VOGONS

Search

Common searches

My 486 UMC8886/8881 Project (Version 2.0)

Topic actions

First post, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

I've switched from an ASUS PVI-486-SP3 to UMC based boards since I wanted to use 1024K of L2 cache.

Current config:
Shuttle HOT433, BIOS American Megatrends Inc. 433AUS2C from Shuttle homepage
AMD 5x86-133 ADZ @ 160 MHz
256 MB RAM FPM in SIMM 1 and SIMM 3 slot - or 128 MB RAM FPM in SIMM 1 and SIMM 3 slot
HighPoint HT370A PCI UDMA controller @ 40 MHz PCI
Tseng Labs ET6000 graphics card @ 40 MHz PCI
ISA network card @ 10 MHz ISA
Creative Labs SoundBlaster AWE32 Value CT3870 @ 10 MHz ISA
Hitachi Deskstar HDS721680PLAT80 80 GB HDD
Ricoh MP53160DA DVD RW (connected on internal controller)

Sorry for the bad image quality - I'll replace them soon.
I'll add some Benchmark results soon.

Goal is to get an AMD 5x86-133 running at 160 MHz (40x4).
The board should have 1024K of L2 cache.
RAM is 256 MB (with L2 in WT) or 128 MB (with L2 in WB).
Operating Systems should be DOS, Win9x and Windows 2000.

HOT433.jpg
Filename
HOT433.jpg
File size
531.93 KiB
Views
3769 views
File license
Fair use/fair dealing exception

I'm struggling with my current favourite board, the Shuttle HOT433, as described in another topic.

However, there is still a lot of work to be done in this project:
PS/2 Mouse port.
BIOS modifications. (L2 WB, PCI slots / IRQ routing)

I'll repeat my steps so far:

0) Empty Dallas has been modded. Gratefully, Shuttle has added traces for a coin cell battery socket on the PCB already.

feipoa wrote:
PC-Engineer wrote:

Shuttle is aware of the problems with SCSI. In their FAQ they recommend to use PCI SCSI controller only in PCI slot 1 (near ISA slot).

PCI slot 1 is normally where I place the graphics card so it can have some room to breath if I add a fan to the graphics heatsink.

Hi,
I've been working with a Shuttle HOT-433 v1-3 without PS/2 port since a few weeks.
I'm using a HighPoint 370A IDE controller to avoid PIO transfers.
I faced a few problems, when installing Windows 2000 (with Smart Boot Manager and cdrom connected to internal IDE port).
1) Windows 2000 failed when recognizing keyboard controller with 433AIP16 BIOS.
Solution: BIOS upgrade from 433AIP16 to 433AUS2C.
2) F6 disks worked in first step, but failed when copying the files.
Solution: Setting printer port back to ECP. Without, Windows assigns floppy DMA channel to paralell port (in SPP mode)!
3) Windows 2000 hard disk access has been halted and probably resetted for several seconds. Several entries in eventlog. Same with other IDE and SCSI controllers. No change when placing in other PCI slot.
Solution: Assign IRQ 12 to controller.
4) No way to install a PCI network card without removing the PCI mass storage controller.
Solution: Use ISA NIC.
5) When copying SP4 to the HDD and doing a verify over network, 4-byte-transfer errors occured by random but with system: address: 0x*3F? 0x7F? 0xBF? 0xFF? the integrients were replaced by first DWORD from following 16-byte-block (=cacheline): for example 0x*3F0-0x*3F3 by 0x*400-0x*403 or 0x*3F8-0x*3FB by 0x*400-0x*403 and so on... No difference with L2 256K 2banked, 256K 1banked, 512K 1banked, 1024K 2banked. No difference with 32M, 64M or 128M RAM (cacheable area respected).
Solution: Switching L2 from WB to WT. L2 WB is not working with a foreign BIOS. Please read this update: Re: My 486 UMC8886/8881 Project
6) System does not read a boot sector at 40 MHz FSB. Even with slowest settings, even with both caches disabled, even from floppy.
Solution: 33 MHz. Problem solved. Solution unknown. Board runs with 40 MHz FSB. PCI runs at 40 MHz without any issue. Additional information: HighPoint controller does not like clock frequencies below 33 MHz. It does not boot with 20 or 27 MHz.

