Can you possibly try it in another board?
Printable View
Can you possibly try it in another board?
this would be so awsome if its figured out how to unlock them
This kind of resembles stem cell research, in a way. Us curious people represent the scientists who know that being able to research/experimentstem cells could make break throughs in science and help alot of people, but then there is people with morals, in this case, the man of the hour with the unlocked venice. Not that he has morals, but he has this vital piece of information that could benefit many other people if it were in the right hands, or if he grew some right hands. anyway, i think that you should find a nice guide, take the ihs off, just so you can see everything, and take the pictures, also pictures of the socket to see if there is anything strange!
and about selling it to amd for an fx-57, that would be the most self-centered greedy thing in history! im ashamed!
welcome to xs xfx! is it cuz of me u joined? just admit it. :P
lol, i joined because i was going to ask about v-modding my mobo, but i decided i screw things up easily and soldering isnt for me. ill say its because of you though http://img3.imageshack.us/img3/8307/81649yay25kd.gif
Mr Tinker, think beyond 2 dimensions. The laser cuts the bridges...it MISSED or barely cut one and heating-cooling cycles/manipulating/fondling/handling the chip managed to get the connection re-established AFTER it left the factory.
yeah......maybe it got dropped, even from say 1 inch into the ashtray, rattling just the right part to reconnect a certain something.
i think it needs to be tried in a different motherboard!
and handled with extreme care!Quote:
Originally Posted by welshtom
hell!!! Why dont we all cover the bottom of our CPUs with something conductive (AS5 or dust) and see if it unlocks......or explodes :D :D
you first
AS5 isn't conductive...it's capacitive, it will have a different effect.
I'm still game for someone with a Venice to compare it to taking a REALLY close look.
on my s754 setup came some AS5 accidentally, I could buy a s939 setup... it died :(
on topic: :slobber: :slobber: :slobber:
As I earlier said I got a s939 +300. venice.
first it's a crap overclocker (stuck at 2700mhz)
second: locked multi :(
xyus
I tryied my venice in ASUS A8N-SLI Deluxe and it is too unlocked.Quote:
Originally Posted by welshtom
ok great :).
Didnt do anything abnormal or anything did ya mate? stick it near a huge magnet or something ?
Tom
I understand you now. So to paraphrase your previous post, this is what needs to be done to find out how to unlock A64's:Quote:
Originally Posted by uwackme
1. This chip would have to have the black epoxy removed, and
2. the same done to a standard locked Venice.
3. Compare the two under a low power microscope.
do you mean the black epoxy I've circled here:
signed
@pirs , where do you live ?
He is from Czech Republic, look at his Location.
(inac Pirs, ja som zo susednej krajiny, mas fajne stastie s tym procakom :D )
english please :p
:)
nvm, you are probably asking if he is from west or east ...
In fact, if some one just compares a 3000+ venice with a 3200+ venice, he would also see a difference, because the have different multiplier, and so people could mod there 3000+ to a 3200+ for exapmle.
But that would be a easy job to do, so I think it might be a little harder then that.
How do we get the epoxy off? It's time to start looking into that.
I'm trying with my IHS free, 30 broken pin 2.66b....it's tough!! (and I have nothing to lose!)
where can we find out what that epoxy is made of and how to dissolve it? Might there be anything that dissolves only the epoxy and nothing else on the cpu? How would you google that epoxy?
would this help?
http://www.jasco-help.com/products/prod_rem.htm
I'd be careful using that close to the core though. IIRC the core is glued to the chip with epoxy as well.
Maybe a bit of that stuff on a q-tip, lightly and take a lot of time.Quote:
Originally Posted by Gogar
The public A64 materials docs just say "epoxy." They also say BT Resin:
ADDITIONAL INFORMATION / ADDITIONAL MATERIALS
The following materials are not listed in EIA’s Material Declaration Template, but are contained in this package.
