Sorry, forget point 2). I just needed to search here on Vogons, there is a patched asus bios 0402.002 and one more other bios.
I just need to try, but I have to check if a have a spare 27c512 and a working programmer.

But question 1 is still interesting for me. Anyone tried to put 128MB RAM on the SV2GX4?

Sorry, not quite what you've asked, but keep in mind that this side of a very sluggish WinXP, you won't need more than 64 MB of RAM.

Officially, the SIS 471 chipset supported 128 MB.
If many sources claim the board only supports 64, that might be true due to other constraints.
Or, it merely was assumed because there were no 32 MB modules known to man at that moment.
If it eats one 32 MB module, it should eat four of them.

Cache: Same here, the mobo won't do just anything the chipset was capable of. IIRC, "write back" also depended on the CPU and most DX2 ones did not offer that.

this mobo maxes out at 64mb system memory.

pshipkov wrote on 2022-03-27, 19:27:

this mobo maxes out at 64mb system memory.

Thought as much.
Chipset maximum often means that you'd need a board with 8 sockets of PS2 which basically doesn't exist.

Cuttoon wrote on 2022-03-27, 19:44:

pshipkov wrote on 2022-03-27, 19:27:

this mobo maxes out at 64mb system memory.

Thought as much.
Chipset maximum often means that you'd need a board with 8 sockets of PS2 which basically doesn't exist.

I understand what you mean, but that is not the case here. According to the chipset, it would even support 64MB modules.

Not quite sure any more how the whole "double sided / single sided" business used to work but that usually was what made the differnce.

I am not aware of 64Mb FPM 72-pin SIMM modules. Only EDO, but the board/chipset does not like them.

Now a third question 😁

VL/I-486SVGOX4 Revision 1.2:
Anybody has that board?
The question is how to enable L1 write back for an Am5x86.

The Bios on that board was super old (Revision 2), but is now replaced with the new version with the "write back fix".

Enabling L1 write back does not work for me.
Please note, that the jumper block for setting the CPU type is different to Revision 2.x.

When I configure the jumpers for P24CT (including the 2 for the CPU trap) => not only L1 is in WT, L2 is disabled then. WTF.
JP19 is set to 1-2 always that gives multiplier 4, that works!

I post the "manual" = writings on the mainboard:

So the correct setting SHOULD be P24CT and cpu trap => P24T (so JP7 and JP8 both 1-2)

I found a big manual pdf for a lot of mainboards here on Vogons, cannot find it anymore now 😁 There was a setting for Rev.1.1 and Rev.1.2, and it was slightly different.
There was a P24CT56 (or something like that).
JP18: x
JP19: x
Jp20: 1-2, 4-5
JP21: 1-2

Threas is not of high priorty, because I also have the other board with the revision 2.0, there L1 write back works with Am5x86.

In case somebody stumbles over that thread again.
I found the "The PC Engineer's Reference Book" again .....
Here are settings for the Rev.1.2 board:

It is a bit different to what is printed on the board, 🤣.

So no dice for L1 WB.
But the DX4 2x - setting is working fine for the Am5x86, with L1 WT only of course (EDIT)

I've seen your posts in the other thread so I know you've got Write-Back working with your 1.2 board 😀 I dug out my VL/I-486SVG0X4 today to see what I'd find since I also have a 1.2 revision and there's so little information, it must be rare - looks like you already found that JP18 pin 2 goes to the WB/WT pin: Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?

I couldn't figure out how the regulator goes from 5v to 3.4v, I couldn't see any jumpers specifically for CPU voltage or specific voltage, so they're hiding in the JP18>JP22 block. dopefish went through how the voltage regulation works while going over how to upgrade boards without voltage regulators:
Another high-end 486 back in action (ASUS GX4 conversion success! Tons of pictures)
Since my board has the voltage regulation parts, I tested out VOLDET without the CPU fitted since it was set up for the DX/2-66 I had got the board running with. Once I confirmed the voltage would go down to 3.4, I was okay with fitting one of my AM5x86-P75 CPUs to the board. I think that the voltage detection is enabled with jumper JP20 at 4-5, can't recall if there was another one but jumpering for a DX/4 and fitting a 3.45v CPU does result in the lower voltage.

