder
https://github.com/raszpl/FIC-486-GAC-2-Cache-Module (AT&T Globalyst)
https://github.com/raszpl/386RC-16 ram board
https://github.com/raszpl/Zenith_ZBIOS Zenith Z-386 MFM-300 ZBIOS disassembly
Crazy RAM speed at 40MHz. I would say a polite thanks & goodbye to the cache if I had that board.
Mike and rasz are right.
Do some stability tests in the configuration @ 160 MHz with EDO and no L2.
Run all benchmarks from DOSBENCH.