• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

xp-m's running different multi

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

ihrsetrdr

Señor Senior Member
Joined
May 17, 2005
Location
High Desert, Calif.
I've got a couple 2500 xp-m's with the L5 mod running in a Gigabyte GA-7DPXDW motherboard, and have used wpcre to change the binary and use CrystalCPUID build 4.4.8.261 to change the multi's. CPU1 changes o.k. to 16x on up, but CPU2 won't take any multi higher than 14x which is its default. I re-applied conductive paint to the L5 bridge, but it still defaults to 14x. I doubt if I have the hand-eye coordination to do a wire mod, and frankly don't feel like tearing this rig apart again.

I backed down CPU1 to 14x to match CPU2...however, does it really matter if the processors are running different multipliers? The SOLE purpose of this machine is to fold 2 instances of fAh.
 
ihrsetrdr said:
I've got a couple 2500 xp-m's with the L5 mod running in a Gigabyte GA-7DPXDW motherboard, and have used wpcre to change the binary and use CrystalCPUID build 4.4.8.261 to change the multi's. CPU1 changes o.k. to 16x on up, but CPU2 won't take any multi higher than 14x which is its default. I re-applied conductive paint to the L5 bridge, but it still defaults to 14x. I doubt if I have the hand-eye coordination to do a wire mod, and frankly don't feel like tearing this rig apart again.

I backed down CPU1 to 14x to match CPU2...however, does it really matter if the processors are running different multipliers? The SOLE purpose of this machine is to fold 2 instances of fAh.
Isnt there 2 L5's you have to change? One is the MP bridge and the other makes it a mobile. I could be wrong though.
 
Since mobile CPUs have both the "mobile" and "SMP capable" bridges joined you do not need to do any modifications to make them work in the Gigabyte. I assume then that the bridge you are talking about painting is the one remaining uncut brige, and joining that sets the maximum software multiplier to 24x. Without rejoining this bridge, you will be limited to the default maximum (14x) for PowerNow transitions.

So it sounds like you need to redo the L5 bridge paint on the chip that does not want to accept higher multipliers.

Alternatively, painting the pins on the bottom of the chip is much easier than painting the bridges, so you might as well just do that to set the startup multiplier to 16x or whatever you want.

As a side note, Windows does not properly support PowerNow transitions on K7 SMP motherboards and you'll get wierd effects and possible stability problems should you try to do so. Linux can be configured to work correctly, though most "out of the box" kernels from the common distros will not be correct. Similarly, Windows will generlly take a dump on you if you run with different speed CPUs, though you can make Linux accept it (more or less). In any case, it's definately not recommended and can cause a large number of annoying stability problems.
 
emboss said:
Alternatively, painting the pins on the bottom of the chip is much easier than painting the bridges, so you might as well just do that to set the startup multiplier to 16x or whatever you want.

Painting the pins? I've only heard of doing a pin mod with a wire loop. If painting the pins is easier than painting the bridges, I'd consider redoing the processor. Would you have a link for the pin painting procedure?
 
Thanks for the link...I shall definately try it out! ;)


EDIT:
Forgot to ask- will just alcohol remove the paint, or need something like acetone?
 
Back