Do you have any suggestions or similar problems?

I don't even know what happens when I try to add a PS/2 mouse port because of the IRQ 12 issue.

GA486IMv2.jpg
Filename
GA486IMv2.jpg
File size
373.64 KiB
Views
3769 views
File license
Fair use/fair dealing exception

I've a Gigabyte GA486IM REV.2 here, I don't use it yet. I'm using foreign BIOS AWARD 2A4X5G03 (from GA486AM).

MB-8433UUD-Av3_1.jpg
Filename
MB-8433UUD-Av3_1.jpg
File size
429.96 KiB
Views
3769 views
File license
Fair use/fair dealing exception

I've bought a BIOSTAR MB-8433UUD-A VER:3.1 which I've got without RTC/CMOS, BIOS and Clock generator. However, I haven't had time to use the original BIOS yet - consider it as WIP. (L2 not running with foreign BIOS AWD 2A4X5G03 or AMI 433AUS2C)
Perhaps I'll do the same amazing cache mod that feipoa did when I do not get the mouse port running on my HOT-433.

Cards1.jpg
Filename
Cards1.jpg
File size
561.78 KiB
Views
3769 views
File license
Fair use/fair dealing exception

I'm using this cards:
PCI ET6000 (with 2 megs yet)
PCI Highpoint HT370A
ISA Network card
ISA Sound Blaster 32 AWE Value CT3780
an ISA / PCI POST card

The RAMs are all double sided:
32x 16 MBit x 1 + 4x 16 MBit x 1 -> 64 megs (1 rank)
16x 16 MBit x 4 + 8x 4 MBit x 1 -> 128 megs (2 ranks)

Hints:
Booting CDROM from a floppy disc with Smart Boot Manager.
Partitions: 500 MB DOS FAT16, 5500 MB Win2000 NTFS, 2000 MB Logical FAT16 volume in 2000 MB Extended Partition. Above DOS' 8 GB limit: 66 GB Win98SE FAT32. (FAT32 partition created with h2format from Heise / c't)
Boot Manager: Added Bootsectors to boot.ini on NTFS partition.

L1 is running at WB mode without any problems. (L1 WB / WT has to be changed by the blue jumper on the motherboard (below the coin cell battery)
L2 cache timing is 2-1-1-1 (2-1-2).
RAM timing is 1 WS Read, 0 WS Write.

Current problems:

  • BIOS / L2 WB
    L2 is running in WB with foreign BIOS AWD 2A4X5G03 read Re: My 486 UMC8886/8881 Project
    AWARD BIOS has a Y2K bug -> year is fixed at 2094. foreign BIOS taken from BIOSTAR board
    AWARD BIOS does not allow Win2000 to boot since a loop during an IRQ issue when initializing the HighPoint driver. IRQ routing fixed by BIOS hack
  • BIOS / IRQ routing
    L2 is running in WT with BIOS AMI 433AUS2C
    The HighPoint controller does busmaster DMA and it seems to work with IRQ 12 only. IRQ routing fixed by BIOS hack
    More than one PCI cards do not function. (I do not count the graphics card.) IRQ issue. IRQ routing fixed by BIOS hack
    BIOS and differences have to be examined. Done!

Tests:
For distinguishing L1 between WB and WT: ctcm 1.7a from Andreas Stiller (Heise / c't) (WB is > 100 MB/s, WT is < 100 MB/s)
For stability of L2 @ WB: comparing a ~ 130 MB file (W2K SP4) with fc /b over network. (see above for details)

Last edited by Disruptor on 2020-01-19, 11:36. Edited 3 times in total.

