2500+ unlocking effort, strictly directions

[c627627]

There are still some possibilities not yet identified.

Knowing what you know up to this point, can you name a few ways in theory, that would make this a reversible lock.

 

[wolverine I]

1 thing i have noticed is that,i dont know if this an nvidia quirk,
but on the previously locked processors if you tried to change the multi,it would invariably not post.on these new locked processors
it simply boots at the default multi,no matter what you set it to.
on the gigabyte 7n400pro mobo at any fsb above 200 fsb if you set the multi to 10-10.5 it posts as 11 the default multi.this is with an unlocked processor.the same thing the locked cpus do.
in the original bios it posts at the 10 multi over 200fsb,if they have changed a resistance value sufficiantly to keep the multi from being implemented by the mobo or by programming this would explain the similaritys.as the bios changes in the later bios change voltages slightly and ram timings electrically all they would need to do is as youve found raise the resistance 10-20 ohms to prevent multi changes.if youve ever had the cold start problem with cpus running on extreme high voltages,where the mobo does not provide the needed voltages to sustain the overclock because the voltage signal ok to the mobo is not fast enough, 4 ns ,at a very cold start, you would know what im getting at. on the 7n400 pro under 200fsb in later bios it posts
at the 10 -10.5 multi no problem.a very small voltage change in the different bios configurations is the difference between
posting at the desired multi or not.maybe adding resistance in parallel to the L1 or L3 to halve the resistance so the voltage from the mobo would be sufficient to set the multis again.

 

[PDL]

Here's another theory...I'll probably get shot down hard on this one!

So far I have not seen that anyone with a 2600+ or higher Bart has gotten a locked one.
And if I am not mistaken the 2500+ Bart was the first Bart release!
What if....

The production process was real good and a high percentage of the cores would do much better than 2500+..and..
Since the multiplier setting for 11x is to have all of the connections set to ground, there was no need for pull-up resistors on this package. So none were included for PCB packages that were destined to be 2500+ Barts because they were very confident it would at least do 1.83 gig.

If we see 2600+ and higher ones locked either now or in the future(past week 39) then this is a mute point.

A test may be to take a locked 2500+ and cut one of the bridges and install a pull-up resistor in it's place and see if the multi changes.

Disclaimer::
I've read through this thread and others several times and pardon me if someone has already tried the above or has proven it false. Since I do not have a Bart this is the only way I can try to help!

 

[Vaguerant]

I got a AQXEA 0341 VPMW 2600+ retail box yesterday from ZipZoomFly and it's locked

 

[PDL]

Well...there goes my theory down the drain...

I think we are SOL on this one!

I'll bet AMD gets rid of all their old TBreds and Thortons now!

 

[Audioaficionado]

Hoot, how about wet sanding your PCB down layer by layer and taking digital images of all the layers including the via connections. Maybe we can develop an elecrical network for the circuitry and see how it all works.

 

[Hoot]

I did. It was a painstaking process since only one L3 line ran along the top layer. The other four dropped to lower layers using via's before they went over to their respective grid array spots. The routing was exactly as portrayed in the Fab51 diagram, with about the only vaguity being the location, if any, of the pull-up resistors, but then they could have been internal to the core, or not even needed since most gate inputs default to logic high when they float.

 

[phaedrus]

hoot, by the look of it, i suspect there are few here who could kill a cpu with better hope of finding an answer. if you have a paypal account, i'll drop you $10, thereby putting my money where my keyboard is. maybe a few more will do the same, and the pain will go away pm me with a paypal, and you'll see some $

this really has been a fine effort

cheers, phaed

 

[Virtex]

Has anyone tried this on a VIA chipset board and not just an nForce2? Don't c**p on me I'm just trying to shoot ideas out there.

 

[hitechjb1]

"No hidden pin" is not yet conclusive for the locked Barton:

- All NC pins are indeed isolated, i.e. > 40 M ohm to VSS or DMM infinity.
- Each VCC pin to VSS has small resistance in the order of 10 ohm, i.e. less than 100 ohm (putting a measurement margin).
- Each VSS pin to VSS has 0 ohm.

AFAIK, all these three conditions have not been confirmed positively.

NC, VSS, VCC pins are defined in this AMD tech doc for Barton (model 10)
http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/26237.PDF
Chapeter 11 is about the CPU pins, page 53-78.
Page 55 is the pin layout view from bottom.
There will describe all the signal pins, VCC, VSS, NC pins and where they are, ..., plus much more.

 

[LBJGH]

I was thinking about getting another 2500+ but now that AMD locked 'em there is no reason to buy one. I'll wait for the 64bit cpu's to come down in price before I spend anymore money on AMD.

Unlocking the palominos was tedious but prevented all but the hardcore o/c's from unlocking the chips. AMD would have no problem identifying a palomino that was screwed with if someone tried to RMA it.

I can understand the t-bred's and bartons being harder to nail down because they were unlocked.

AMD is gonna kill the enthusiast market for their product which will only hurt AMD in the long run. I've personally convinced 6-8 people to choose AMD over Intel but I don't know if I'd push AMD when my friends and family ask now.

