Shuttle Hot-541 IDT Winchip -> any chances ?

[zyga64]

Hello all!

This is my first post to this forum.

I have Shuttle Hot-541 mainboard (Intel FX chip), and last time I bought IDT Winchip to fit it's maximal performance (processor is single voltage, 200 Mhz (66Mhz x 3), has MMX extensions - just great

As I expected procesor works, and TestCPU and other Windows diagnostic programs indentifies correct type of CPU and displays CPU frequency.

So what is exactly the problem ?
BIOS doesn't knows this CPU and displays " cpu at 66Mhz".
Version of BIOS: 541wuq10 (according to information from Shuttle,
it's last version that can be flashed to "intel-flash-module" which I have).

I also look to bios image using cbrom, and noticed that old (P1) bioses doesn't have cpu microcode in separated module (like P2 and later) - so easy "exchange microcode" way is not the solution (as I think before). Probably also CPU frequency table must be changed.

And now I have big request to apple_rom (and maybe someone else who is able to do) to add IDT Winchip support to his famous patcher
(i have used it with success on many P1 and PII bioses, and no one patch fails !!!).

On Centaur technology (constructors of Winchip) homepage there are datasheets and even BIOS Writer Guide, so it's pretty good documented processor.

http://www.centtech.com/designs.html

...but I'm not-so-good programmer, so my last chance is posting to this forum. Maybe someone...

best regards
Szymon Zygmanski
zyga64@ssi.com.pl

ps. yes, I know that this bios doesn't support int13 extension but this is
not problem for me...

[BiosMan]

Hello Szymon,

That would be a job for me.
I've been patching BIOSes for added CPU support (mainly K6+) for the past few years. I will check the 541WUQ10 BIOS to see if it can be patched for the Winchip and report back.

Jan.

[zyga64]

Hello Szymon,

That would be a job for me.
I've been patching BIOSes for added CPU support (mainly K6+) for the past few years. I will check the 541WUQ10 BIOS to see if it can be patched for the Winchip and report back.

Jan

Thank you for answer !
It will be great if you can do it.
Btw, I have looked into 541WUQ10 bios image and when compare to 541WUQ13 and 541WUQ20 the decompression module "decomp_blk.bin" is only 8192 bytes long where in the later bioses length is = 12288 bytes.
This probably makes impossibility to flash (smaller boot block in intel
module ?). Am I right ? Is it possible to get "original.tmp" from newest
bios and exchange with this in 541WUQ10 ?

best regards
Szymon Zygmanski
zyga64@ssi.com.pl

[BiosMan]

Szymon,

I had some time this week to disassemble the 541WUQ10 BIOS, and I believe I can patch it for the Winchip. Can I send it to the address below your name?

Yes, I believe you are correct and that the bootblock of the Q13 and Q20 BIOSes are incompatible with the Intel flash chip.
And using original.tmp from those BIOSes won't help because they don't have any IDT support either.

Greetings,
Jan.

[zyga64]

Hello Jan

>I had some time this week to disassemble the 541WUQ10 BIOS,
>and I believe I can patch it for the Winchip.
>Can I send it to the address below your name?

Thank you for interesting my problem. I'm happy that this wasn't
hopeless, and you manage it. Email address below my name is correct.

>Yes, I believe you are correct and that the bootblock of the Q13 and Q20 BIOSes are
>incompatible with the Intel flash chip.
>And using original.tmp from those BIOSes won't help because they don't have any IDT
>support either.

But generally changing 'original.tmp' (in example by AWDBedit
http://awdbedit.parodius.com/) should work ? I'll check It

And Yes, I know that Q20 and Q13 bioses doesn't include support
for IDT (finally I found MX chip in old broken Pentium mainboard,
then flashed Q20 bios and... ...this bring me problems with serial ports
So I returned to old, good Q10 revision.

What is exactly the problem ? Except 2 onboard serial ports, I also have
2 additional cards: ISA (PnP) modem, and old ISA multi I/O card.
There weren't problems so far even though Com1 share IRQ with Com4,
and Com2 with Com3. After update I have seen "conflict in resources".
It's possible to choose IRQ from 3/4/5/7 for modem, so I set it to 5, but there
was still conflict between 2 ports. I need additional ports as it's my service
computer and various interfaces like eprom and AVR programmer
must be connected at the same time (well, not 'must be', but it's very
comfortable, and I'm used to this).
So I give up, as didn't noticed any positives in new BIOS.

Now about IDT support. I belive that it must contain: displaying
correct processor name and frequency, turning on write combining support
when detected, and turning on Advanced Power Management (after
change from Pentium-S to Winchip, APM option in BIOS becomes inactive,
also in Windows). Is it possible to make all this things ?

In refer to write combining support: I took sample source code from
datasheet and compiled to .com file, but didn't compare differences
in performance yet (I intend to place it in autoexec.bat at the beginning
- will it work in that way ?).
But in the meantime I have compared performance of Pentium 133, and Winchip 200.
Low efficiency of Winchip suprised me, 133 Mhz pentium beats Winchip
in Dhrystone ALU, and Whetstone FPU tests. Winchip wins only Multimedia
(MMX) test as it has MMX unit which Pentium-S doesn't.
I tested this by SiSoft Sandra 2001.
But instead of pure numbers I noticed increase of smoothness
in Windows environment (hehe - name says "WinChip"). As I wrote before,
I'm not intend to transform this to so called "multimedia machine",
i treat this computer like service tool (nice if this tool
could be more useful anyway).

regards
Szymon

[BiosMan]

Szymon,

Patched BIOS sent...
Please let us know how it works.

Jan.

[zyga64]

Thank you, very much Jan !!! Sorry for long delay in answer.

I have reflashed Bios chip with your patched version, and no suprises
Everything works very well. Procesor invites itself as 'IDT Winchip-S"
at 200 Mhz. That was exactly I want to see !
Also Power Management works, now I can change settings for APM, and
in windows "device manager" there's no problems at all. Great job.

Your patch is great and now is tested, works perfectly.
Everything what you do so far for me, satisfy me completely.
But in free time you may look into write-combining support.
I'll be even more gratefull for it.

greets from Poland
Szymon Zygmanski

ps. I was wanted to thank you personally, but I've got:
"Remote host said: 553 5.3.0 <y.ourname@net.HCC.nl>
... Blocked transfer for this recipient"