Unlocked Venice 3000+

I've just been looking at a dead 3000 venice (donated to me by BazX). The components on the top of the CPU PCB, which have been assumed to be capacitors in the past appear to possibly be some sort of bridge. I've only done some very brief tests on them but they appear to have some sort of internal connection.

Does anyone want to investige the components on the top of the CPU? I would do it myself, but I've only got one (dead) A64 CPU here.

Edit: After some further investigation, it looks likey they are just supply smothing capacitors between Vcore and ground :(

subscribed to this thread!

you could always sent it to me to analyze.. I got the tools...

Maybe a resistor pack, instead of bridges it's like high resistance is off and low is on and a combination gives us our values for multis and vcore and such?

i must say, this is the most interesting yet funniest threads i have ever read. All the technical stuff is way over my head, but i just wanna give you guys the thumbs up for being pioneers! And good god man, dont sell that chip on ebay!

well i went through each of the SBZ bits in C0010015 with hwdirect (thanks for that btw). i wrote each one to 1, but none of them had any affect on the writeability(sp?) of the bits in C0010041 or C0010042. :( i did notice something strange about bit 5 in C0010015, though. in amd's guide, it isn't grayed out in the register diagram like the other reserved bits are, and i can't write it back to 0 (all the other SBZ bits went back to 0 fine). might be nothing though.

maybe it has to be done physically like tegflyer's idea. i did see something in another tech doc from amd talking about JTAG pins.

if I had the money I would make him a very nice offer for it...
double sounds fair doesn't it...

Look, if i've understood things correctly,

(this is just pseudo-:banana::banana::banana::banana:)

setMultiplier(number x)
if(x < MAX_MULTI_FOR_THIS_CPU)
currentMulti = x;
else
dont_do_jack_:banana::banana::banana::banana:;



right?

this means that the register where currentMulti is stored, must be writeable, right?
So what needs to be done is to set the currentMulti-register, WITHOUT going in the main door (the one that checks the max multi available), right?

The only other option as I see it is to manipulate the MAX_MULTI and set it to a higher value, which most likely is not possible.


Now, I might be only a software architect, but sure as hell one must be able to address that section of the cpu.

Quote:

---

Originally Posted by Ref

guide to unlock venice: put it between ashtray and mousepad for 2 weeks :banana::banana::banana::banana:

---

:LOL:
:ROFL:
:LMAO:

tegflyer

This is unreal (about CPU and JTAG) because of you have no any PDF info regarding A64 CPUs describing JTAG Protocols and Commands for these cpus.

People measure your CPUs, do not be lazy fat mices. ;) Maybe Pirs CPU lost connection with one of STRAPs or another side - shorts to GND such pin and get free multi...

If no, it is also possible to write bit to 0 to CPU PROM (because when PROM manufactured all it's memory bits are 1 and when it is programmed, necessary bits become 0. But there is possible to re-programm that bits that remain 1 to 0.

what exactly do you want mesured? ill be pulling my CPU to install my new waterblock sometime next week, if you can tell me what you need i will gladly test it for you.

everything if possiable, because I plan to clone a Venice of my own to the exact unlocked specs

Acrazar

I describe it few page earlier. There is a pic.
You have 3000+ x9max cpu so your data will not be very important. but if you do if will be another piece to resolve this maze.

I need cpu owner with not venice3K cpu. I need to compare x9max cpu with others. Then I plan to isolate different pins and look what we get.

Hi ALL

anybody can send me dead 939 CPU ?
Only important that cpu core is not mechanicaly damaged

I already done some researches.

Also I isolated such pins:

STRAP_E11 - nothing changed. CPU start and work as usual.
STRAP_F11- nothing changed. CPU start and work as usual.
RESET_L - boot fail. black screen
POWER_OK - nothing changed. CPU start and work as usual.
TDO - nothing changed. CPU start and work as usual.
TMS - nothing changed. CPU start and work as usual.
TCK - nothing changed. CPU start and work as usual.
(BTW, I think JTAG pins are only for CPU testing during MFG process, they aren't used by regular CPU work)

I'll try to isolate each NC_xx pin one by one and see what happens.

But I need some hardware :confused:

1. any dead cpu with multi other that 9
2. any dead FX even 51
3. any dead mobo for with good CPU VRM and not damaged socket (for powering up system and look thru boot up sequence)

If anybody have such things and can help tell me to tin@topmods.net.

We are going to do FX from Venice 3000. :slobber: Every small detail can help us. So, don't go thru topic and just laugh. :rolleyes:

Quote:

---

Originally Posted by TiN_EOF

Acrazar

I describe it few page earlier. There is a pic.
You have 3000+ x9max cpu so your data will not be very important. but if you do if will be another piece to resolve this maze.

I need cpu owner with not venice3K cpu. I need to compare x9max cpu with others. Then I plan to isolate different pins and look what we get.

---

i already am waiting to compare notes..
i got a 3000+ ready and waiting to be messed with...

nn_step

can we meet in ICQ ? minee 86080981

Oobi Doob

PLZ, send me full version of HWDIRECT

Because now I have no HW for experiments I'll try something software-level tricks.

you are added and you just need to buy a $5 multi meter and have a digital camera...

understood, will post results on wed or thurs when my stuff comes in

if there is any difference we will find it...

i have a locked venice 3000+, how can I :help:

:)

KoHaN69

You can kill it and send to me :)

Kidding ;)

Now you can't help, because your CPU is x9 :)

Nice researchs

PS: Suscribed :D

Jus a thought, not sure if anyone has considered it. Someone said that the venice die could have been put on the wrong package, ie an FX one... this was discounted because dies are mounted on the package, speed binned, then have their max multi programmed. Thats all well and good. One thing i think has been overlooked (came to me in my sleep :P)
FX CPU's are of the SSOI process, not the standard SOI process. So they kno to a certain extent, which of their cpu dies are gonna be of the FX nature cus theyre strained silicone. With that in mind, could they have packages that are primarily used for the FX (SSOI) format?
Maybe this package was used by accident?
Dont kno too much about all that resistances and pins n all that. Plus i own a winnie so sorry

the second the data I requested is posted..
i will Find a way to unlock these things even if it costs my whole CPU budget for the YEar...

I dont see any progress here...

so I am ready to sell my Venice on ebay or here, maybe someone will be more succesful ;)