The only problem with this idea is that I cannot see a documented option to set the AMD 5x86 P75 chip to a lower multi of 3x.

Try using the "AMD DX4 Enhanced" setting.

retro games 100 wrote:

3DBench = 0.00 ?
3DBench2 = 94.3

Yea that machine is too fast for 3dbench 1.0 🤣

94.3 is brilliant for a 486 😀

retro games 100 wrote:

The only problem with this idea is that I cannot see a documented option to set the AMD 5x86 P75 chip to a lower multi of 3x.

Am5x86 can not be set to lower multi than 3x
I think the 2x60 setting is only possible with a DX4-120 chip.

retro games 100 wrote:

Inside the mouse driven AMI BIOS set up area, there is an option that says "PCICLK-to-ISA SYSCLK Divisor", and its possible values are PCICLK/2, PCICLK/3 and PCICLK/4. What does this mean?

That sets that the ISA bus clock speed to a fraction of the PCI bus clock.
E.g.: if you run the PCI bus at the standard 33 MHz, PCICLK/2 sets the ISA bus to 16 MHz, PCICLK/3 sets it to 11 MHz, etc...

Thanks very much for the info people! Re: setting the mobo's CPU jumpers to "AMD DX4 Enhanced" - yep, that was absolutely perfect. It's interesting to note that the jumper configuration for this setting, and the 5x86 setting, only differ by one jumper. I guess that specific jumper is the 3x - 4x multi option.

Using ChkCPU, I now know that the 5x86 chip is running at 3x multi. The next step is to determine if there are any undocumented FSB jumper settings, and what speed they are. Question please people: what's the most appropriate CPU to use, when searching for FSB settings which may go up to as high as 66 FSB? Thanks a lot.

I set the mobo's bus speed setting to 50 FSB, using an Intel SX-33. It POSTed OK, even with a) the BIOS timings set to their most aggressive, and b) the mobo's CPU voltage jumpers were only set to 3.47V.

Tomorrow, I will set the BIOS timings to their most "fail safe", and set the mobo's CPU voltage jumpers to 5V, and begin experimenting with different undocumented FSB jumper settings. I'm hoping that this SX-33 chip will POST OK, even if I "unwittingly" set the FSB to a high value of either 60 or 66 FSB...

this is kind of a dumb question, but where's the heatsink on your CPU? 🤣

Yeah fan that or DIE, though it is the ADZ variant which is rarer and more heat tolerant

Hehe, yes I removed all cooling paraphernalia before taking that photo. I've got 4 fans cooling all areas of this test mobo, just to be on the safe side! It's interesting, but just as an experiment, if I just use fans and no heatsink, then press my fingers on the CPU itself, I don't feel too much heat. In contrast, on some SX chips without a heatsink or fan used, these have felt much hotter.

I'm not using this equipment for too long, perhaps a maximum of 8-10 minutes at a time, so I doubt if any harm can be caused.

Wow, those Speedsys results are impressive for a 486 board! Looks like you found a real gem there....congrats!!!!!!!!!!!! 😁

How much L2 cache does the board have, and does it run in write-back or write-through?

Thanks a lot, it seems like a very good board. The manual mentions that the PS/2 port is "optional", so perhaps there were two SYL8884PCI-EIO models manufactured? I think that if my board had PS/2, it would be the icing on the cake.

The board has 256K L2 cache. (Please see the Speedsys image above.) The board is capable of using up to 512K of L2. The cache mode can be set inside the AMI BIOS set up area to either WB or WT. For the Speedsys image above, I was using WB.

I did notice that when I removed the AMD P75 chip, and replaced it with an SX-33 @ 50 FSB, that the BIOS POST display said something like "now using accelerated cache", or words to that effect. I wonder whether the BIOS settings were automatically altered to WT cache, when it detected that the SX-33 was installed, and the mobo was running at a high FSB rate? I also noticed that the Speedsys memory bandwidth score had increased to about 137 (from 120).

On your images from mars I noticed that your ribbon cable has seen better days. Look after them, bent cables like that can one day cause issues which are hard to track down 🤣

Mau1wurf1977 wrote:

On your images from mars I noticed that your ribbon cable has seen better days. Look after them, bent cables like that can one day cause issues which are hard to track down 🤣

🤣 I like the images from Mars reference! That cable was brand new, only 2 weeks ago. I do lots of testing, and things get a bit "out of shape" quite rapidly!

You have to keep that camera as these images will gain cult status. Spirit couldn't shoot pictures as nice as these 🤣

Doesn't it have a PS/2 mouse header just in the side of the KB connector ? A straight row of 5 pins.

Immediately next to the keyboard socket, there is a row of 5 "solder points/spots" on the PCB where (presumably) the five PS/2 mouse header pins ought to be located. There is also a 6th solder point, immediately next to these 5 solder points, and it is marked CN8 on the PCB.

Quick update. I have located the 60 FSB jumper setting. It was the very first undocumented FSB jumper setting I tried! If you have this board, it's jumper J28. You need to close the middle pins only, and the leave both outer pins open. Like this:

oo
--
oo

For my quick test mentioned above, I used an SX-33. At first, the BIOS POST display appeared, but then froze. I then realised I had the BIOS set up settings "maxed out". I set them to "fail safe", and that allowed me to get to the C:\ prompt OK. I ran CPUChk.exe, and it told me that the chip was running at 60 MHz. The SX-33 has a multi of 1x, so the simple equation is 60 FSB * 1x CPU multi = 60 MHz CPU speed.

I will now replace the SX-33 with the P75 (with the mobo's AMD CPU jumpers set to 3x multi, and not 4x multi), and see what happens. At first, I will try a lower voltage of 3.47, then work my way up to 3.6, then 4, and then 5...

retro games 100 wrote:

Question please people: what's the most appropriate CPU to use, when searching for FSB settings which may go up to as high as 66 FSB? Thanks a lot.

I tackled this very same thing about a year ago when testing a couple of my boards for undocumented settings. Granted, I was looking for lower fsb's but knew I might hit higher ones by accident.
Before setting about to do testing, I set everything at it's lowest timings and used an ISA card. This way I had the best chance of a successful boot.

I think one of the best (if not- the best) cpu would be the Intel DX-50 but as those are quite uncommon (and I only have one) I decided to use my ugliest AMD DX-40 instead 😉
There may be better options but if I ended up frying a cpu, I wanted it to be one I could easiestly spare.
Another option could be a DX4-120 set with a 2x multi, if there are any. I didn't have such a cpu a year back so the AMD DX-40 it was!

Edit: I think I have the same mobo that you have, though not sure if it has the same features or not. Doing some researching I stumbled upon an unofficial site about this board. It did mention it has some flaws, most notibly:
-unreliable secondary IDE port
-PS/2 mouseport wouldn't work
-Intel DX4-100 wouldn't work

Link: http://th2chips.freeservers.com/syl8884/

Edit2:
the one I have is the middle left one in this pic:
http://i941.photobucket.com/albums/ad254/inte … um/DSC00201.jpg

You have exactly the same board as I have! 😀 I've seen that unofficial page before when I first got this board about 18 months ago, but I forgot to look it up again. Hehe, no wonder the PS/2 header pins have been "soldered out" on our mobos - it sounds like they wouldn't have worked if the pins were present.