Material Wt% ppm mg
Aluminum < 0.39 < 3,900 < 155
Barium Titanate < 0.27 < 2,700 < 108
BT Resin < 2.6 < 26,000 < 1,036
Epoxy < 0.87 < 8,700 < 347
Iron < 0.087 < 870 < 35
Phosphorus < 0.0019 < 19 < 0.76
Silicon < 0.85 < 8,500 < 339
Silicone elastomer < 0.34 < 3,400 < 136
SiO2 Filler < 0.30 < 3,000 < 120
SiO2 glass mesh < 3.2 < 32,000 < 1,275
Tin < 0.14 < 1,400 < 56
Zinc < 0.0038 < 38 < 1.5
Zinc oxide < 0.14 < 1,400 < 56
Also, on page 375 of the "BIOS and Kernel Developer’s Guide for the AMD Athlon™ 64 and AMD Opteron™ Processors" it says this:
"New FID (NewFID)—Bits 5–0. This field is the new FID to transition to. If an attempt is made to write a NewFID value greater than MaxFID in the FIDVID_STATUS register then the MaxFID value is written instead. See Table 80 for FID code translations."
So we want to hide the maxFID value, or change it. I'm sure I'm not saying anything that hasn't been said.
hahaha, if 2.7 sucks then just give it to me, ill trade you my 2400+m 35w :toast:Quote:
Originally Posted by Xyus89
i just removed my ihs from me venice 3000 2,8 RS but 2,7 Daily my vitesta 566 can't go 310 1:1 but can handle 300 1:1 1T :D
http://img144.imageshack.us/img144/8051/image2032sy.jpg
http://img144.imageshack.us/img144/7165/image2024io.jpg
http://img144.imageshack.us/img144/7715/image2050rr.jpg
http://img144.imageshack.us/img144/2029/image2045xv.jpg
and gain'd 5/6°C and 3 on the chipset :banana:
omg its just laying on the bed... i hope it didnt get shocked... or maybe it got shocked and becomes unlocked:D.
i've done it because i've red this thread..but no gain on multiplier..but atleast now i'm ready when someone will find out how to unlock it :D :woot:
unscharp pictures, i think your camera is to close. increase distance between camera and object and use a magnifying glas to zoom into the details of the proc. if you have a magnifying glass offcourse...
The pictures are not clear, and in such a unique case, we would need very detailed pictures. Try to use another digital camera and try to set the focus better this time.
Here are my 3 chips my Venice, San Diego and my 3800 X2. I hope these help!
There's something very solid white paste under the two black strips under the HS. And under the white strip - absolutely nothing.
The smaller black slab removed with a (sharp) surgical 10-blade:
http://www.xtremesystems.org/forums/...id=35945&stc=1
go deeper :D
Nothing under the bigger one either... :(
This slaughtered 3000+ has a text "J9" in one of the corners. I'm just guessing but could it be that the package the core is glued to specifies the multi(es) that it can use?
J9 = FID ≤ 9?
http://www.xtremesystems.org/forums/...id=35944&stc=1
What does a 3200+, 3500+ or FX's read?
well, poop.
Maybe this should be called "plastic surgery" :D
Thats interesting, maybe pirs venice was glued to the wrong package?Quote:
Originally Posted by largon
Tom
We need to find out what other models have in that corner.
I can check what's on my winnie 3200+, but atm I don't want to remove the IHS of my Venice :)
sorry guys i have no digital camera
it was made from my nokia 6600 only that's why so crappy :p:
I guess we want to see Pirs' CPU w/o HS. Or *any* other unlocked Venice.
|SiLA|'s / or anyone elses pics are quite useless if the proc on them isn't Pirs'...
If anyone is wondering, the proc I sliced 'n diced - is still alive.
Interesting thought... what multiplier was your Sandy running at? Would the bios then attempt to run your cpu with this multiplier after you swapped cpus over? And was this enough to change the cpu?
ie. Was the mobo's bios forceful in telling the cpu what multiplier to run at, and the cpu chucked a hissy-fit and started carrying on like "fine, from now on, YOU can tell me what multiplier to use... Im sick of doing all the thinking for myself!! How do you like that, Mr. ImAFussyBiosWhoHasToHaveMultipliersHisOwnWay?"
