Athlon 64 3000+ with 11x multiplier

Hello,

One of our readers discovered that his Athlon 64 3000+ (Venice) with LBBLE-stepping (0518) has multipliers unlocked between 5 - 11x. Normally 9x is maxium with 3000+ CPU (9 x 200 MHz). Is this just one case or are there similar CPUs around like this one?

http://www.soneraplaza.fi/GetImages/...0,62665,00.jpg

looks a bit suss. the "3000" is a different size to the rest of the text in the first line? Failed 3500+?

And the "05" and "8" in "0518" look a bit off aswell - maybe it just me for the 05 and 8 though...

Is the CPU-Z bugged or is it unrecognisable and hence a new revision? :confused:

More pics:

http://koti.mbnet.fi/aop13/athlon64/

CPU-Z becomes bugged if the multiplier is raised.

I have two processor with same stepping, can't wait to test :D

They've tried the CPU on two different motherboards and the multipliers is unlocked in both motherboards..so it can't be a bug in the CPU-Z.

:argue:

i wish , mine 939 3.0+ [@300X9 stock voltage @ & 310X9 stock voltage with chiller phase change] winchester could have 11multi :hehe:

Strange stuff - but hey - wouldnt mind buying it if he decides to sell ... lol ...

is this true of all 3000+ then?

or just these certain steppings?

Quote:

---

Originally Posted by don_vercetti

is this true of all 3000+ then?

or just these certain steppings?

---

certainly not true of them all, its strange how those letters appear to be different sizes, all the ones i've seen are pretty much bang on all the time

What if he mounted 3000+ IHS on 3500+? And changed specification?

or perhaps it really is a 3000+ with 11x multi? don't be so quick to judge

There is no valid reason for the information stamped on the IHS to be in different sized letters is there? Look at the pics of the other two 3000s with the same codes...those are not sized differently.

That looks a little weird, possibly as mentioned above, someone is playing around with procs and switching IHS. That is certainly plausible, more so than for there to be a 3000 with an 11 multi. By design doen't that makes it a "3500" (2.2GHz)...by that I mean that a "3000" with an 11 multi would have to be running at less than a 200 MHz for it to be a "3000" (1.8GHz).

not neccessarily...if it is real and "unlocked," the default multi could still be 9x, but you can adjust up to 11x...assuming this is true

Quote:

---

Originally Posted by aznblueeboi

...assuming this is true

---

Isn't that the real question?

I don't really care one way or another, but just making an observation about the "3000" in question.

AMD's only known unlocked cpu is the very high end FX. The thought of AMD making this option available on a few of its entry level 3000 cpus just doesn't seem very likely, does it?

I'm sure stranger things have occurred, but again, I just find it unlikely that its true and more likely that its someone playing around.

I know there was a bug with a particular Gigabyte board in the past....

This smells of a gag...... code doesn't look right and what are the odds that a cpu with a funny looking code has a default multi unlocked up two multis? I'm not ready to believe it just yet.

Quote:

---

Originally Posted by TMM

And the "05" and "8" in "0518" look a bit off aswell - maybe it just me for the 05 and 8 though...

---

Sorry but what the heck are you talking about?

05 = year 18 = week

or what are you trying to say? :confused:

The 8 is kind of floating away from the rest of the numbers for one. :D

Oh well, whatever insight sampsa inputs I'll believe. edit: Oh sounds like he just got the pic from some one else. So should we picking it apart then? I don't really have a tin-foil hat. The letters thing is definitely visably strange. Beyond that, well he'd be as good a judge as anyone.

Quote:

---

Originally Posted by sai

Sorry but what the heck are you talking about?

05 = year 18 = week

or what are you trying to say? :confused:

---

He is saying that if you look at the numbers for the dates and compare them to the other numbers on the codes the size of the numbers is not the same, just like the "3000" in the first line is larger than the numbers and letters on the other sides.

Compare the size and shape of the codes on the IHS shown in post #1 to those pictured in post #4 and you will (should) see a clear difference in them. Those pictured in post #4 are consistent in their size, as opposed to those pictured in the first post.

It's a secret conspiracy by AMD, first "frequency doesn't matter" now "multis don't matter". :D

Seriously, looks liek were are being rooked, time to ignore this one. Speculation is interesting, but the most probable explanation is the likely explanation.

Yeah, I doubt that's real..

Could AMD gives us 11x for venice 3000+ passed 05 = year 18 = week ??????

that would stop it being a 3000+ wouldn't it. It would become a 3500+.

Quote:

---

Originally Posted by sai

Sorry but what the heck are you talking about?

05 = year 18 = week

or what are you trying to say? :confused:

---

look before you shoot smart ass.

looks like it was just a mess up huh? maybe a hung over worker from the weekend not paying attention to what he has doing on the assembly line.

I don't believe that there is some guy sitting there at a factory arranging letters to stamp info on an IHS and picking the wrong numbers in addition to the wrong size numbers. You have to believe that all this stuff is automated...the batch's information is entered into the equipment stamping the IHS and then you just walk away while it stamps a few thousand or so. Nobody "messed up" on the assembly line.

