Controlling vCore on nForce2 (was limited to barton unlock)

rubber_boat · Post by **rubber_boat** » Mon Jan 12, 2009 1:33 am

Edit: I put the followup at the bottom. No use in starting a new thread.

Hi,

Tried my luck and I admit I am not too good at searching these forums. I am not a software person at all.

Need some help. I don't make any money from running my computer system: I only read websites and basic stuff.

Had a Shuttle AN35 Ultra 400 (NF2 chipset), and a week 43 or so Athlon XP Barton. The Barton was 'super-locked', and probably if you are reading this post and familiar with the forum you will know better than I do about it.

I bought a used [ASUS N7-something] Correction: MSI K7N2 Delta2 fancy MB, and a slightly older AQXCA0310MPMW Athlon XP Barton, which could be multiplier adjusted.

From my humble perspective, there is a lot of confusing information on how to unlock this processor. Including mention of having it run from 5.5 to 12.5 multiplier, or 13 and higher multiplier. Other limitations are mentioned.

I can cut, or alternatively, bridge the pins. Physically doing the work is a hardware issue and not much of a trouble for me. I would simply like to know how to 'un-&%/' the processor from what AMD did to disable useful options. Most of the time to underclock it. I -do not- want to do on-the-fly FSB changes, only multiplier changes and voltage changes. I had some bad corruption problems of my software raid on my Shuttle/Superlocked combination spread out over time I think due to FSB changes.

Now I've a 3ware 9500B for real raid, in the 'new' [ASUS] MSI kit. The Shuttle stuff goes in an assembled system to my mother-in-law to replace her P3-700. A nice boost.

I am headed for minimum power consumption. With FSB regulation of the Shuttle, the power consumption was down to about 80 watts at the 'idle' setting with a FSB of 100mhz. The multiplier was fixed around 11. The PC has an 80plus power supply. We pay about 20 Euro cents per kilowatt hour. I have my computer, plus a DSL router and bridge, and an APC UPS, plugged into an outlet through a power meter to check consumption changes immediately.

Probably TMI. If there is anything left out of the information above, write it as a PM or reply; I'll fix it.

-- Edits: I bought an MSI board, not ASUS, to replace the Shuttle.
I tried to bid I think several times on several ASUS boards and was not a good enough ebay bidder, so for some reason I switched to an MSI. Sorry for any confusion; the 'new' parts were sitting in a box for a month while I waited the 3ware 9500B and time to do this switch.

psiu · Post by **psiu** » Mon Jan 12, 2009 4:03 am

Going off memory here, but basically pre-week 35 processors should be unlocked. There is nothing you need to do physically to change them, the motherboard will do it. Weeks 36-43 *may* be unlocked.

On older (Thunderbird, I think Palomino cores also) models there were different bridges visible you could cut or connect and also you could fiddle with the pins. By Thoroughbred and Barton days it was handled internally.

Which specific motherboard? A7N8X? What operating system do you use? If XP, then check out either S2KCtl or 8rdavcore. 8rdavcore will do more if it works with your motherboard, S2KCtl simply uses the halt/interrupt command to drastically lower temps (like 10C) at idle. Which should be saving energy right?

You should be able to determine if your multiplier adjustments are having any effect pretty easily, leave your FSB where it is and then change multiplier, save, and reboot. If the speed has changed you're good.

Hope this helps!

edit: actually, if you could try S2KCtl through the power meter and compare idle numbers versus stock idle numbers, that would be great. I've always wondered what the power savings would be. It only affects idle performance, so performance when you need it is unchanged--you don't need to drop the FSB down to accomplish the idle changes.

lm · Post by lm » Mon Jan 12, 2009 4:40 am

I did a physical voltmod to my barton core athlonXP 2800+ several years back, to allow lower voltages. My mobo back then (Asus A7N8X Deluxe Gold) did not allow me to undervolt at all, but it did allow overvolting, so I had to change the voltage ID of the cpu by cutting one bridge on the pcb surrounding the die itself. This lowered the voltage ID to some low number, and then I could "overvolt" from bios to get the voltage I wanted. However I set the ID so low that even with max "overvoltage" I could not get the default voltage of the cpu anymore. Stupid limited bios. Obviously the mobo was capable of providing just about any voltage, but the lower and upper limits set by software were just braindead.

The voltmod itself was simple to do, and there were nice resources on the net about it, but I can't remember any details of it anymore.

But I am pretty sure that I even saw a configurator that hilighted parts of an image of a barton to show what you needed to cut or connect depending on what you wanted to do.

rubber_boat · Post by **rubber_boat** » Mon Jan 12, 2009 5:17 am

psiu wrote:Going off memory here, but basically pre-week 35 processors should be unlocked. There is nothing you need to do physically to change them, the motherboard will do it. Weeks 36-43 *may* be unlocked.

lm wrote:... online configurator ...