Or something like that.
prueg,
AMD's tech docs describe how a multiplier is applied. Board, bios or previous proc can't affect that. The process is totally internal to the cpu.
But, if Pirs' core is glued to a package meant for FX, it would seem to explain why the multis are enabled.
There are some reports of partially unlocked multipliers (2 different 3000+ with multis up to 11). These 2 could have package of a 3500+.
Why would cores end up being glued to wrong packages is anyone's guess.
:2cents:
Guys go measure resistance on the pins, that's the only way.
Everything done with the mulipliers is done inside the core.
Yes Mr Tinker, exactly.
Based on Largon's picture nothing "obvious" seems to be there. I'd need to take a very close look at a dead Venice to try and ferret out the bridges. You'd need to also use a powerful light from underneath, as the bridge could be on an inner layer that the laser reached into.... an infrared laser could work without actually pocking a hole in the upper surface of the substrate.
So a light table to backlight the PCB/substrate material may show something on inner layers in the black covered areas.
Or they could be total dicks and put the bridges UNDER some of the resistor/capacitor SMD packs, and we need to remove all the caps/res's packs to figure out which the bridges are hidden under. Not a biggie but 1st time should be a DEAD Venice :fact:
But again, the chips are placed on the substrate BEFORE powerup-and-testing and are then speed binned. They cant possibly test the raw cores as the current requirements are so high no test fixture could handle the needs on a raw core level. So first the cores go on the substrate, then they speed-grade/test them, then they set the MAX-fid value somehow. It's always possible they put a PROM like circuit inside the core that gets "blown" once the chip is tested.... but that would be going overboard.
Nice pic's Largon, wish I could get my camera's macro down that fine!!
Well, On my Winnie there wasn't a J9, so prolly this will be only on Venice's... But Pirs, can't you take some pictures of it?
he's not gonna remove the IHS tho' :slapass:Quote:
Originally Posted by FnF
but lucky him with that :banana: cpu
http://www.pirs.vyrobce.cz/Images/venice/predek.jpgQuote:
Originally Posted by FnF
http://www.pirs.vyrobce.cz/Images/venice/zadek.jpg
I dont know without looking but the pins on the corners is that normal? I thought only one corner looked like it was missing pins but all 4 hmmm let me go look.. OOPS sorry guys I just looked and it looks normal!
largon could u pls post pics of yer pins?:D i jsut wanna check sth :D
what are those circle'd with reds?
http://img335.imageshack.us/img335/1...ited9bp.th.jpg
I think your onto something cuz the 2 on the right look connected or burnt or something. The pcb looks a little black.Quote:
Originally Posted by |SiLA|
i hope largon can check his venice or make pic of itQuote:
Originally Posted by njkid32
mine is here running atm don't wanna remove it and change water again lolthere's like 2 bridges 1 white and 1 black bridges
6 up and 3 across from the topmost left circled one there is a funny pin also.
Tom
I'll post pics of my 3000 when i change over to my vapo ls next week.
Do you have access to the equipment needed? If you think you're up to it, I'd be willing to donate $10 toward this research. Get about 11 more people and we could grab a venice from ewiz, or find a dead venice. Maybe we should make a post in the WTB section for a dead venice.Quote:
Originally Posted by uwackme
check this 1 also from the data-sheet of AMD
i circle'd the two circle'd "bridges" on pir pins pic
http://img247.imageshack.us/img247/4...ited7bj.th.jpg
That's a Winchester in my pics too.Quote:
Originally Posted by FnF
But, in the end because a proc is binned after assembling, packages can't be pre-built to be core-multiplier-specific.
Circled pins from up to down:Quote:
Originally Posted by |SiLA|
1st: VDD
2nd: MEMDATA[5]
3rd: MEMDATA[4]
4th: L0_CADIN_L[5]
5th: L0_CADIN_H[6]
this is getting very interesting....although can someone tell me what the pins circled in red are? to me they look like the rest??
i circled the pins of Pir's venice coz they look like bridgedQuote:
Originally Posted by mnewxcv
I'm sorry, but are you sure that's right? I think you got the orientation of the chip wrong.Quote:
check this 1 also from the data-sheet of AMD
i circle'd the two circle'd "bridges" on pir pins pic
Look at the 4 pins gone at the top left of the table, the 3 pins gone on the top right.. And then the 2 sets of 2 empty pins, on the same height horizontally.