Reply 1 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

BIOS settings

WB160_BIOS_Chipset.jpg
Filename
WB160_BIOS_Chipset.jpg
File size
516.51 KiB
Views
3692 views
File license
Fair use/fair dealing exception

Some Benchmarks
SpeedSys 4.78

WB160_Speedsys.png
Filename
WB160_Speedsys.png
File size
25.12 KiB
Views
3692 views
File license
Fair use/fair dealing exception

Landmark

WB160_Landmark.jpg
Filename
WB160_Landmark.jpg
File size
355.51 KiB
Views
3692 views
File license
Fair use/fair dealing exception

Superscape 98.4

WB160_Superscapek.jpg
Filename
WB160_Superscapek.jpg
File size
315.35 KiB
Views
3692 views
File license
Fair use/fair dealing exception

Quake 1.06 54.0 / 17.9 fps

WB160_Q3.jpg
Filename
WB160_Q3.jpg
File size
291.98 KiB
Views
3692 views
File license
Fair use/fair dealing exception

Chris 119.8 / 71.9 fps
Chris SVGA 27.5 / 16.5 fps
PCP VGA 27.6
PCP VESA 8.5
DOOM 1239 realticks
SI 8.0 345,7
TOPBENCH 405

Last edited by Disruptor on 2019-12-30, 12:27. Edited 1 time in total.

Reply 2 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

Running my Shuttle HOT-433 with 128 MB RAM in 1024 K L2 in WB (write back) mode:

WB160_CPUZ.png
Filename
WB160_CPUZ.png
File size
6.89 KiB
Views
3690 views
File license
Fair use/fair dealing exception
WB160_Cachechk.jpg
Filename
WB160_Cachechk.jpg
File size
480.06 KiB
Views
3690 views
File license
Fair use/fair dealing exception

Update:

Note: The board runs Windows 98 with L2 in WB mode because the HighPoint driver runs the controller in a kind of fallback mode since there is a problem with the BIOS' IRQ routing table.
Since the routing table is now fixed by a BIOS hack, it is no longer possible to run it in this mode.

Benefits:
HighPoint controller runs as busmaster with IRQ enabled.
I'm able to populate all 4 PCI slots now.
I can use all IRQs.
I can run Windows 2000 with a AWARD BIOS.
I can use a Plug & Play BIOS.

Last edited by Disruptor on 2020-01-19, 11:34. Edited 2 times in total.

Reply 3 of 27, by Nemo1985

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

That's very interesting, thank you for sharing such benchmarks.
I am interested to do a head to head ASUS PVI-486-SP3 vs BIOSTAR MB-8433UUD-A, basically SIS vs UMC.
I'm using an Amd 5x86 at 133 with write back on both l1 and l2, 32mb ram and a Virge 4mb.
Actually with my Asus and fastest memory timings are lower than yours on speedsys:
L1: 117,74
L2: 51,03
Memory: 38,25

Out of curiosity what is the mb you prefer so far and how do you get such detailed speedsys screens, mine lacks all the lower part of the screen (Cache memory benchmark, memory map and PCI\AGP Device information)?

Attachments

  • photo_2020-01-12_17-52-15.jpg
    Filename
    photo_2020-01-12_17-52-15.jpg
    File size
    196.63 KiB
    Views
    3585 views
    File comment
    Cachechk
    File license
    Fair use/fair dealing exception

Reply 4 of 27, by pshipkov

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

Good read. Thank you.

@Nemo1985
Not far ago i actually compared the Biostar UMC and and ASUS SP3 mobos here: 3 (+3 more) retro battle stations (page 3).
I think there is a big thread (the ultimate 486) that has comparison of the two as well, only basic dos tests, results are fused into uber charts, but still - a great source of information.

