XP running on a 486 cpu

[roytam1]

and if you want cheaper, you may try some VT6421/VT6421a based cards, some of them having a PLCC LPC Flash chip socketed/soldered.

[Dietmar]

I tested about 20 different PCI to IDE/Sata cards with own Bios Rom. I flash a lot with EEProm writer.

Always after this procedure I can see the connected HD or CD-Rom drive or both at boot time.

But I never succeed with booting XP SP3 via any of this cards on the Shuttle Hot 433 board with 486 cpu.

Also I cant install XP SP3 via any of this cards. Before I have to hit F8, I always get Bsod 0x7B, even the driver for this harddisk is to 100% correct and there.

So I start to think, that the Bios of this board  is faulty as much as possible or

it is even impossible to boot via any of the PCI slots of this board.

Only the own 2 integrated IDE slots work and the Floppy connector

Dietmar

Edited May 8 by Dietmar

[roytam1]

3 hours ago, Dietmar said:

I tested about 20 different PCI to IDE/Sata cards with own Bios Rom. I flash a lot with EEProm writer.

Always after this procedure I can see the connected HD or CD-Rom drive or both at boot time.

But I never succeed with booting XP SP3 via any of this cards on the Shuttle Hot 433 board with 486 cpu.

Also I cant install XP SP3 via any of this cards. Before I have to hit F8, I always get Bsod 0x7B, even the driver for this harddisk is to 100% correct and there.

So I start to think, that the Bios of this board  is faulty as much as possible or

it is even impossible to boot via any of the PCI slots of this board.

Only the own 2 integrated IDE slots work and the Floppy connector

Dietmar

I wonder if BIOS' PnP handles SATA cards correctly.

for non-PnP device, windows handling them in a tricky way.

you may hook the harddisk and card to another newer PC to see if it can be booted or not.

see also: https://www.betaarchive.com/forum/viewtopic.php?t=45422

[Dietmar]

@roytam1

After soso many fails I succeed to install XP SP3 to the Shuttle Hot 433 board with 486 cpu

via an PCI slot using the PCI to IDE controller card Dawicontrol DC-133 and

fresh flashed Bios from Silicon

https://ufile.io/c41pi7rw

Videocard is in Slot 1 (near to the ram) and DC-133 card is in 3. slot (near to the ISA).

Until now I use a 8GB IDE harddisk but this I will change soon against an SSD via an IDE adapter

Dietmar

EDIT: The jumper for RAID I put away from the DC-133 card.

The jumper for Bios Rom I set.

Edited May 10 by Dietmar

[Dietmar]

Now I also succeed to install XP SP3 on the Shuttle Hot 433 board to a 486 cpu using the

Dawicontrol DC-150, with newest flashed Bios for NON-Raid modus.

This card is a PCI to Sata card, this means for the very first time I can boot XP SP3 for a 486 cpu via a fast Sata HD,

nice

Dietmar

 

EDIT: Strange, during install via PCI to IDE/Sata card the install lasts much longer than with the integrated IDE ports?!

A HD speed test with CrystalDiskMark5 using the integrated IDE ports, always crashes.

I hope, that I can get some valid results via the PCI cards.

EDIT2: Oh..crazy, the Bios puts in Standard Settings the speed of the PCI bus to halve of its original value..

Edited May 11 by Dietmar

[user57]

that with the drive wasnt complicated to program 

you had a buffer to progress (that buffer being the data that is being written to that hdd disc)

the hdd drive itself had certain commands it understand "like write or read" 

you processed it directly via IO command (aka "in" and "out" assembly)

you had to a in BYTE(8 bits each)/WORD(16 bits each)/DWORD (32 bit each)  - this got repeated with the REP command (repeat command if easy speacing)

 

so here let me take this approach -> that IN OUT command write/read not directly to the HDD drive rather it write that to the CACHE of the HDD

the firmware ... then progress it to a physical movement and electric pulse to the HDD head (ect.)