From there I found Jan Steunebrink's page which gave some more information on how WB/WT works: http://www.steunebrink.info/amd5x86.htm I'm so glad this page is still active 😀
I checked out that JP18 pin 2 goes to the B13 WB/WT pin while JP18 pin 1 goes to T1 which is the PODP WB/WT pin, Jan's page confirms that connecting B13 to a chipset WB/WT pin should work and this does, so I guess the PODP WB/WT goes to the chipset also.
I was initially skeptical and tried connecting pin2 to VCC with a jumper wire - which also works:

Since reading your last posts in the other thread, there's now just a regular jumper on JP18 between 1-2 and that gives write-back mode every time as well:

So it seems that the only thing that needs changing from the manual page that we have (from the PC engineers reference book), to have the AMD 5x86-P75 working at full speed on the VL/I-486SVGOX4 Revision 1.2 is (along with newest SV2G bios):

1CPU			| JP18			| JP19			| JP20		| JP21		| JP22
2AMD-X5-133	| 1-2,3-4,5-6	| 5-6 (2x=4x)	| 4-5		| 1-2		| 2-3,4-6

It's definitely working in write-back mode, benchmarks are noticeably quicker with it enabled

Thermalwrong wrote on 2023-09-18, 01:31:

I've seen your posts in the other thread so I know you've got Write-Back working with your 1.2 board :) I dug out my VL/I-486SVG […]
Show full quote

I've seen your posts in the other thread so I know you've got Write-Back working with your 1.2 board 😀 I dug out my VL/I-486SVG0X4 today to see what I'd find since I also have a 1.2 revision and there's so little information, it must be rare - looks like you already found that JP18 pin 2 goes to the WB/WT pin: Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?

I couldn't figure out how the regulator goes from 5v to 3.4v, I couldn't see any jumpers specifically for CPU voltage or specific voltage, so they're hiding in the JP18>JP22 block. dopefish went through how the voltage regulation works while going over how to upgrade boards without voltage regulators:
Another high-end 486 back in action (ASUS GX4 conversion success! Tons of pictures)
Since my board has the voltage regulation parts, I tested out VOLDET without the CPU fitted since it was set up for the DX/2-66 I had got the board running with. Once I confirmed the voltage would go down to 3.4, I was okay with fitting one of my AM5x86-P75 CPUs to the board. I think that the voltage detection is enabled with jumper JP20 at 4-5, can't recall if there was another one but jumpering for a DX/4 and fitting a 3.45v CPU does result in the lower voltage.

From there I found Jan Steunebrink's page which gave some more information on how WB/WT works: http://www.steunebrink.info/amd5x86.htm I'm so glad this page is still active 😀
I checked out that JP18 pin 2 goes to the B13 WB/WT pin while JP18 pin 1 goes to T1 which is the PODP WB/WT pin, Jan's page confirms that connecting B13 to a chipset WB/WT pin should work and this does, so I guess the PODP WB/WT goes to the chipset also.
I was initially skeptical and tried connecting pin2 to VCC with a jumper wire - which also works:
vli486svgox4-5x86-jumpers.jpg
Since reading your last posts in the other thread, there's now just a regular jumper on JP18 between 1-2 and that gives write-back mode every time as well:
vli486svgox4-5x86-wb-enabled.jpg

So it seems that the only thing that needs changing from the manual page that we have (from the PC engineers reference book), to have the AMD 5x86-P75 working at full speed on the VL/I-486SVGOX4 Revision 1.2 is (along with newest SV2G bios):

Copy code to clipboard

1CPU			| JP18			| JP19			| JP20		| JP21		| JP22
2AMD-X5-133	| 1-2,3-4,5-6	| 5-6 (2x=4x)	| 4-5		| 1-2		| 2-3,4-6

It's definitely working in write-back mode, benchmarks are noticeably quicker with it enabled

Hello.