When we see 4 or 5 of these popping up then maybe it won't be viewed as someone having fun...until then...well...I think its just someone having some fun. :)

that cpu is not a fake... it's real,not switched ihs... two users has tested that cpu...

About the authenticity of this cpu, I dont know if the cpu is real or not but what I can tell you is the information that is "stamped" on the IHS is actually laser etched (and not manually!) in Singapore so how you could get variation in size I dont know

Quote:

---

Originally Posted by Reefa_Madness

Look at the pics of the other two 3000s with the same codes...those are not sized differently.

---

All three processors have similarly sized codes. :rolleyes:
You can verify it by measuring the characters and/or drawing some lines in Photoshop.
It would be easier if we had a close-up pic of the code like the other two.

Quote:

---

Originally Posted by Reefa_Madness

That looks a little weird, possibly as mentioned above, someone is playing around with procs and switching IHS.

---

First you say the code on the IHS doesn't look right (basically meaning you think the code has been edited).
Then you say the IHS has been changed.
Why would they edit the code if they change the IHS? :rolleyes:

Quote:

---

Originally Posted by Reefa_Madness

That is certainly plausible, more so than for there to be a 3000 with an 11 multi. By design doen't that makes it a "3500" (2.2GHz)...by that I mean that a "3000" with an 11 multi would have to be running at less than a 200 MHz for it to be a "3000" (1.8GHz).

---

Did anyone say that the default multiplier was 11? :rolleyes:

Did anyone just write AMD and ask?

we don't want them to stop doing it, do we.

as if they wouldn't know

Quote:

---

Originally Posted by sai

Sorry but what the heck are you talking about?

05 = year 18 = week

or what are you trying to say? :confused:

---

i mean that the printing of those 3 numbers looks different then the other numbers on the IHS

If it was a fake, and it really was a 3500+ it was supposed to show 3500+ in cpu-z. Wich it doesn't. It shows as a 3000+ in the internal cpu name, or whatever they come factoried with.

Seems real to me, tough strange it comes with those multi unlocked.

What about trying measuring the pins resistence with a multimeter looking for something that would unlock all AMD64s?

Quote:

---

Originally Posted by Viper8

All three processors have similarly sized codes. :rolleyes:
You can verify it by measuring the characters and/or drawing some lines in Photoshop.
It would be easier if we had a close-up pic of the code like the other two.

Response:
All three have similarly sized? Poster #2 and #9 before me at post #12 and posters #16, 18, 20 and 21 after seemed to disagree with you. You may be right, but then again you may be wrong and have shown no proof, just your opinion...which is all we did.


First you say the code on the IHS doesn't look right (basically meaning you think the code has been edited).
Then you say the IHS has been changed.
Why would they edit the code if they change the IHS? :rolleyes:

Response:
That point is valid if you go with the notion that I said both where done at the same time, a switch and an edit. I only indicated that the suggestion made by the guy before me was plausible, that the sink was changed or alternatively that it was altered...never said both. Once again, the original suggestion was first made my someone else but yet you choose to attempt to ridicule me with your response. Please note the key word here being "attempt".

Which leads me back to your first comment...if the letters and numbers are correctly sized, as you've asserted, then the "switched IHS" is still a valid explanation. Do you know the guy that has the cpu and have you inspected it personally? If not, then you are strictly speculating...just like I'm openly speculating. I'm just as entitled to that speculation as you are. If you have inspected the cpu and are trained in this area, having removed and re-intalled many IHS on A64s then your opinion gains validity, but without that training and experience don't hold yourself out to be the expert, especially when simply looking at pictures.

Did anyone say that the default multiplier was 11? :rolleyes:

Response:
That point is valid, no one stated it was a default multi and therefore that one comment of mine should be ignored, however, if you continue to "rolleyes" on every post they are going to roll back into your head and you'll be able to see even less than you do now.

Instead of choosing to give your opinion, you've chosen to attack mine, like I'm some kind of ignorant arse. Keep the comments on topic and factual (or identified as an opinion) and I won't have any problems...start making them personal and the mods will deal with it.

I don't really give a cat's behind whether or not you agree with me and I make no apologies for not being in "awe" the minute someone posts a picture of an oddity in this age of Photoshop and other digital masteries.

I remain skeptical but certainly would love to be wrong and find that there are in fact unlocked Venice 3000s out there.

To quote myself:
"I don't really care one way or another, but just making an observation about the "3000" in question.

AMD's only known unlocked cpu is the very high end FX. The thought of AMD making this option available on a few of its entry level 3000 cpus just doesn't seem very likely, does it?

I'm sure stranger things have occurred, but again, I just find it unlikely that its true and more likely that its someone playing around."

Nothing that you've presented in your post changes my opinion, as stated above.

---

Enjoy the 3000/3500/3700/FX or whatever.

There is nothing wrong with the text "3000" on the proc...
It is very common that numbers are a bit bigger than letters.
Random pic from a Venice OC thread:

http://www.xtremesystems.org/forums/...6&postcount=89 http://shared.sploitz.com/system/A64-v3200-0515.JPG

That is a very good pic and clearly shows the numbers to be larger than the letters so it appears that you can say that the IHS may very well be unaltered. My winnies (3) are all the same size so maybe it (numbers sized differently) is this way on the newer cpus, such as the Venice core.