Last edited by pshipkov on 2020-01-12, 20:57. Edited 1 time in total.

retro bits and bytes

Reply 5 of 27, by Nemo1985

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie
pshipkov wrote on 2020-01-12, 17:47:
Good read. Thank you. […]
Show full quote

Good read. Thank you.

@Nemo1985
Not far ago i actually compared the Biostart UMC and and ASUS SP3 mobos here: 3 (+3 more) retro battle stations (page 3).
I think there is a big thread (the ultimate 486) that has comparison of the two as well, only basic dos tests, results are fused into uber charts, but still - a great source of information.

Very interesting thank you!

Reply 6 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie
Nemo1985 wrote on 2020-01-12, 16:47:
That's very interesting, thank you for sharing such benchmarks. I am interested to do a head to head ASUS PVI-486-SP3 vs BIOSTAR […]
Show full quote

That's very interesting, thank you for sharing such benchmarks.
I am interested to do a head to head ASUS PVI-486-SP3 vs BIOSTAR MB-8433UUD-A, basically SIS vs UMC.
I'm using an Amd 5x86 at 133 with write back on both l1 and l2, 32mb ram and a Virge 4mb.
Actually with my Asus and fastest memory timings are lower than yours on speedsys:
L1: 117,74
L2: 51,03
Memory: 38,25

I've a PVI-486-SP3 too. However, since I got fastest timing @ 160 MHz only with 256 K double banked cache, and I'm limited to 32 M RAM with WB L2.

Nemo1985 wrote on 2020-01-12, 16:47:

Out of curiosity what is the mb you prefer so far and how do you get such detailed speedsys screens, mine lacks all the lower part of the screen (Cache memory benchmark, memory map and PCI\AGP Device information)?

It's merged from 3 speedsys pics using Paint Shop Pro.
You can switch to the other screens using I and M with speedsys.

Reply 7 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

Update

Disruptor wrote on 2019-12-29, 11:38:

BIOS / L2 WB WT!
L2 is running in WB with foreign BIOS AWD 2A4X5G03

This is only true, when running Windows 98 with a BIOS that does not assign an IRQ to the HighPoint controller
It seems that the controller's Win98 driver changes to a fallback method when it detects the IRQ problem.
When using another foreign BIOS (from BIOSTAR MB-8433UUD-A VER:3.1 board) with fixed IRQ mapping, Win98 uses HighPoint controller in normal operation mode. There I have read errors like I have described in my post above.
I no longer can say this board runs PCI busmasters properly in L2 WB method.
Conclusion: Switch to L2 WT. Use full 256 MB RAM.

Reply 8 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

Gigabit Ethernet on a 486?

I've tested copying a 130 MB file from a 486 into a Windows 10 computer's NUL device.

The opponents were three network interface cards (NIC):
D-LINK DGE-500T 1 GBit/s
3COM 3C905B-TX 100 MBit/s
3COM 3C905C-TX 100 MBit/s
VIA VT6122 1 GBit/s

The server was:
5x86 133 @ 160 MHz in a Shuttle HOT-433 with 1024 KB L2 WT Cache.
PCI Clock 40 MHz (Note: Gigabit card is officially capable to run at 66 MHz)
RAM: 256 MB (file was fully in Windows 2000' disk cache)

TCP offloading was enabled on the D-LINK DGE-500T.

The result was for the first three cards 48 MBit/s.
The VT6122 drops 49 MBit/s out of Gigabit but 47 MBit/s out of 100 MBit/s.

Conclusion: There is almost no difference in performance in a 486 whether you use a Gigabit NIC or a 100 MBit NIC.

Reply 9 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

Finally, the PS/2 mouse port is running.

After several hours of work a mate got an external keyboard controller running in the HOT-433.
It was a bit tricky with some soldering work and patching a BIOS.

We'll add more details this day.