so to make people understand why i have to say this first 

the CPU is a lot faster then the HDD is and the CPU as might i said in the past can translate assembly commands to a "different progressing"

with different progressing i mean it can detect a 8 bit move and translate it to a 64 bite (or a lot more) move

 

so now that we know that the HDD cant progress the speed power the CPU has (and therefore use a cache (+ i must say this method got increased up to fast bigger cache like 12 MB or more cache))

you actually read out the status code for the HDD (the hdd then tells you like "im doing this im busy (do not send me data again for now")

this is a direct hardware approach

 

 

the "new" ones a "engine" - yes engine you heared right - i say so because everything today is going through like 8 engines before doing anything (and yes i say it again stop that bullc... everything it filled up with engines today - use less of engines programmers)

this "new" is done via a driver (pre driver) and a IRP 

https://learn.microsoft.com/en-us/windows-hardware/drivers/ddi/wdm/ns-wdm-_irp

this IRP is then being progressed - but it still do that classical IN and OUT command just to say 

 

this is not a direct hardware approach but it gives the operating system a "standart code" it can operate with - its far less effective and therefore also have less speed (the reason you dont see this so directly is for example that cache reason)

 

 

 

 

 

but enough about the disc and back to dietmars problem more directly i dont think we can see the reason so directly without having the right tools but the pci bus is fast enough 

there are certainly differences between PCI standarts either

https://en.wikipedia.org/wiki/Peripheral_Component_Interconnect

 

the SATA to IDE was already called out to be a idea  

that however raise the question what standart or norm for that older cpu is working

https://de.wikipedia.org/wiki/ATA/ATAPI

https://www.elektronik-kompendium.de/sites/com/0501021.htm

https://de.wikipedia.org/wiki/Programmed_Input/Output

 

this like we creating our first HDD 

[Dietmar]

@user57

There must be something going complete wrong,

using the PCI bus for installation of XP SP3 in compare with the original implemented IDE slots

on the Shuttle Hot 433 board with 486 cpu.

For the first TXT-Setup of XP I need about 45 min with the IDE slots, same cpu.

With a PCI to IDE or PCI to Sata card, which should have full performance of 132 MByte/sec of the PCI bus

(33 MHZ, 32 bit with 32 bit=4*8bit, gives 132 MByte/sec).

But Now I need about 9(!) hours for the same TXT mode. This is a factor of 12 compared to the original IDE slots.

So, the real speed of the PCI Bus is not 132 Mbyte/sec but for sure less than 11 MByte/sec,

even the connected Sata or IDE harddisk is about 60 MByte/sec.

And the PCI cards fulfill all to be faster than 132 MByte/sec

Dietmar

 

EDIT: The speed of this PCI bus is less than an original CD-ROM with speed 1, means 150Kb/s.

Whole install of XP via Floppy disk with 200kB/sec would be faster..

Edited May 12 by Dietmar

[roytam1]

6 hours ago, Dietmar said:

@user57

There must be something going complete wrong,

using the PCI bus for installation of XP SP3 in compare with the original implemented IDE slots

on the Shuttle Hot 433 board with 486 cpu.

For the first TXT-Setup of XP I need about 45 min with the IDE slots, same cpu.

With a PCI to IDE or PCI to Sata card, which should have full performance of 132 MByte/sec of the PCI bus

(33 MHZ, 32 bit with 32 bit=4*8bit, gives 132 MByte/sec).

But Now I need about 9(!) hours for the same TXT mode. This is a factor of 12 compared to the original IDE slots.

So, the real speed of the PCI Bus is not 132 Mbyte/sec but for sure less than 11 MByte/sec,

even the connected Sata or IDE harddisk is about 60 MByte/sec.

And the PCI cards fulfill all to be faster than 132 MByte/sec

Dietmar

 

EDIT: The speed of this PCI bus is less than an original CD-ROM with speed 1, means 150Kb/s.