I am a bit confused about my own posting:
Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?
What picture? There is no picture attached. WTF

EDIT: Adding pictures is not possible at the moment.
But my jumper setting is: (JP18 and JP19 are 1-6, JP20 to JP21 are 1-5 only)

1CPU			| JP18			| JP19			| JP20		| JP21		| JP22
2AMD-X5-133	| 1-2,3-4,5-6	| 2-3,5-6       | 2-3,4-5	| 1-2		| 1-2,4-5

MORE EDIT: Typo corrected. JP22 is 1-2, 4-5.
Picture added. In case of differences, the setting in the picture counts, haha.

CoffeeOne wrote on 2023-09-19, 18:42:

Hello. […]
Show full quote

Thermalwrong wrote on 2023-09-18, 01:31:

I've seen your posts in the other thread so I know you've got Write-Back working with your 1.2 board :) I dug out my VL/I-486SVG […]
Show full quote

I've seen your posts in the other thread so I know you've got Write-Back working with your 1.2 board 😀 I dug out my VL/I-486SVG0X4 today to see what I'd find since I also have a 1.2 revision and there's so little information, it must be rare - looks like you already found that JP18 pin 2 goes to the WB/WT pin: Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?

I couldn't figure out how the regulator goes from 5v to 3.4v, I couldn't see any jumpers specifically for CPU voltage or specific voltage, so they're hiding in the JP18>JP22 block. dopefish went through how the voltage regulation works while going over how to upgrade boards without voltage regulators:
Another high-end 486 back in action (ASUS GX4 conversion success! Tons of pictures)
Since my board has the voltage regulation parts, I tested out VOLDET without the CPU fitted since it was set up for the DX/2-66 I had got the board running with. Once I confirmed the voltage would go down to 3.4, I was okay with fitting one of my AM5x86-P75 CPUs to the board. I think that the voltage detection is enabled with jumper JP20 at 4-5, can't recall if there was another one but jumpering for a DX/4 and fitting a 3.45v CPU does result in the lower voltage.

From there I found Jan Steunebrink's page which gave some more information on how WB/WT works: http://www.steunebrink.info/amd5x86.htm I'm so glad this page is still active 😀
I checked out that JP18 pin 2 goes to the B13 WB/WT pin while JP18 pin 1 goes to T1 which is the PODP WB/WT pin, Jan's page confirms that connecting B13 to a chipset WB/WT pin should work and this does, so I guess the PODP WB/WT goes to the chipset also.
I was initially skeptical and tried connecting pin2 to VCC with a jumper wire - which also works:
vli486svgox4-5x86-jumpers.jpg
Since reading your last posts in the other thread, there's now just a regular jumper on JP18 between 1-2 and that gives write-back mode every time as well:
vli486svgox4-5x86-wb-enabled.jpg

So it seems that the only thing that needs changing from the manual page that we have (from the PC engineers reference book), to have the AMD 5x86-P75 working at full speed on the VL/I-486SVGOX4 Revision 1.2 is (along with newest SV2G bios):

Copy code to clipboard

1CPU			| JP18			| JP19			| JP20		| JP21		| JP22
2AMD-X5-133	| 1-2,3-4,5-6	| 5-6 (2x=4x)	| 4-5		| 1-2		| 2-3,4-6

It's definitely working in write-back mode, benchmarks are noticeably quicker with it enabled

Hello.

I am a bit confused about my own posting:
Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?
What picture? There is no picture attached. WTF

EDIT: Adding pictures is not possible at the moment.
But my jumper setting is: (JP18 and JP19 are 1-6, JP20 to JP21 are 1-5 only)

Copy code to clipboard

1CPU			| JP18			| JP19			| JP20		| JP21		| JP22
2AMD-X5-133	| 1-2,3-4,5-6	| 2-3,5-6       | 2-3,4-5	| 1-2		| 1-2,4-5

MORE EDIT: Typo corrected. JP22 is 1-2, 4-5.
Picture added. In case of differences, the setting in the picture counts, haha.