You have to see the chip as if it were turned over with the golden corner on the top left.
jup i'm sureQuote:
Originally Posted by Gogar
just turn the amd sheet pic to make it look like Pirs pic
/¯¯¯¯¯¯¯¯¯¯¯¯\
| |
| |
| |
| O|
| |
| |
____O______
that's Pirs' pic
amd sheet is
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯\
| |
| |
| |
|O |
| |
| |
_______O_________/
edit
looks ugly when it's posted lol
http://img349.imageshack.us/img349/840/pirs4ln.jpg
http://img349.imageshack.us/img349/2118/amd5px.jpg
we need someone with a extra 3000+ venice to join those pins that are circled on their own CPU and see if it gets unlocked.
I, however, dont have any idea how we would do that because the space in between the pins is so small it is hard to hook them up :D
This is my conclusion of the proverbial pins.
easy 1 peice of wire into the socket itself, just like the uwire dothan mod voltQuote:
Originally Posted by Rezag1000
similar to this. http://www.xtremesystems.org/forums/...id=31255&stc=1
You mean just rotate pir's pic 90 degrees clockwise right?
Look at the empty pins in the middle of the chip, and where they would end up.
It can't be right, because on the pic the two open spots at the bottom would be in one straight line:
AE X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
AF X X X X X X X X X X X X X X X X X X X X X X X X X X X
AG X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
while on the table they are like this:
AC X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
AD X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
AE X X X X X X X X X X X X X X X X X X X X X X X X X X X X X
well the probs then is the amd sheet :slap:Quote:
Originally Posted by Gogar
it's a pic from the top(core resistances etc...) not from the "bottom"(pic of the pins)
:D
Surely if we know all the pinouts with the diagram, we can work out which ones could do multiplyers? Just by looking at the labels etc? Someone must be brainy enough
*runs and hides*
Tom
so who wanna try bridge'n those pins if they're not "bridged" on their cpu? :toast:
i bet some ppl from amd are lookin at this thread and :banana::banana::banana::banana:ting their pants. :p:
They are saying holy sh** they are getting close to cracking the code. LOLQuote:
Originally Posted by mnewxcv
Quote:
Originally Posted by mnewxcv
:woot:
for how long now the A64 is out?since the s754..and noone discovered yet how to unlock the cpu :(
perhaps they are sending out these teaser chips just to show its possible :p
or so everyone is gonna try to find it out and killing the cpu so we have to buy new cpus :rolleyes:Quote:
Originally Posted by welshtom
:nono:
Quote:
Originally Posted by |SiLA|
i think we're onto something....
lol
hehe, or that !!
Quote:
Originally Posted by mnewxcv
no more rats from the amd?who can help us by guidin us where and what to do to unlock it :stick:
anyway let's stay on topic :slapass:
anymore ideas? can't check mine..who can post another 3000 Venice pins pic? :stick:
Quote:
Originally Posted by |SiLA|
I can post that on monday/tuesday. But it will be covered in dielectric grease, not much use. I'll look for the "J9" thing though.
Tom
ill try and find some pics on the web.
Looking at some tech docs, the CADIN pins that appear to be connected belong to the hypertransport link i think.
I found some other interesting pins though.
http://www.xtremesystems.org/forums/...id=35897&stc=1
What could those be for? Not normally used? Then when are they used? Hector, tell me.. what are they for?!?! :slapass:
These might also be interesting. I'd compare an FX with a locked cpu and try to find differences in resistance between the pins. (not that i have an FX or the time to do it right now though)
http://www.xtremesystems.org/forums/...id=35898&stc=1
yeah, why hasn't someone compared pin resistances of an fx to a locked A64? Oh right: time. Gotcha. Well, someone needs to do that. Seriously.
