dj_pirtu wrote on 2022-02-26, 08:30:

But, if you want that ISA slot... And I do because of Gravis Ultrasound. But all the later VIA-chipsets are really good, 100% SB PRO support in DOS with ESS SOLO-1.

Hey, since you had luck with PCISET before, maybe you can figure out why this won't work with the SOLO-1:

pciset 1969 125D 52 00000001 00000010

This is supposed to toggle SB-LINK to on (change register 52), but it just returns a "0" result... If I could get that to work I would replace my YMF724 with it on my Northwood system... The SB-LINK on the YMF724 works just fine, but I would much rather have the SOLO because of the internal wavetable header and full ADPCM support.

I'm off now to test SETMUL with your trick, will report back shortly.

cde wrote on 2022-02-25, 17:51:

I have modded the AXMH2400FQQ4C successfully (CrystalCPUID reports the original multiplier as 18x) however even after changing register 55h from 89 to 8D I still get a system hang when changing the multiplier or voltage in CrystalCPUID (version 2.0). Same behaviour in DOS when changing multiplier after using PCISET. Any idea? (what BIOS version and CPU voltage do you use, btw)?

@cde

It worked great for me. Using the BIOS I linked to earlier and my KT7A 1.1 board, I was able to go as low as 133 x 3 (400Mhz, since 3x is the lowest multiplier allowed. Obviously, a lower FSB will yield even lower results. I would assume 66 x 3 or 200Mhz would be the lowest allowable clockspeed, but I don't know if you can change FSB on the fly, but there are other mechanisms to reduce speed along with a multiplier reduction, such as disabling caches)...

This is fantastic. Goodbye K6-II/K6-III/K6-III+ systems...

Has anyone seen how much those Aladdin and Apollo motherboards have been fetching recently?

Hi,
Just to make things clear: The requirement to set reg. 55, bit 2 in case of KT133A is known for a long time. Gerwin is also aware of it. Just a few posts where we discussed it on SetMul's topic:
Re: SetMul - Multiplier control for VIA C3 / AMD K6 Mobile / Cyrix 5x86
Re: SetMul - Multiplier control for VIA C3 / AMD K6+7+8 Mobile / Cyrix 5x86
Re: SetMul - Multiplier control for VIA C3 / AMD K6+7+8 Mobile / Cyrix 5x86

The most likely reason why it is not part of Setmul is because these settings are very chipset specific. Every chipset requires different bits to be set:
https://web.archive.org/web/20170822100906/ht … html/Tweaks.htm

Well, unfortunately THROTTLE.EXE doesn't work very well at all in conjunction with SETMUL.... No matter what I set the multiplier at, THROTTLE severely decreases performance to a 386 no matter what I pick from the list...

Can someone else confirm please? I'll give CpuSpd a shot as well, since it also has multiplier control built in.

mockingbird wrote on 2022-02-27, 03:56:

Well, unfortunately THROTTLE.EXE doesn't work very well at all in conjunction with SETMUL.... No matter what I set the multiplier at, THROTTLE severely decreases performance to a 386 no matter what I pick from the list...

Can someone else confirm please? I'll give CpuSpd a shot as well, since it also has multiplier control built in.

THROTTLE by default always disables L1 cache, too. You have to tell it explicitly no to do it: give it the -c command line parameter.

Falcosoft wrote on 2022-02-27, 02:07:

Hi, Just to make things clear: The requirement to set reg. 55, bit 2 in case of KT133A is known for a long time. Gerwin is also […]
Show full quote

Hi,
Just to make things clear: The requirement to set reg. 55, bit 2 in case of KT133A is known for a long time. Gerwin is also aware of it. Just a few posts where we discussed it on SetMul's topic:
Re: SetMul - Multiplier control for VIA C3 / AMD K6 Mobile / Cyrix 5x86
Re: SetMul - Multiplier control for VIA C3 / AMD K6+7+8 Mobile / Cyrix 5x86
Re: SetMul - Multiplier control for VIA C3 / AMD K6+7+8 Mobile / Cyrix 5x86

The most likely reason why it is not part of Setmul is because these settings are very chipset specific. Every chipset requires different bits to be set:
https://web.archive.org/web/20170822100906/ht … html/Tweaks.htm

ok, this was very helpful since i can realize that this trick could be applied to every mobo as long as it's based on a specifical chipset, being so the case of an abit kt7a or an epox 8kta3 ( each one kt133a )

cde wrote on 2022-02-26, 08:36:

I have tried the ESS Solo-1 in a much newer system (MSI MS-7253 / K9VGM-V (KM890, VIA8237A) , Athlon 64 X2 5050e, Radeon X800 XL) that still retains DOS compatibility and can be slowed down to 386 levels. It's a really nice card coupled with the X2GS or WP32 McCake.