Too bad AMD didn't resort to a palomino-like locking method... it would have protected their a$$es while letting the tweakers have some fun...

Just my 2cents.

 

[hitechjb1]

I was thinking about getting another 2500+ but now that AMD locked 'em there is no reason to buy one. I'll wait for the 64bit cpu's to come down in price before I spend anymore money on AMD.
...

Even for locked Barton, as long as you have a rev 2.0 nforce2 motherboard, it is still the best choice CPU for overclocking, except the lower priced Tbred B 1700+/1800+ DLT3C (default multiplier x11, x11.5), ...
Here is why.

Locked Barton 2500+ or locked 1700+ DLT3C are still the best choice in terms of CPU overclocking, FSB overclocking and price

This was written for Barton 2500+ w/ 11x multiplier, same argument would hold true for a locked Tbred B 1700+ DLT3C.

For a locked 1800+ whose multiplier is 11.5, it becomes borderly good as the locked multiplier limits the FSB overclocking by about 4% (or about 8 MHz).

As long as you have a nforce2 rev 2.0 motherboard (running at FSB 220 MHz +- 10 MHz), keep the locked Barton, its x11 multiplier works nicely with good HSF on air, from 2.31 - 2.53 GHz depending on how high the FSB and memory can be pushed.

Don't get the higher PR rating CPU, it costs more and it may make things worse, if you have a 210 or higher FSB motherboard. Same argument, don't get locked CPU with multiplier higher than 11x or 11.5x

E.g. assuming the Barton series can be overclocked to around 2.5 GHz,
if you get 2800+, if it is locked at x12.5, and your FSB would be limited to 200 MHz. Memory intensive applications and benchmarks may not perform as well as a Barton running at 2.475 GHz using 225 FSB x 11. The lower 25 MHz in FSB in the case of 2800+, is about 25 (8) (0.95) = 190 MB/s lower in effective memory bandwidth.


1700+/2500+ default multiplier 11
1800+/2600+ default multiplier 11.5
2800+ default multiplier 12.5
3000+ default multiplier 13
3200+ default multiplier 11

To summarize,
Assuming the Barton series can be overclocked to around 2.5 GHz using a high end copper HSF such as SLK-800/900/947U,
A 2500+, locked at 11, FSB would be limited to 227 MHz
A 2600+, locked at 11.5, FSB would be limited to 217 MHz
A 2800+, locked at 12.5, FSB would be limited to 200 MHz.
A 3200+, locked at 11, same as 2500+, FSB would be limited to 227 MHz.

There is a good chance to achieve 220+ FSB w/ a NF7-S rev 2.0, with winbond memory BH5/CH5, enough Vdimm and chipset Vdd.

So Barton 2500+ is still the best choice for locked CPU in terms of CPU overclocking, FSB overclocking and price.
Goal for 24/7 system:
- 2.5 GHz 512 L2 cache Barton 2500+ (SLK- HSF)
- 227 FSB x 11 (NF7-S rev 2.0)
- memory w/ winbond CH5/BH5 chips
- $90 CPU
That is, the FSB and the CPU would have a good change to be maximized out at the same time, without using the more expensive 3200+.



Impact of Barton 2500+ w/ locked muliplier on overclocking

Putting things into perspective, currently for system with NF7-S rev 2.0 with CPU on air cooling, the impact of locked 2500+ is minimal, at most 3% from best overclocking (2.5 GHz), if the motherboard and memory can do 220 MHz.

If aiming for super FSB to 240-250 MHz for 3Dmark benchmark, then the CPU is limiting the FSB due to the fixed 11x multiplier.

If motherboard or memory is limiting at the low 200-210 MHz, then the FSB is limiting the CPU to 2200-2310 MHz.


If there is no workaround for the locked Barton, these would be the scenarios:

1. It would mostly affect users with KT266A, KT333A motherboads, since those boards can achive max FSB around 150 MHz and 190 MHz respectively, due to 4:1 and 5:1 PCI lock, .... As a result, a Barton 2500+ with 11x multiplier, the max CPU overclocking frequency would be 1650 MHz and 2090 MHz respective. These number are way below the norm of 2.2 - 2.4 GHz for Barton 2500+.

2. It would also affect users using extreme coolings. Assuming they are using nforce2 which can do 220 MHz FSB on the average, and as high as 240+ if using enough Vdd, Vimm and chipset cooling, and "good" RAM, ... So the Barton 2500+ in these systems with 11x multipliers would be limited to overclocked frequency of 2420 MHz (220 FSB) to 2640 MHz (240 FSB). Not 2.7 - 2.9 GHz as hoped for. One would need a Barton with 13x multiplier to achieve 2860 MHz running 220 FSB.

3. For air cooling using a good nforce2 motherboard such as NF7-S rev 2.0, whose FSB averages around 220 MHz +- 10 MHz. The Barton 2500+ should still be able to run at 220 x 11 = 2420 MHz, which is above the norm 2.3 GHz for Barton 2500+. At FSB 230 MHz, which is doable, the Barton would then run at 2530 MHz.