Whole install of XP via Floppy disk with 200kB/sec would be faster..

maybe you need to check/tweak BIOS settings?

EDIT: you may try the Award BIOS update: https://theretroweb.com/motherboards/s/shuttle-hot-433-ver.-4.0#bios

Edited May 12 by roytam1

[roytam1]

On 5/12/2024 at 12:54 AM, Dietmar said:

A HD speed test with CrystalDiskMark5 using the integrated IDE ports, always crashes.

you may try the retro version: https://crystalmark.info/en/software/crystalmarkretro/

 

for normal version, you should try 3.0.4. as newer versions comes with CMOV.

Edited May 13 by roytam1

[Dietmar]

@roytam1

Oh,

thanks, I am going to test.

https://www.youtube.com/watch?v=zqsKbE0qH2I

Just I read this

https://www.pcgameshardware.de/Windows-Software-277633/News/Intel-i486-CPU-XP-1447805/

https://www.xda-developers.com/windows-xp-running-cpu-ten-years-older-itself/

https://www.tomshardware.com/software/windows/enthusiast-gets-windows-xp-running-on-an-i486

https://www.reddit.com/r/windowsxp/comments/1ctlglw/windows_xp_finally_works_on_486_processors/

https://itc.ua/en/news/a-combination-of-two-eras-windows-xp-launched-on-intel-i486-processor/

Do you have an idea, why the speed on a HD is so slow and using the PCI-slots with PCI to IDE/Sata is even more slow

Dietmar

Edited Sunday at 07:28 PM by Dietmar

[roytam1]

10 hours ago, Dietmar said:

@roytam1

Oh,

thanks, I am going to test.

https://www.youtube.com/watch?v=zqsKbE0qH2I

Just I read this

https://www.pcgameshardware.de/Windows-Software-277633/News/Intel-i486-CPU-XP-1447805/

https://www.xda-developers.com/windows-xp-running-cpu-ten-years-older-itself/

https://www.tomshardware.com/software/windows/enthusiast-gets-windows-xp-running-on-an-i486

https://www.reddit.com/r/windowsxp/comments/1ctlglw/windows_xp_finally_works_on_486_processors/

https://itc.ua/en/news/a-combination-of-two-eras-windows-xp-launched-on-intel-i486-processor/

Do you have an idea, why the speed on a HD is so slow and using the PCI-slots with PCI to IDE/Sata is even more slow

Dietmar

no idea, could be chipset problem or driver problem.

BTW neowin reported this as well. lol.

https://www.neowin.net/news/intels-legendary-i486-cpu-gets-unofficial-windows-xp-support-iso-too/

Edited Monday at 04:07 AM by roytam1

[Dietmar]

Hi,

until now, I do all my tests on real hardware:

The 486 Shuttle Hot 433 board, vers. 1..3 with 1024kB Cache and 256MB Ram.

There I flashed the Bios from Award 2001(!) . It has a mod, for to disable the PS/2-mouse by default.

mio.bin

https://ufile.io/trzn0n6x

 

Now I try to use the nice Emulator for x86 cpus

https://github.com/86Box/86Box

 

Is there a way, to add this mio.bin Bios to the 86Box?

 

Because this Bios allows boot from Cdrom and large harddisks

Dietmar

PS: By the way, which is the last version of this Emulator, that works under XP SP3?

[roytam1]

3 hours ago, Dietmar said:

Is there a way, to add this mio.bin Bios to the 86Box?

for new machine(not ever started), replacing roms\machines\hot433\433AUS33.ROM with yours.

for already started machine, replace nvr\433AUS33.ROM with yours.

[Dietmar]

@roytam1

I tried this, but get only a lot noise at start.

No enumerating of memory.

I think, that those roms are hard wired

Dietmar

[roytam1]

29 minutes ago, Dietmar said:

@roytam1

I tried this, but get only a lot noise at start.

No enumerating of memory.

I think, that those roms are hard wired

Dietmar

I replaced hot433.bin with yours, and it works.