Specs (preliminary):
HOT-433 (Ver 1-3)
External Keyboard Controller
FSB 40 MHz
PCI 40 MHz
1 MB L2 Cache
256 MB FPM PS/2 DRAM (fully cached in write-through)
AMD 5x86 133 MHz @ 160 MHz
PCI HighPoint HT370A PCI UDMA controller @ 40 MHz
PCI USB 2.0 adapter @ 40 MHz
PCI Tseng Labs ET6000 graphics card @ 40 MHz
PCI 3COM 3C905C-TX (the classic model) @ 40 MHz
ISA SoundBlaster 16 CT2910 (with OPL) @ 10 MHz
HDD PATA Hitachi Deskstar HDS721680PLAT80 80 GB (on controller with ATA/100)
DVDRW 16x (on internal port with PIO only)
PS/2 optical mouse
DIN adapter for PS/2 keyboard

Last edited by Disruptor on 2021-05-21, 08:20. Edited 4 times in total.

Reply 13 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie
pii_legacy wrote on 2020-06-21, 13:03:

Wow someone else here has a HOT-433 with a 5x86! That is awesome, I'll be very excited to hear details and see if the same PS2 mouse mod is do-able on the HOT-433 v4.0.

What date is your BIOS?

Source was an AWARD with name 2A4X5H21
However, the mouse code has to be enabled with MODBIN

I don't have a HOT-433 v4.0

Reply 14 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie
Disruptor wrote on 2020-06-21, 09:29:
Finally, the PS/2 mouse port is running. […]
Show full quote

Finally, the PS/2 mouse port is running.

After several hours of work a mate got an external running in the HOT-433.
It was a bit tricky with some soldering work and patching a BIOS.

We'll add more details this day.

Success: Adding a PS/2 mouse port to the Shuttle HOT433 (using a dedicated keyboard controller chip)

Reply 16 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

So, I've played PCI card puzzle.
Conclusion: The network card (or the driver) does not like IRQ sharing.

PCI 1 (near ISA): Network card (IRQ 11)
PCI 2: Storage card (IRQ 10)
PCI 3: Graphics card (IRQ 9; no IRQ in Windows 2000)
PCI 4: USB 2.0 card (takes 2 different IRQs, 9 for 1.1, 10 for 2.0)

Reply 17 of 27, by Disruptor

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

Playing a MP3 file with WinAMP 2.81 on my SB16 with 44 kHz, 16 bit, stereo is no problem.
File source was a 192 kbit/s MP3 file over network from a Windows 10 share (SMB1 enabled).

I've changed the output plugin from DirectSound to waveOut.

Reply 18 of 27, by pii_legacy

Posted on 
User metadata
Rank Newbie
Rank
Newbie
Disruptor wrote on 2020-06-27, 22:32:
So, I've played PCI card puzzle. Conclusion: The network card (or the driver) does not like IRQ sharing. […]
Show full quote

So, I've played PCI card puzzle.
Conclusion: The network card (or the driver) does not like IRQ sharing.

PCI 1 (near ISA): Network card (IRQ 11)
PCI 2: Storage card (IRQ 10)
PCI 3: Graphics card (IRQ 9; no IRQ in Windows 2000)
PCI 4: USB 2.0 card (takes 2 different IRQs, 9 for 1.1, 10 for 2.0)

I just had this same trouble trying to add a USB 2.0 card! It wouldn't boot without an error unless I rearranged the order of them physically. Wish the UMC chips had some documentation.

Reply 19 of 27, by citronalco

Posted on 
User metadata
Rank Newbie
Rank
Newbie
Disruptor wrote on 2019-12-30, 12:21:

Update:

Note: The board runs Windows 98 with L2 in WB mode because the HighPoint driver runs the controller in a kind of fallback mode since there is a problem with the BIOS' IRQ routing table.
Since the routing table is now fixed by a BIOS hack, it is no longer possible to run it in this mode.

I have a similar problem with that board.
Do you still remember what kind of hack you did? Or: Could you upload your modded BIOS here?

Go to top of page Go to top of page
Back to System Specs
Close window
Close window