Lol, now the picture is shown in my original posting. Never mind.

CoffeeOne wrote on 2023-09-19, 18:42:

Hello. […]
Show full quote

Thermalwrong wrote on 2023-09-18, 01:31:

I've seen your posts in the other thread so I know you've got Write-Back working with your 1.2 board :) I dug out my VL/I-486SVG […]
Show full quote

I've seen your posts in the other thread so I know you've got Write-Back working with your 1.2 board 😀 I dug out my VL/I-486SVG0X4 today to see what I'd find since I also have a 1.2 revision and there's so little information, it must be rare - looks like you already found that JP18 pin 2 goes to the WB/WT pin: Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?

I couldn't figure out how the regulator goes from 5v to 3.4v, I couldn't see any jumpers specifically for CPU voltage or specific voltage, so they're hiding in the JP18>JP22 block. dopefish went through how the voltage regulation works while going over how to upgrade boards without voltage regulators:
Another high-end 486 back in action (ASUS GX4 conversion success! Tons of pictures)
Since my board has the voltage regulation parts, I tested out VOLDET without the CPU fitted since it was set up for the DX/2-66 I had got the board running with. Once I confirmed the voltage would go down to 3.4, I was okay with fitting one of my AM5x86-P75 CPUs to the board. I think that the voltage detection is enabled with jumper JP20 at 4-5, can't recall if there was another one but jumpering for a DX/4 and fitting a 3.45v CPU does result in the lower voltage.

From there I found Jan Steunebrink's page which gave some more information on how WB/WT works: http://www.steunebrink.info/amd5x86.htm I'm so glad this page is still active 😀
I checked out that JP18 pin 2 goes to the B13 WB/WT pin while JP18 pin 1 goes to T1 which is the PODP WB/WT pin, Jan's page confirms that connecting B13 to a chipset WB/WT pin should work and this does, so I guess the PODP WB/WT goes to the chipset also.
I was initially skeptical and tried connecting pin2 to VCC with a jumper wire - which also works:
vli486svgox4-5x86-jumpers.jpg
Since reading your last posts in the other thread, there's now just a regular jumper on JP18 between 1-2 and that gives write-back mode every time as well:
vli486svgox4-5x86-wb-enabled.jpg

So it seems that the only thing that needs changing from the manual page that we have (from the PC engineers reference book), to have the AMD 5x86-P75 working at full speed on the VL/I-486SVGOX4 Revision 1.2 is (along with newest SV2G bios):

Copy code to clipboard

1CPU			| JP18			| JP19			| JP20		| JP21		| JP22
2AMD-X5-133	| 1-2,3-4,5-6	| 5-6 (2x=4x)	| 4-5		| 1-2		| 2-3,4-6

It's definitely working in write-back mode, benchmarks are noticeably quicker with it enabled

Hello.

I am a bit confused about my own posting:
Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?
What picture? There is no picture attached. WTF

EDIT: Adding pictures is not possible at the moment.
But my jumper setting is: (JP18 and JP19 are 1-6, JP20 to JP21 are 1-5 only)

Copy code to clipboard

1CPU			| JP18			| JP19			| JP20		| JP21		| JP22
2AMD-X5-133	| 1-2,3-4,5-6	| 2-3,5-6       | 2-3,4-5	| 1-2		| 1-2,4-5

MORE EDIT: Typo corrected. JP22 is 1-2, 4-5.
Picture added. In case of differences, the setting in the picture counts, haha.

I acquired a faulty VL/I-486SVGO.
The problem was ISA D6 was open-circuit, because the trace had rotted under the battery holder.
I added the missing parts to make it the X4 as well as I had them in my parts drawers and wanted to use Am5x86-P75.
Using your (CoffeeOne) jumper settings, the system freezes after POST if L2 cache is enabled.
Using s.mouse's jumper settings for Rev1.2, L2 cache works but CPU voltage is 5V. (First photo here: Asus VL/I-486SVGOX4 Amd 5x86-133 with 30 pin simms )
Installing JP22 4-5 fixes the voltage, allowing it to autodetect.

