VOGONS


First post, by Anonymous Coward

User metadata
Rank l33t++
Rank
l33t++

Does anyone have any documentation about register programming for this chipset? In the backup of the American Megatrends website that has BIOS programming for all kinds of chipsets, documents for Acer parts are strangely missing. I say this is strange because AMIBIOS was quite popular on 386 and 486 boards with Acer (ALi) chipsets.

Something else that's pretty interesting is that in this thread: "Bioteq" chipset, Eep386 claimed that while working on a Biostar MB1433-AEA based on the FTDI chipset, he discovered "Acer" stenciled into the PCB under the chipset while doing what I guess must have been hot air work. He tested an Acer M1419 BIOS, which apparently worked. So it seems that the FTDI chipset must be some kind of clone of the M1419. I have since tested an Award BIOS V4.50 for M1419 in my Biostar MB1340-AEA, and can confirm it works perfectly.
American Megatrend's website interestingly did have BIOS programming for the FTDI (FTD, mislabled FDT) chipset. Similarly Microid Research also did not make a BIOS for the ALi M1419, but did for the FTDI, which is sadly missing from our archives.

In the M1419 AMIBIOSes I have tried, there is nothing in the setup that allows adjustment of the RAM or cache timings. Similarly, in the AMI documentation for the FTDI chipset I can find no registers programming information for these features. Does anyone know if either or both of these chipsets handles this automatically, or if there are some hidden/undocumented chipset registers?

"Will the highways on the internets become more few?" -Gee Dubya
V'Ger XT|Upgraded AT|Ultimate 386|Super VL/EISA 486|SMP VL/EISA Pentium

Reply 1 of 11, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie

I just looked too and the AMI, Award, and Acer BIOSes for this look pretty sparse.
Since they are uncompressed you can run a "strings" utility (e.g. on Linux or the Sysinternals version for Windows) and look for interesting strings/possible hidden options even if you don't have or can't open them in modbin, amibcp, etc.
It looks like the Acer BIOS has a hidden memory type settings, where "Slow" is for 80 ns or slower. That's probably part of what you are looking for. [edit: it's not hidden, I just didn't know about Page Down] But I see nothing at all about 486 cache burst settings.

Another indication that not much is known about this chipset is there is no emulation for it in PCem or 86box.

Reply 2 of 11, by Anonymous Coward

User metadata
Rank l33t++
Rank
l33t++

I didn't see the Acer BIOS for the M1419. Do you know what kind of system it came from?
Too bad there aren't any utilities for dealing with Acer BIOS. I have a pretty sweet dual Pentium system that has an Acer BIOS, and I'm not aware of any replacement options.

"Will the highways on the internets become more few?" -Gee Dubya
V'Ger XT|Upgraded AT|Ultimate 386|Super VL/EISA 486|SMP VL/EISA Pentium

Reply 3 of 11, by Anonymous Coward

User metadata
Rank l33t++
Rank
l33t++

In the FTDI chipset "manual" I did find some registers to control 486 burst mode timings:

36h
Bit 7 Internal local bus control
0 Disabled
1 Enabled
Bit 6 486 burst mode
0 Disabled
1 Enabled
Bits 5-4 Low speed rate selection
00 1:6 01 1:2
10 1:3 11 1:4
Bit 3 Cache write algorithm
0 Wrt/Thru
1 Wrt/Back
Bit 2 Refresh type
0 AT Bus
1 Hidden
Bits 1-0 486 burst mode timing
00 3-2-2-2 01 3-1-1-1
10 2-2-2-2 11 2-1-1-1

This may or may not work on the Acer M1419 chipset.

"Will the highways on the internets become more few?" -Gee Dubya
V'Ger XT|Upgraded AT|Ultimate 386|Super VL/EISA 486|SMP VL/EISA Pentium

Reply 4 of 11, by Eep386

User metadata
Rank Member
Rank
Member

Thank you for finding that snippet about the write back/thru bit!
I'd like to try punching that into CTCHIP to see if I can get these ALi M1419 386DX boards to stop using broken write back. 😅

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 5 of 11, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

This is all I get in Advanced Chipset menu on https://theretroweb.com/motherboards/s/acer-i … gxi-model-i433a

GxiSettingvgabench.jpg

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 6 of 11, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie

I got a Bioteq 82C3480 version of this. I'm running a 486DX CPU, L2-cacheless for now as it came with 25 ns chips.

In debug,
-i fca0
appears to be the RAM speed register, seems autoconfigured to 15h for 486DX-25 MHz, and 11h for 486DX-33 MHz or greater. That setting seems pessimistic.