I have Asus A8V Deluxe with Athlon64 X2 @2,8GHz, display cards are Radeon 3850 AGP AND Voodoo3 PCI. Sound Blaster X-Fi and ESS SOLO-1 for the sound. It can do DOS-games really well, Glide games in Windows98SE and yes, it runs Crysis 😁

It has triple-boot, DOS/Win98SE, WinXP and just for the kicks: Windows 10.

That's a really nice setup 😀 How well does Windows 10 work though? In my experience it is slow as hell on the Athlon 64. All other OSes work just fine, including the latest Debian/Ubuntu.

Falcosoft wrote on 2022-02-27, 08:06:

THROTTLE by default always disables L1 cache, too. You have to tell it explicitly no to do it: give it the -c command line parameter.

Thanks! That did the trick... The flexibility is amazing now... Multiplier reduction plus 16 throttle levels with THROTTLE...

I played around abit more with the SOLO... the PCI register shift for register 52 with command pciset 1969 125D 52 00000001 00000010 does in fact seem to work, but when running ESSOLO.COM, it shows it as "TDMA" and the TSR is loaded, but that's on a different system so not really relevant to this thread... ESSOLO.COM shows "VIA DDMA" on the KT7A though FWIW.

I wonder if the SB-LINK implementation on the Solo actually works.

cde wrote on 2022-02-27, 11:42:

That's a really nice setup 😀 How well does Windows 10 work though? In my experience it is slow as hell on the Athlon 64. All other OSes work just fine, including the latest Debian/Ubuntu.

It's usable but I can't get cache to work on VIA SATA-controller so disk access is slow and only 2GB RAM so not so fast.

Damn, my tbred 2600+ (AXDA2600DKV3C) seems to be locked ( i cannot change multiplier neither from bios nor via software ) to my epox 8kta3 pro+...
So i tried a mobile barton 2600+ (AXMG2600FQQ4C) that came factory unlocked, but as i feared, even with a 05-2003 bios release, i've got troubles: i've tried so far 5x and 13x ( which is automatically set to 12,5x ) and win98 seems to work, but when i try to install winxp, the system inexplicably keeps rebooting just at 6% of copying files process )

Now, being my epox apparently not working with xp-m, which desktop t-bred model should i get in order to pick an unlocked one?
NB: as you could understand, i wouldn't mess around with L bridges, ecc...

Kaisersoze wrote on 2022-03-12, 17:48:

Damn, my tbred 2600+ (AXDA2600DKV3C) seems to be locked ( i cannot change multiplier neither from bios nor via software ) to my […]
Show full quote

Damn, my tbred 2600+ (AXDA2600DKV3C) seems to be locked ( i cannot change multiplier neither from bios nor via software ) to my epox 8kta3 pro+...
So i tried a mobile barton 2600+ (AXMG2600FQQ4C) that came factory unlocked, but as i feared, even with a 05-2003 bios release, i've got troubles: i've tried so far 5x and 13x ( which is automatically set to 12,5x ) and win98 seems to work, but when i try to install winxp, the system inexplicably keeps rebooting just at 6% of copying files process )

Now, being my epox apparently not working with xp-m, which desktop t-bred model should i get in order to pick an unlocked one?
NB: as you could understand, i wouldn't mess around with L bridges, ecc...

Have you tried lowering the voltage? In my case, Windows XP freezes too at a random point (the higher the voltage, the sooner). At 1.35v it does succeed in the initial file copy phase. I believe this could be caused by either a defective component on the MB or a underpowered PSU (mine is modern and can only output 24A on the 5V rail).

cde wrote on 2022-03-12, 18:06:

Kaisersoze wrote on 2022-03-12, 17:48:

Damn, my tbred 2600+ (AXDA2600DKV3C) seems to be locked ( i cannot change multiplier neither from bios nor via software ) to my […]
Show full quote

Damn, my tbred 2600+ (AXDA2600DKV3C) seems to be locked ( i cannot change multiplier neither from bios nor via software ) to my epox 8kta3 pro+...
So i tried a mobile barton 2600+ (AXMG2600FQQ4C) that came factory unlocked, but as i feared, even with a 05-2003 bios release, i've got troubles: i've tried so far 5x and 13x ( which is automatically set to 12,5x ) and win98 seems to work, but when i try to install winxp, the system inexplicably keeps rebooting just at 6% of copying files process )

Now, being my epox apparently not working with xp-m, which desktop t-bred model should i get in order to pick an unlocked one?
NB: as you could understand, i wouldn't mess around with L bridges, ecc...

Have you tried lowering the voltage? In my case, Windows XP freezes too at a random point (the higher the voltage, the sooner). At 1.35v it does succeed in the initial file copy phase. I believe this could be caused by either a defective component on the MB or a underpowered PSU (mine is modern and can only output 24A on the 5V rail).