Thanks. I'll have to look into the 'Friend of SPCR'.

The online configurator may be here:
http://www.ocinside.de/go_e.html?http:/ ... kshop.html

I was wrong about the 'new' board. It is an MSI K7N2 Delta2 MS-6570E v1.0

Here is a photo of the CPU. It is a scan with a Canon N650. Some scanners have better depth of field: this one does not. The traces and cuts are visible. It would be really, very nice to have the cuts/jumpers done before assembling the entire PC. Everything is apart (2 PCs) in my tiny kitchen at the moment.

http://img133.imageshack.us/img133/3671 ... tonkt2.jpg

Mats · Post by **Mats** » Mon Jan 12, 2009 5:22 am

IIRC, changing the FSB in Windows worked on some nVidia boards,
and changing the multiplier in Windows worked on some VIA boards.
Not the other way around like you want it.

psiu · Post by **psiu** » Mon Jan 12, 2009 7:10 am

I never got into it at all--my main chip was a Tbred-B 1800+ JIXIB stepping on an Asus A7N8X board. Ran at 2600 speeds without a hiccup. Have a locked 2600 Barton and had a locked 2200 Thorton.

rubber_boat · Post by **rubber_boat** » Mon Jan 12, 2009 7:33 am

http://www.ocinside.de/go_e.html?http:/ ... nting.html

Seems to show cutting the first bridge of the L5, when selecting 'Unlock'.

Alternatively, http://www.ocinside.de/go_e.html?http:/ ... inmod.html
shows tying a certain pair of pins together (select AMD Barton from the second pulldown, and Unlock multiplier from the fourth pull-down).

What do you all think?

P.S. I threw 10 bucks into SPCR. Don't have the 25 for a full friend membership now.

Mats · Post by **Mats** » Mon Jan 12, 2009 7:42 am

Useful links from good old fab51!

http://fab51.com/cpu/barton/athlon-e24.html

http://fab51.com/cpu/barton/athlon-e23.html

rubber_boat · Post by **rubber_boat** » Mon Jan 12, 2009 10:53 am

Mats thank you.

Do I only need to close L5 bridge 2 then, and set the processor to Mobile mode, to make the processor 'flexible' again?

PS I also have a third board which is a VIA chipset I can test. In any case, I want to get the processor(s) correctly setup.

But AMD forgot (thankfully) to lock the L5 bridges which lets us change our cpu to a mobile XP or a Athlon with MP support.

Luckily, one of OCW members, Petr, has made a software that allows chipsets that support mobile AMDs to change multiplier in windows...

rubber_boat · Post by **rubber_boat** » Mon Jan 12, 2009 11:03 am

It looks like http://www.cpuheat.wz.cz/html/AXP_multi ... iplier.htm
says I must also cut/connect bridges on L11 to set the Startup Voltage.

psiu · Post by **psiu** » Mon Jan 12, 2009 4:46 pm

If the Barton is old enough, just pop it in and see what you get??
Build it out of the case with just the minimum to see what you get. Run your FSB up, try setting CPU voltage low, and leave at stock multiplier to start. See what your idle power draw is, download S2KCtl and run it, see what the power draw does.

link to S2KCtl

Just saying you should see what have first before tinkering too much.

rubber_boat · Post by **rubber_boat** » Wed Jan 21, 2009 4:15 pm

psui

b4 I saw u said that I patched L5 Mobile and cut all of L6.

CPU shows as a mobile. For kicks. It's in a nForce2 MSI K7N2 Delta2.

Here is the setup and followup question

It WAS an:
Athlon XP Barton 2600+ 0343 in a Shuttle AN35 Ultra 400.
Can change the FSB, vCore, other values with 8rdavcore on the fly.

It NOW is:
MSI K7N2 Delta2 Platinum and Barton 2500+ 0310.
The FSB is adjustable, but the vCore is not!! ?

No vCore slider available in nTune and 8core error log indicates some problem, mentions selecting a driver (not mentioning GiveIO.sys, GiveIO initializes okay). XP is exactly the same as b4 plus a couple of drivers added to support the 3ware (the box also went from software raid to hw raid). I installed a 'rescue' copy of XP on another partition just to check the vCore problem. Same results. Just for kicks, I swapped out the 2500+ 0310 and put in the 2600+ 0343. No difference, no vCore adjustment available.

Is the MSI K7N2 vcore not adjustable on the fly?

Thanks.

mattthemuppet · Post by **mattthemuppet** » Wed Jan 21, 2009 7:36 pm

long time since I did this and I no longer have the chip or the pictures, but my ancient post here has some good info. Basically there's a pit where one of the contacts is cut (by a laser I think) that locks the multiplier in any direction. So to unlock the chip, you need to get a fine scalpel or screwdriver and scrape each end of the pit (but not the bottom). Then, once you've exposed the trace at either end, paint some conductive silver between the 2 traces.