Now the burden becomes putting more demand on the FSB, memory, and how to tune/optimize the FSB reaching 220 - 230 MHz, so that the motherboard, good memory (modules with WinBond CH5/BH5 chips) and FSB are not holding back the Barton 2500+ due to the fixed multiplier 11. If one can only achieve FSB 210 MHz, the Barton would be running at 2320 MHz, which is still above the PR rating of a 3200+ AMD processor.

To achieve 2.3 - 2.5 GHz speed for a 2500+ on air, a copper HSF such as SLK-800/900/947U and a high CFM adjustable fan such as TT SFII would be the choice. It is doable with a TT SFII fan running 3000-3200 rpm at which the noise should be acceptable for 24/7 run. Barton 2500+ at 2.3 - 2.5 GHz can run few degree C cooler and less power than a Tbred B 1700+/1800+ DLT3C delivering the same overall performance.

So I would say at multiplier 11x, currently with high end air cooling, for system w/ nforce2 rev 2.0 board, there is not much an impact for 24/7 usage:
- The impact is at most 12% (2200 MHz at stock FSB 200 MHz) on CPU frequency from the best Barton 2500+ overclocking on air (assuming 2500 MHz). Even in this worst case situation, it is already running as a 3200+ CPU.
- An average system at FSB 220 MHz can run Barton at 2420 MHz (about 3% off the best Barton at 2500 MHz on air).

 

[LBJGH]

... ya but I ain't gonna buy a new mobo.

 

[Hoot]

I was thinking about getting another 2500+ but now that AMD locked 'em there is no reason to buy one. -snip-

This has been said ad-nauseum, but there are probably tens of thousands of unlocked Bartons still out there to be purchased. Will you get one for $83.00? Those are luck-of-the-draw low end prices and time is the enemy as their numbers drop. Can you get one right this minute for say $99.00? Most definitely. Searches within the CPU section of pricewatch for strings containing stepping codes are growing every day as resellers realize the importance of specifying them. That is, at least, from the ones who want your business. Local stores, both "chain type" and "ma and pa types" still have them. From my searching, it seems like the AQXEA 0331's are quite prevalent.

Can you hymn and haw around trying to decide what to do? I wouldn't advise it too long, because as the number of remaining unlocked Bartons dwindle, you can bet your bottom dollar enterprising people, with money to invest, are going to start gobbling them up and sell them at a premium.

Aside from the incredibly myopic, these observations are realily apparent to anyone by simply looking around on the Web, or visiting local merchants.

Hoot

 

[emboss]

Here's a screnshot of the AMD specs with the important pins highlighted. I'm just about to start working on the unknown NC pins.

Gnah, who put a the stupid 640x480 limit on? Anyhow, I've uploaded it to:
http://www.emboss.co.nz/images/amdpins.gif

 

[hitechjb1]

Can't wait to see the measurement results.

You may use a corner pin (pin B2) for VSS, and measure every pin (NC, VSS, VCC) with respect to it.

 

[emboss]

OK, here's the "connected" NC pins, values are resistances to ground in Ohms. Measured using an analogue multimeter, so not 100% accurate (though if two values are different, they almost certainly are actually different):
A31 - 80
E33 - 80
L35 - 80
U37 - 80
H8 - 50
H10 - 50
K8 - 50
H6 - 50 (EDIT FROM 40)
J5 - 40 (EDIT FROM 50)
F8 - 50 (EDIT FROM 40)
C13 - 80
A19 - 80

edit: fixed a few mistakes in the resistances ...

 

[hitechjb1]

So all of them are from the group of UNKNOWN NC pins.


Can you also measure with the two probes reversed, and list them side by side?

Are you using the lowest setting of the ohm meter?

What is "inifinity" resistance measured in the REAL NC pins?

What do you mean by
"though if two values are different, they almost certainly are actually different".

What is the resistance between VCC and VSS measured?

 

[hitechjb1]

I measured from an unlocked 1800+ the five NC-converted-control pins (AN25, AN27, AL25, AL27, AJ27) which are BP_FID pins, the resistance to VSS is about 70 ohm.

They resembled your measurement of UNKNOWN NC pins which is 40/50/80 ohm, interesting.

From the pin map, there are six unknown NC pins F8, H6, H8, H10, K8 and also J5, from your measurements w/ 40/50 ohms, are placed very close to each other. They are candidates to look into further as control pins. They have paths to VCC and/or VSS.

Other unknown pins you found should be looked at too.

 

[emboss]

OK, repeated the measurements with the probes reversed. All came out at pretty close to 100 Ohms. Also, for a reference I did the same on a XP1700 palomino (unlocked, of course ) and the results were the same (give or take 5 Ohms, which is not surprising seeing it's a vastly different core ...), with the exception that the Palomino also had pin AM8 connected to something (80 and 100 Ohms). I'm going to check the numbers, but it looks as though no new NC pins have been connected to anything.