The only thing i can try is lowering the V from the bios by 0.1 which is the max negative that the bios allows. I'll give it a chance, but i'm quite pessimistic about that, expecially considering the luck i've got so far with this stuff 😒

Anyway, if i turn on a desktop cpu, can you confirm that is the L3 5# bridge that should be closed in order to be multiplier unlocked?

Unfortunately I only have the mobile version, so I'm unsure how to confirm that 😒

cde wrote on 2022-03-13, 04:01:

Unfortunately I only have the mobile version, so I'm unsure how to confirm that 😒

Nope. I'm going to test a t-bred 2200+ desktop which seems unlocked ( L3 #5 bridge closed ). Anyway, if unsuccessful, i've a PII klamath 300mhz ( 440BX ) which i'd use for DOS/win98 speed sensitive games 😉

Got myself another KT7A, this time v.1.3. Maybe three caps are bulging a little, need to replace all some day.

But it has boot problems. Now it's 133FSB and 3,5V, if I raise FSB a bit it won't boot anymore. No picture. Reset or power button won't help. Only if I boot it up from PSU swith and pressing INS at the same time so it resets CPU settings.

BUT, if I can get it to windows I can set it up to 145FSB with setfsb and it's stable, no problems at all. Memory and AGP settings maxed out (but fast writes off).
My previous KT7A got only 140FSB stable.

Maybe it's the caps... CPU is Athlon XP-M 2400+ (Barton).

dj_pirtu wrote on 2022-03-19, 07:31:

Got myself another KT7A, this time v.1.3. Maybe three caps are bulging a little, need to replace all some day. […]
Show full quote

Got myself another KT7A, this time v.1.3. Maybe three caps are bulging a little, need to replace all some day.

But it has boot problems. Now it's 133FSB and 3,5V, if I raise FSB a bit it won't boot anymore. No picture. Reset or power button won't help. Only if I boot it up from PSU swith and pressing INS at the same time so it resets CPU settings.

BUT, if I can get it to windows I can set it up to 145FSB with setfsb and it's stable, no problems at all. Memory and AGP settings maxed out (but fast writes off).
My previous KT7A got only 140FSB stable.

Maybe it's the caps... CPU is Athlon XP-M 2400+ (Barton).

These kt133a's are a real p.i.t.a.
Anyway, before messing around with caps, try even a PSU switch ( a strong 5v line, starting from 30A and above ). Indeed, a recap is mandatory.

Falcosoft wrote on 2022-02-10, 08:27:

Repo Man11 wrote on 2022-02-10, 05:36:

Maybe this is becoming commonplace as these boards age? Another person is having an identical problem: ABIT KT7A 1.3 rev and Agp slot ( apparently ) not working...

It seems identical person is having identical problem 😀 (Kaisersoze)

Waaaaait. cde and kaisersoze are the same person? 😮
https://en.wikipedia.org/wiki/Sock_puppet_account ?

Tetrium wrote on 2022-03-21, 10:21:

Falcosoft wrote on 2022-02-10, 08:27:

Repo Man11 wrote on 2022-02-10, 05:36:

Maybe this is becoming commonplace as these boards age? Another person is having an identical problem: ABIT KT7A 1.3 rev and Agp slot ( apparently ) not working...

It seems identical person is having identical problem 😀 (Kaisersoze)

Waaaaait. cde and kaisersoze are the same person? 😮
https://en.wikipedia.org/wiki/Sock_puppet_account ?

No. Repo Man11 referenced another topic as ' another person is having the same problem' but that another person was the same person namely Kaisersoze.

Kaisersoze wrote on 2022-03-21, 09:56:

dj_pirtu wrote on 2022-03-19, 07:31:

Got myself another KT7A, this time v.1.3. Maybe three caps are bulging a little, need to replace all some day. […]
Show full quote

Got myself another KT7A, this time v.1.3. Maybe three caps are bulging a little, need to replace all some day.

But it has boot problems. Now it's 133FSB and 3,5V, if I raise FSB a bit it won't boot anymore. No picture. Reset or power button won't help. Only if I boot it up from PSU swith and pressing INS at the same time so it resets CPU settings.

BUT, if I can get it to windows I can set it up to 145FSB with setfsb and it's stable, no problems at all. Memory and AGP settings maxed out (but fast writes off).
My previous KT7A got only 140FSB stable.

Maybe it's the caps... CPU is Athlon XP-M 2400+ (Barton).

These kt133a's are a real p.i.t.a.
Anyway, before messing around with caps, try even a PSU switch ( a strong 5v line, starting from 30A and above ). Indeed, a recap is mandatory.

Did some googling and seems like this problem is quite common with these mobos. It won't post sometimes when doing restart from windows or ctrl+alt+del but after pressing reset-button it comes to life. And yes, hwmonitor says that my +5V and +12V are both under desired and with bad caps...