Ta da, you now have an XP Mobile chip. Never tried increasing the multiplier, but it went down to 5x (I think) quite happily. Unfortunately the NForce2 board I ended up with (Abit NF7-S) couldn't dynamically adjust the multiplier (this applies to ALL NF2 boards) though it did undervolt well. The Gigabyte Via board I had before that did adjust the multiplier but wouldn't let me undervolt either the CPU OR the fans, which was why I ditched it.

If you can't find any conductive silver let me know your address and I'll post you some - cost me a fair bit for about a million times what I needed, so I have plenty spare

Good luck!

~El~Jefe~ · Post by **~El~Jefe~** » Thu Jan 22, 2009 7:24 am

I think someone should donate an old 939 or 754 to this dood

scalpeling chips ............

rubber_boat · Post by **rubber_boat** » Sat Jan 24, 2009 10:47 am

Hmm. Thanks.

Now there are many different versions of Bios collected in the FLASH folder.

I've plenty of the silver paint to repair windshields.

Updated the board to Bosskiller's B72c, and could connect to ATXP1 and control vCore - once. After going into the bios and swapping the CPU, could not anymore.

Now it's running MSI's stock (but not available from MSI's website) B71. Also named 11.71

Going to try and flash the Shuttle AN35 Ultra 400 bios on it next. Just for giggles. Someone did this somewhere else...

Gojira-X · Post by **Gojira-X** » Sat Jan 24, 2009 2:45 pm

Hey rubber_boat,
I donno if this helps but I still have a load of the Speed Strip - Athlon XP unlocker strips featured in this Overclocker Cafe review.

I got them for use on the AXP 2600+ (T'Bred B) in my sig, with a couple to sell on. In the end I kinda missed the boat and have a load of Speed strips that I can't really do much with.

If you are interested, I can send the whole lot (for a small price to cover postage).

rubber_boat · Post by **rubber_boat** » Wed Jan 28, 2009 12:06 pm

bangers, I cannot control vCore. It's always greyed out, except for one time after flashing 11.72. Does not matter if I used the Insert key to reset cmos or jumper or whatever.

I now have official bios's:

11.3
11.5
11.6
11.71
11.72 (although this is a modified version from bosskiller of 11.5)

The last two official bios's are not available on MSI's website.
The log of 8rdavcore says 'SMI Not responding' several times.
That is not an exact quote because neither the machine nor the log is handy at the moment. I do not have any log from when 8rdavcore was working; it overwrites any existing log when it starts up.

vCore manipulation worked on the Shuttle AN35 Ultra 400. All the parts on the new rig are moved over, plus a 3ware card which should not make any difference. CPU and RAM are the same.

Does anyone have an MSI K7N2 Delta2 rig working they could run 8rdavcore or even nTune on it and check if they can control the vCore?

Can you post your 8rdavcore log?

Is there another bios from a similar nVidia reference design spec board, like the Epox 8rda6+ (I think I got that model right)? I found some link to the Epox bios from 2006, but their ftp server is down. Then I found out some offices closed up awhile ago.

I do not need a lot of the hardware working at boot, and if this is all the BIOS does, the core functions of controlling vCore are more important to me.

Thanks thus far.

rubber_boat · Post by **rubber_boat** » Tue Feb 03, 2009 6:33 am

Has anyone see this: http://sandeen.net/fwfix/index.html ?

"How to flash a PLCC, using an Intel Pro 100 network card with a PLCC socket as a flash reader/writer."

I am ordering one for a couple of bucks. Should be here in about 4 days.

I was trying to flash the 8rda3+pro/6+Pro bios onto the 512k Pm49FL004T-33JC PLCC in the MSI board. It currently has the MSI K7N2 Delta2 bios on it. The MSI bios is an AWARD bios image; the 8rda3+Pro seems to be an AMI BIOS. The Award flasher does not want to flash it, the AMI flashers I have used do not want to, and UNIFLASH 1.4 (latest) doesn't like the whole mainboard in general.

For testing, I took the 256k W49V002AP PLCC out of the Shuttle AN35 Ultra 400 board and it boots in the K7N2 Delta2. I can get access to the USB drive via a DOS boot CDROM I constructed. All the flash images and flashers on are the USB drive. So at least I know if the 8rda3+Pro failed, it's possible to boot the mainboard and flash back; but I can't flash-forward yet.

Because I had so many problems with trying to cross-flash, I think I could brute-force it by using a PLCC flasher, and the Intel card would seem to do this with fboot.exe, much cheaper than a dedicated PLCC flasher.

Controlling vCore on nForce2 (was limited to barton unlock)

Controlling vCore on nForce2 (was limited to barton unlock)

I'm getting mixed messages.

tinkering

Re: vcore probs

vcore

more info