Using the 0402.002 patched BIOS (Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?).
Using a GD5428+PDC20230 VLB combo card.

DOSBENCH Doom-max benchmark gets around 1400 realtics.
I'm using 4x4MB 60ns FPM 30-pin SIMMs. My 128MB 72-pin FPM SIMMs wouldn't POST in this board, nor would any EDO SIMMs.
Using 1MiB 15ns cache. Will make some 10ns chips with adapters if I determine that the L2 cache system is preventing a possible performance improvement.
I have another VLB VGA (Vision864) and VLB IDE (EIDE2300Plus) card coming I can try as well.

nztdm wrote on 2023-12-08, 02:24:

I acquired a faulty VL/I-486SVGO. The problem was ISA D6 was open-circuit, because the trace had rotted under the battery holder […]
Show full quote

CoffeeOne wrote on 2023-09-19, 18:42:

Hello. […]
Show full quote

Thermalwrong wrote on 2023-09-18, 01:31:

I've seen your posts in the other thread so I know you've got Write-Back working with your 1.2 board :) I dug out my VL/I-486SVG […]
Show full quote

I've seen your posts in the other thread so I know you've got Write-Back working with your 1.2 board 😀 I dug out my VL/I-486SVG0X4 today to see what I'd find since I also have a 1.2 revision and there's so little information, it must be rare - looks like you already found that JP18 pin 2 goes to the WB/WT pin: Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?

I couldn't figure out how the regulator goes from 5v to 3.4v, I couldn't see any jumpers specifically for CPU voltage or specific voltage, so they're hiding in the JP18>JP22 block. dopefish went through how the voltage regulation works while going over how to upgrade boards without voltage regulators:
Another high-end 486 back in action (ASUS GX4 conversion success! Tons of pictures)
Since my board has the voltage regulation parts, I tested out VOLDET without the CPU fitted since it was set up for the DX/2-66 I had got the board running with. Once I confirmed the voltage would go down to 3.4, I was okay with fitting one of my AM5x86-P75 CPUs to the board. I think that the voltage detection is enabled with jumper JP20 at 4-5, can't recall if there was another one but jumpering for a DX/4 and fitting a 3.45v CPU does result in the lower voltage.

From there I found Jan Steunebrink's page which gave some more information on how WB/WT works: http://www.steunebrink.info/amd5x86.htm I'm so glad this page is still active 😀
I checked out that JP18 pin 2 goes to the B13 WB/WT pin while JP18 pin 1 goes to T1 which is the PODP WB/WT pin, Jan's page confirms that connecting B13 to a chipset WB/WT pin should work and this does, so I guess the PODP WB/WT goes to the chipset also.
I was initially skeptical and tried connecting pin2 to VCC with a jumper wire - which also works:
vli486svgox4-5x86-jumpers.jpg
Since reading your last posts in the other thread, there's now just a regular jumper on JP18 between 1-2 and that gives write-back mode every time as well:
vli486svgox4-5x86-wb-enabled.jpg

So it seems that the only thing that needs changing from the manual page that we have (from the PC engineers reference book), to have the AMD 5x86-P75 working at full speed on the VL/I-486SVGOX4 Revision 1.2 is (along with newest SV2G bios):

Copy code to clipboard

1CPU			| JP18			| JP19			| JP20		| JP21		| JP22
2AMD-X5-133	| 1-2,3-4,5-6	| 5-6 (2x=4x)	| 4-5		| 1-2		| 2-3,4-6

It's definitely working in write-back mode, benchmarks are noticeably quicker with it enabled

Hello.