Pirs 3000=TQuote:
Originally Posted by Onepagebook
http://www.pirs.vyrobce.cz/Images/venice/predek.jpg
Don't know how true it is about the Ts only on FX grade.I have SD 4000 with T
Quote:
Originally Posted by RAMMIE
Just today I got a Venice E6 3000+ which also has the "T", so I don't think that's something special. ;)
Wow, I just finished reading this whole thread, and all I can say was I was on the edge of my seat the whole time. It may be worth getting a 939 system and a 3000 venice, taking up smoking, following his exact way of living for 2 weeks, then get unlocked multi's. Nah, my sempron system has an 8x multiplier, so we need to change that. It may be time to check out all of those pins.
Are there others willing to shell out say $10 apiece for a community project to get to the bottom of A64 unlocking?
Quote:
Originally Posted by Mr. Tinker
yes
I am
Thanks
Why would you find any more information comparing his venice to other venices than you would comparing a san diego to a FX?
I got $10 on my paypal so count me in ;)
Have the cpu registers been compared to a locked 3000+ yet? I've only seen Pir's dump come by in this thread.
I would deff give $10.00!
Well, because no body posted some pics, here are some pics of a locked Venice 3000+ which are original posted by mklaassen:
http://home.student.utwente.nl/j.m.k...e/IMG_1311.jpg
http://home.student.utwente.nl/j.m.k...e/IMG_1318.jpg
http://home.student.utwente.nl/j.m.k...e/IMG_1305.jpg
very very interesting...
unlocked venice that used to be locked...
magic man, that reminds me a P3 550mhz I once checked and it was unlocked... ahhh old times.
AMD datasheets refer to CPU's "Model-Specific Registers" that store the useable multis, default voltages, id bits...
Page 324 in "BIOS and Kernel Developer's Guide for AMD Athlon™ 64 and AMD Opteron™ Processors":
MaxFID value is stored in a register called "FIDVID_STATUS":Code:Address Register Name Description
0410h MC4_CTL page 207
0411h MC4_STATUS page 207
0412h MC4_ADDR page 207
C001_001Fh NB_CFG page 365
C001_0041h FIDVID_CTL page 374
C001_0042h FIDVID_STATUS page 376
C001_0048h MCi_CTL_MASK page 207
If we could reprogram this "FIDVID_STATUS register" we could rewrite the "MaxFID" value to, for example 25...Quote:
Originally Posted by Page 375
btw. What's that "Clock Control Register Correction" section in the "Tweaks" tab of Central Brain Identifier.
Huh?
Edit:
Id for FID 9x seems to be 00_1010b, 25x is 10_1010b. Only the first bit is different...
Conclusion:
AMD techs must've overclocked their PROM machine - and occasionally it puts a wrong bit in the wrong place. Pirs' proc was programmed wrong...
:D
Quote:
Originally Posted by largon
Yet it was x9 when he first tested it..
Anyone try sticking in a FX x15 and then w/o flashing bios stick in a locked cpu? :slapass:
Wow.... :eek:
so maybe we could "trick" or reprogram the "FIDVID_STATUS register" using possibly a modded BIOS or maybe a home-made proggie??
If so.. Who is good at programming in assembily? :p:
Anyways really looking forward to seeing some experiments over the next few weeks.
Alec
Pirs proc shows 11_1010 as its max multiplier.
All the bits in that register are read only, so I really don't know how do you intend to write those registers. The only way I can think of is on system power up (maybe), when you just turn it on, maybe you can access that MSR and change it's value.
Footnote in "Table 80. FID Code Translations" page 375:Quote:
Originally Posted by The Coolest
http://www.xtremesystems.org/forums/...id=35943&stc=1
11_1010 is not included in this table... What if this "unpredictable behavior" includes all multipliers becoming enabled?Code:Encodings not listed are reserved. Using reserved encodings will cause
unpredictable behavior.
It won't most likely.Quote:
Originally Posted by largon
11_1010 results in a x33 max multiplier.
You can calculate the multi using this formula:
(Multi Code / 2) + 4 = Multiplier.
(Multiplier - 4) * 2 = Multi Code.