The battery is new. The previous owner replaced the old one before I bought this board.
I will not remove it again anytime soon.

So here is the result from the Ali chipset:

509.1s 1.9fps in quake bench.

Cachechk:

BIOS:

Even with Auto HDD detection.

Can anyone confirm that their ALi 386 board works with a busmastering SCSI controller? I had an ALi board and it didn't sit well with any of my AHA 1542 cards. I'm not sure if it was bad board design or just a "feature" of the chipset. Also, does anyone know if the ALi 386 chipset supports page-interleave mode?

Elianda, does the SiS board also have the internal cache option in the BIOS?

Nope:

If I enable Fast A20 some memory managers are unable to control the A20 line, so I left it disabled. Doesn't seem to be a problem, since A20 is usually just on.

And here another slight tune of the memory performance:

I used the hyperram tool to set refresh to 600 microseconds, which seems to work. (not long time tested yet)
But since old SIMMs have Parity, errors come up quite fast.
650 was not stable. But no reason to complain as 120 is usually already considered slow. The performance impact is probably more noticeable if you start from 60.

I'll have to try out this hyperram program on my working SiS board, which has slow memory performance (SpeedSyS shown previously). I beleive I found it online in a package called sk470s.zip

Something else which caught my eye in the BIOS of your SiS board, what function does "Mouse Support" have?

feipoa wrote:

I'll have to try out this hyperram program on my working SiS board, which has slow memory performance (SpeedSyS shown previously). I beleive I found it online in a package called sk470s.zip

Something else which caught my eye in the BIOS of your SiS board, what function does "Mouse Support" have?

According to the internet, it enables PS/2 Mouse Support on IRQ12, but since I have no PS/2 mouse in this system, this setting does not matter for me.
I am not sure, why this has to be enabled specifically.

Was this internet site referencing a 386 or a 486 computer? To the best of my knowledge, there is no way to add a PS/2 mouse port to a 386. I have not seen a PS/2 port on a stand-alone 386 motherboard, though there may be some proprietary Compaq, laptop, or similar designs which had one.

This information wasn't given. I don't have the manual for the mainboard and the option doesn't seem to influence serial mouse operation.

The official printed manual I have does not discuss the "Mouse Support Option", although it has it listed under the Advanced CMOS Setup screen on page 43. The manual does not go through any of the BIOS settings, with the exception of the AMI hard disk types, cache select, cache write wait state, and bus clock select. Mouse Support may remain a mystery...

There is a way to get a PS/2 port on a 386. Just talk to your friend chuck(g) at the vintage-computer forum. There was a discussion a few years back about designing an ISA card with a PS/2 port. The only catch is that it likely wouldn't work on IRQ12.
I don't know why few people in those forums seemed interested in buying one. If enough people ask for a card, we could probably have one built.

Anonymous Coward wrote:

There is a way to get a PS/2 port on a 386. Just talk to your friend chuck(g) at the vintage-computer forum.

This is the link,
http://www.vintage-computer.com/vcforum/showthread.php?23251

I replied to it, but I don't think anyone is willing to take lead on it, nor put in the brain grease. I would personally be interested in doing the design in my spare time, but I no longer have any (spare time). I still beleive that the best approach is with an improved RS-232 to PS/2 type of deal, similar to the photos I uploaded on the above link.

At this point you might as well just go for a USB interface. Is PS/2 protocol still supported in modern products? It seems to be hit or miss, implying that PS/2 is on the way out.

DCA2, if you get a chance can you test your M326 motherboard with either 8MB or 32MB of RAM. I am curious to know if you board supports page-interleave mode. You need to fill all 8 slots for it to work (if supported)

It took 8 hours and a babysitter, but I managed to repair the damaged traces on my Chaintech 333SCD/340SCD. There were 24 traces not making a connection. I also added a coin cell battery for CMOS battery backup. I was surprised when it actually turned on. While it runs cachechk, 3dbench, and pcpbench, it hangs on Speedsys during the Year2000 bug detection. Also, when loading the XMS/HIMEM in DOS, I get "Unable to Control A20 line. XMS driver not installed." This XMS/HIMEM error only occurs when the Fast Gate A20 Option is enabled in the BIOS. Elianda also report this Fast Gate A20 issue.

elianda & old thrashbarg, what is the date of your BIOS on this board? Mine is 7/7/91.

Aside from missing a broken trace or wiring a trace to the wrong location, any ideas on how to fix this Fast Gate A20 option and how to get Speedsys running? Anyone else have Speedsys hault on Year2000 bug detection when using a 386? How did you overcome it?

The board has a 66 Mhz oscillator, so I have left the Intel i386 DX-33 installed. If I can get this board functioning properly, I will change the oscillator to an 80 MHz piece. 256 KB cache is installed. MemTest v4.00 came back without any errors. I'm using a Trident graphics card for testing. I verified my solder connectivity at least twice.

I have two boards with date 07/07/1991. The topbench result I posted was from the one with eteq386 chipset:

CPU=AMD Am386DX
CPUspeed=40 MHz
BIOSinfo=R(C)1985-1991,American Megatrends Inc.,All Rights Reserved.,1346 Oakbrook Dr.,#120,GA-30093,USA.(404)-263-8181. (07/07/91, rev. 0)
BIOSdate=19910707
BIOSCRC16=F8DE

see also ftp://78.46.141.148/pictures/386dx40_sys1/
<--------------------->
The SIS board has the same date and has the same A20 line problem. I think speedsys does not halt on Y2K bug detection, as I have posted results already.

The difference is that the eteq board has its battery removed while the SIS board has a new battery soldered.

see also ftp://78.46.141.148/pictures/386dx40_sys2/

elianda or old thrashbarg, if you are willing could you remove your co-processor, change your oscillator from 80.0 MHz to 66.0 MHz, install an Intel i386 DX-33 and insert 256 kb of cache. With all these mods, were you able to properly launch Speedsys? If so, then my issue may be more likely related to the trace fixing.

This would be a real effort, as the system looks opened like this: ftp://78.46.141.148/pictures/386dx40_sys2/386desktop.jpg

I have no fitting 66 MHz oscillator here and also not enough working SRAMs with higher capacity. I would have to order this first somewhere. I could try to remove the FasMath though.

You could do a first check and look if the BIOS clock runs properly (not the MS-DOS clock). Maybe it is just not running and speedsys waits forever for the next tick on Y2K check.

What is the best method to see if the BIOS clock is running? Just going into the BIOS and make sure the seconds are ticking up?

elianda or old thrashbarg, if you are willing could you remove your co-processor, change your oscillator from 80.0 MHz to 66.0 MHz, install an Intel i386 DX-33 and insert 256 kb of cache.

Sorry, I can't really help you there. I have neither 256KB worth of cache chips, nor an Intel DX-33... and I'm not sure whether the the soldered AMD CPU can even be disabled anyway.

My BIOS date is 07/07/91, same as elianda's.