I am a bit confused about my own posting:
Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?
What picture? There is no picture attached. WTF

EDIT: Adding pictures is not possible at the moment.
But my jumper setting is: (JP18 and JP19 are 1-6, JP20 to JP21 are 1-5 only)

Copy code to clipboard

1CPU			| JP18			| JP19			| JP20		| JP21		| JP22
2AMD-X5-133	| 1-2,3-4,5-6	| 2-3,5-6       | 2-3,4-5	| 1-2		| 1-2,4-5

MORE EDIT: Typo corrected. JP22 is 1-2, 4-5.
Picture added. In case of differences, the setting in the picture counts, haha.

I acquired a faulty VL/I-486SVGO.
The problem was ISA D6 was open-circuit, because the trace had rotted under the battery holder.
I added the missing parts to make it the X4 as well as I had them in my parts drawers and wanted to use Am5x86-P75.
Using your (CoffeeOne) jumper settings, the system freezes after POST if L2 cache is enabled.
Using s.mouse's jumper settings for Rev1.2, L2 cache works but CPU voltage is 5V. (First photo here: Asus VL/I-486SVGOX4 Amd 5x86-133 with 30 pin simms )
Installing JP22 4-5 fixes the voltage, allowing it to autodetect.

Using the 0402.002 patched BIOS (Re: How to set Asus VL/I-486SV2GX4 (SiS471) into L2:Write-through mode?).
Using a GD5428+PDC20230 VLB combo card.

DOSBENCH Doom-max benchmark gets around 1400 realtics.
I'm using 4x4MB 60ns FPM 30-pin SIMMs. My 128MB 72-pin FPM SIMMs wouldn't POST in this board, nor would any EDO SIMMs.
Using 1MiB 15ns cache. Will make some 10ns chips with adapters if I determine that the L2 cache system is preventing a possible performance improvement.
I have another VLB VGA (Vision864) and VLB IDE (EIDE2300Plus) card coming I can try as well.

I am quite convinced that "my" jumper settings are correct.
JP22 4-5 is part of my setting.
Did you verify that the L1 cache runs in write back mode? It is all about that.
If yes (and the board runs stable of course), then keep your setting, everything is good.

Of course 128MB SIMMs are not working, because the maximum memory size of the board is 64MB.

If you want to run the board with 33MHz or 40MHz, 15ns cache rams are fast enough. you can run the cache with 2-1-2 (read 2-1-1-1, write 2) at 40MHz.
10ns could enable faster timings at 50MHz, but I am not sure.

nztdm wrote on 2023-12-08, 02:24:

My 128MB 72-pin FPM SIMMs wouldn't POST in this board, nor would any EDO SIMMs.

Depends on the board.
First try one single 128 M module in each of the slots.
It that works, try the other 128 M module in each of the remaining slots.
Try each possible combination.

Reason is, in my HOT-433 I can use 2 128 MB modules, but in rank 0,1 and 4,5 only. So the slot with rank 2,3 must be free.
However, with 256 MB RAM I require 1024 k of L2 cache and write through to have all memory cacheable.

Disruptor wrote on 2023-12-08, 19:05:

Depends on the board. First try one single 128 M module in each of the slots. It that works, try the other 128 M module in each […]
Show full quote

nztdm wrote on 2023-12-08, 02:24:

My 128MB 72-pin FPM SIMMs wouldn't POST in this board, nor would any EDO SIMMs.

Depends on the board.
First try one single 128 M module in each of the slots.
It that works, try the other 128 M module in each of the remaining slots.
Try each possible combination.

Reason is, in my HOT-433 I can use 2 128 MB modules, but in rank 0,1 and 4,5 only. So the slot with rank 2,3 must be free.
However, with 256 MB RAM I require 1024 k of L2 cache and write through to have all memory cacheable.

Yes, it depends on the board.
But the maximum amount of memory with the SIS471 chipset is 128MB. So theoretically 128MB is possible.
But the Asus board only supports 64MB.
Just go back to the beginning of this thread, where all possible memory configurations are listed.
So theorectically 2 time 64MB or 4 times 32MB would be possible according the documentation.