Calolo wrote on 2026-10-13, 17:09:

Hello […]
Show full quote

Hello

I have flashed the BIOS of my KT7A v1.1 to A9 (modded for Athlon XP) to use a mobile Barton 2400+.

I need to press reset when the PC cold boot, then the CPU is well recognized but limited to 1666MHz (12.5 x 133).

Has someone found a solution/BIOS option to make it boot directly?

Also if I strap CPU pins to enable high multipliers, how can I access the 3x for DOS slowdown?

Thank you.

Since you have a mobile Barton. You should be able to use pciset (to make setmul work on a KT133A) and setmul commands in your autoexec.bat to make the multiplier whatever the CPU will allow

My XP-M 2800+ will only recognise multipliers up to 16x, probably because it’s a late one.

Link to my thread which has a link to the post about pciset

Re: Socket 7 and Socket A shenanigans. 2 (or possibly 3) retro builds. Attempting to build 600MHz 430TX

I don’t have a kt7a but I have same chipset on a board made by Soltek. It seems to have a lot of the same quirks

I can’t help with the reset trick. The Soltek board POSTs fine with a Barton installed. I used bios patcher to enable the board to recognise the CPU, but it still posted without this patch. Just said “unknown CPU”

in my experience, modifying the socket of the CPU to change the multiplier caused more problems; the system would only post with 17x multiplier, this was too fast for any CPU I had at the time. (2266Mhz) now I have an XP-M I can get to 2350-2400MHz with 16x multiplier by increasing the FSB

FullYes wrote on 2026-10-13, 18:29:

Since you have a mobile Barton. You should be able to use pciset (to make setmul work) and setmul commands in your autoexec.bat […]
Show full quote

Calolo wrote on 2026-10-13, 17:09:

Hello […]
Show full quote

Hello

I have flashed the BIOS of my KT7A v1.1 to A9 (modded for Athlon XP) to use a mobile Barton 2400+.

I need to press reset when the PC cold boot, then the CPU is well recognized but limited to 1666MHz (12.5 x 133).

Has someone found a solution/BIOS option to make it boot directly?

Also if I strap CPU pins to enable high multipliers, how can I access the 3x for DOS slowdown?

Thank you.

Since you have a mobile Barton. You should be able to use pciset (to make setmul work) and setmul commands in your autoexec.bat to make the multiplier whatever the CPU will allow

My XP-M 2800+ will only recognise multipliers up to 16x, probably because it’s a late one.

Link to my thread which has a link to the post about pciset

Re: Socket 7 and Socket A shenanigans. 2 (or possibly 3) retro builds. Attempting to build 600MHz 430TX

I don’t have a kt7a but I have same chipset on a board made by Soltek. It seems to have a lot of the same quirks

Thanks for your reply.

I have tried the pciset command before "cpuspd m3" and now it runs @400MHz 3x133 !

So it will still work even if I strap the CPU pins to enable high multipliers?

You don’t need to ‘strap’ anything. Just try a higher multiplier with your setup as it is now. You should be able to set it to at least the highest rated for your CPU just by using software

So, mine as it is now can be set from 3x to 16x. Modification of the motherboard is not needed 😎👍

FullYes wrote on 2026-10-13, 19:49:

You don’t need to ‘strap’ anything. Just try a higher multiplier!

I have tried using pciset and cpuspd on my Barton 2400+, I can only go up to 13.5x (1800MHz), no overclocking using multiplier.

Ok. What you describe is how my cpu behaves as well - it does not let you exceed the stock multiplier with software. Since I didn’t try this CPU in my board when I modified the board, I cannot say if you will succeed or not, because I put it back to standard now due to the strange posting behaviour.

Another user; Mr.Hunt Did succeed in modifying his Soltek kt133a motherboard (same as mine) to work with his xp-m 2800+ with 17x multiplier. So perhaps it will work for you as well

The thread where we discussed this in more detail is here

Re: Socket 462 SL-75KAV weird posting behaviour after pin mod

Did you guys try to use Wcpuid to change mobile multipier on XP-M?
It worked for me on newer (KT400 or 600) mobos.

luk1999 wrote on Today, 10:37:

Did you guys try to use Wcpuid to change mobile multipier on XP-M?
It worked for me on newer (KT400 or 600) mobos.

I will look into this - thanks for the suggestion!

Edit - i am seeing some information in this thread suggesting that not being able to set a multiplier above the rated multiplier for the CPU is normal for XP-M. I am wondering if it is only possible to aim higher with an early CPU? like with Bartons/Thoroughbred CPUs made before week 38 '03 (or whenever it was)

Has someone else noticed some memory errors with a Barton CPU on KT7A (v1.1) ?

Every good RAM sticks I tried make some errors on test #6 of MemTest86 with my XP-M 2400+. In the BIOS I used "load optimized settings" only.

I don't have this problem using a Duron 700 and SDRAM @133Mhz, same BIOS configuration as the Barton. Unfortunately I don't have other Athlon XPs to try.

FullYes wrote on Today, 11:32:

Edit - i am seeing some information in this thread suggesting that not being able to set a multiplier above the rated multiplier for the CPU is normal for XP-M. I am wondering if it is only possible to aim higher with an early CPU? like with Bartons/Thoroughbred CPUs made before week 38 '03 (or whenever it was)

Max mobile multipier depends on bridges configuration (L6 I think). It was explained on FAB51 page: https://fab51.biosbude.de/cpu/barton/athlon-e23.html
If you use regular Athlon XP or Sempron and connect bridge responsible for mobile mode, then you should be able to set mobile multi in range 3x-24x.

Calolo wrote on Today, 15:03:

Has someone else noticed some memory errors with a Barton CPU on KT7A (v1.1) ?

Every good RAM sticks I tried make some errors on test #6 of MemTest86 with my XP-M 2400+. In the BIOS I used "load optimized settings" only.

Maybe try to rise slightly IO Voltage to 3.4-3.5V.

Calolo wrote on Today, 15:03:

Has someone else noticed some memory errors with a Barton CPU on KT7A (v1.1) ?

I will test and report back to you.

mockingbird wrote on Today, 20:03:

Calolo wrote on Today, 15:03:

Has someone else noticed some memory errors with a Barton CPU on KT7A (v1.1) ?

I will test and report back to you.

Thank you.

I have found that the number of errors reported by Memtest86 depends on the vCore and vIO.
I tested all the voltages between 1.45v and 1.65v for vCore and between 3.3v and 3.5v for vIO.
Increase vIO results in more errors. For vCore 1.575v is the minimum to use.
So the sweet spot is 1.575v and 3.3v, in this configuration Memtest86 report no errors!

I don't know if the problem is with my motherboard or the CPU, I will keep testing to confirm that this configuration is fully stable.