-o fca0 15 will make it faster (at 33 MHz and higher)
-o fca0 17 will make even faster.

It doesn't appear to go beyond 17 or below 11, the machine locks up if you set it as such.

With fca0=17, and at DX-33 MHz, the system's RAM speed appears on-par with a maxed-out UMC 491 with external cache also disabled (to avoid conflating with the "always dirty" issue).
With fca0=17 and at DX-50 MHz, it's around the same speed, 51 us/KB for RAM. As it won't post at 50 MHz without JP10 closed, I suspect JP10 might be the "delay ALE by one cycle" for this board (MB-1433/50AEA).

Another user with a 386DX-33 board found fca0 to be set to 14h, so perhaps the bottom bit is 0 for 386 and 1 for 486

Reply 7 of 11, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie
Anonymous Coward wrote on 2024-07-20, 14:15:

I didn't see the Acer BIOS for the M1419. Do you know what kind of system it came from?
Too bad there aren't any utilities for dealing with Acer BIOS. I have a pretty sweet dual Pentium system that has an Acer BIOS, and I'm not aware of any replacement options.

BitWrangler wrote on 2024-11-01, 16:07:

This is all I get in Advanced Chipset menu on https://theretroweb.com/motherboards/s/acer-i … gxi-model-i433a

Do you still have your M1419 boards?
As the BIOS obviously doesn't tell us, I have to rely on benchmarking to determine the L2 cache speed. These are seemingly hardcoded for 2-1-1-1 when using a 486, when I compare the L2 cachechk for a 486DX-33 in these vs. a comparable UMC. It's too bad as I wanted to use the MB-1433/50-AEA for 50 MHz operation, but seems I'd have a tough time getting external cache to work.

Reply 8 of 11, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++
jakethompson1 wrote on 2025-08-06, 01:03:

Do you still have your M1419 boards?
As the BIOS obviously doesn't tell us, I have to rely on benchmarking to determine the L2 cache speed. These are seemingly hardcoded for 2-1-1-1 when using a 486, when I compare the L2 cachechk for a 486DX-33 in these vs. a comparable UMC. It's too bad as I wanted to use the MB-1433/50-AEA for 50 MHz operation, but seems I'd have a tough time getting external cache to work.

Still have it, but not set up and workbench occupied... had a look through my screenshots though, this is the only one that seemed vaguely relevant, was running at 40Mhz with a U5S in, settings probably same as above BIOS shot. IDK if you can work much out from it...

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 9 of 11, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie

That's either turbo button or something else horribly wrong

Reply 10 of 11, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

Possibly it was the hidden refresh not turned on, there's a lot of stuff that was way faster with it enabled.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 11 of 11, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie

These are stats with an Intel 486dx-33 with the chipset bit hack up above. I'm not able to use Hidden Refresh and External Cache together or the system will lock up. Obviously if you have to pick between the two it's external cache.

 CACHECHK v4 2/7/96  Copyright (c) 1995 by Ray Van Tassle. (-h for help)
CMOS reports: conv_mem= 640K, ext_mem= 19,456K, Total RAM= 20,096K
Clocked at 486 33.0 MHz
Reading from memory.
MegaByte#: --------- Memory Access Block sizes (KB)-----
1 2 4 8 16 32 64 128 256 512 1024 2048 4096 <-- KB
0: 32 32 32 32 37 37 37 37 37 53 -- -- -- µs/KB
1: 32 32 32 32 37 37 37 37 37 54 54 53 54 µs/KB
2 3 <--- same as above.
4: 32 32 32 32 37 37 37 37 37 54 54 54 54 µs/KB
5: 32 32 32 32 37 37 37 37 37 54 54 53 54 µs/KB
6 7 8 <--- same as above.
9: 32 32 32 32 37 37 37 37 37 54 54 54 54 µs/KB
10 11 12 13 14 15 16 17 <--- same as above.
18: 32 32 32 32 37 37 37 37 37 54 54 53 -- µs/KB
19 <--- same as above.

Extra tests----
Wrt 32 32 32 32 32 32 32 32 32 32 32 32 32<-Write mem
This machine seems to have one cache!? [read]
!! cache is 256KB -- 29.7 MB/s 35.3 ns/byte (144%) 4.4 clks
Main memory speed -- 20.6 MB/s 51.0 ns/byte (100%) [read] 6.4 clks
Effective RAM access time (read ) is 204ns (a RAM bank is 4 bytes wide).
Effective RAM access time (write) is 123ns (a RAM bank is 4 bytes wide).
Clocked at 486 33.0 MHz. Cache ENABLED.
Options: -t0