That is all you can say from that pic.

For the record, I am not sitting here saying that the thread starter is a liar...its not even his cpu. However, the authenticity of the cpu remains a question in my mind.

I remain skeptical about a 3000 cpu with unlocked higher multis but like I've already stated...stranger things have happened and its not beyond the realm of possibilities.

It will be interesting to see whether the other two that Benny L. has with the same stepping are also unlocked.

Can they give a CPU-Z verified link for this CPU running @ 11 or 10 multi?

Here you got it.

http://koti.mbnet.fi/aop13/athlon64/...0+_special.png

Here's the link for those that can't be bothered to type it in...

http://valid.x86-secret.com/show_oc?id=29156

:)

Quote:

---

Originally Posted by largon

There is nothing wrong with the text "3000" on the proc...
It is very common that numbers are a bit bigger than letters.
Random pic from a Venice OC thread:

http://www.xtremesystems.org/forums/...6&postcount=89 http://shared.sploitz.com/system/A64-v3200-0515.JPG

---

Hey that CPU looks familiar :P

There are few interesting things.
1. CPU-Z is not able to recognize core revision, code name etc.
2. 11x multi unlocked while CPU name is 3000+ - means that they are not connected to each other
3. different size letters and numbers

Question:
What is the default multi... 9x or 11x ?

There has to be a reason for this ... these are not fake or remaked procesors (IHS has not been removed) ... and it happend more than once ... but did it in the same location or not? does anyone know?

this is not unheard of a guy on here reported having a multi of 11.5 instead of 11 on a cpu of his

i found nemo
I remember that thread. Wasn't it a Gigabyte bios glitch?

Very interesting indeed. I doubt this is a fake, just a very rare freak chip. :)

regarding the fact it claims to be a 3000+;

isn't it possible to alter some code so you can make the CPU say it is whatever you like? As in if you could get an athlon fx to 15x200, you could rename it as an athlon fx 61 and it would look plausible?

I'm not saying I can do this, im just sure it's do-able.

:rolleyes: I can't believe everyone's saying it's fake etc, and not saying congrats on a lucky chip.... :toast: on the 11x 3000+, that's what I call value

Quote:

---

Originally Posted by don_vercetti

regarding the fact it claims to be a 3000+;

isn't it possible to alter some code so you can make the CPU say it is whatever you like? As in if you could get an athlon fx to 15x200, you could rename it as an athlon fx 61 and it would look plausible?

I'm not saying I can do this, im just sure it's do-able.

---

it's really easy to do all you need to do is a simple registry edit

i think its real. i just think it got by without getting locked thats all no big deal. all the cpu's amd makes i sure there are more then just one lucky fool who gets an unlocked cpu thats not an FX chip... i say awsome good find!!

peace perc,

You see the lettering like that a lot. I don't doubt the authenticity because of the lettering.

If you look at the guy who asks which stepping is good for the 3000+ (or was it 3200+) where he posts like 10 pics of different chips, a few of them have lettering like that...

pretty wierd. looks like AMD messed up on one of the chips. was most likely supposed to be a 3500 but it was named a 3000 on accident.

possibly, but the whole process is automated.
So unless the workers have been overclocking the assembly line, and making them glitch, i doubt that is the reason.

Maybe AMD have just decided to make life easier for overclockers!

What a nice lookin CPU. I want one, too. :D
BTW I don't think it's fake, but still wondering how anything like this have happened at all. Never seen anything like this for sure. :confused:

Quote:

---

Originally Posted by don_vercetti

isn't it possible to alter some code so you can make the CPU say it is whatever you like? As in if you could get an athlon fx to 15x200, you could rename it as an athlon fx 61 and it would look plausible?

---

Yes it is possible, but not by us. AMD's the only one that can hard-code the cpuid string of a cpu. Here's my "FX56" :p: It boots @ default FX55 speeds (13x200), but it's cpuid was coded by AMD's Design Verification team as a FX57 to be used for bios testing before the release of the cpu. There's more info in the link in my sig :) :
http://img153.imageshack.us/img153/4202/fx8mh.jpg
http://img167.imageshack.us/img167/4749/579ob.jpg
http://img207.imageshack.us/img207/8...lidated5su.jpg

Quote:

---

Originally Posted by Perc

i think its real. i just think it got by without getting locked thats all no big deal. all the cpu's amd makes i sure there are more then just one lucky fool who gets an unlocked cpu thats not an FX chip.....

---

That's impossible. If it was truly unlocked, multipliers above 11x would be available. And there's also no way an FX die would be produced alongside a venice ;)

you don't need to hard code it. The guy's hardly handing the chip round saying test it. As long as it's in his computer, he can make it look like whatever he wants.

you never know, but its not somthing i will scream to the heavens.

Quote:

---

Originally Posted by don_vercetti

you don't need to hard code it. The guy's hardly handing the chip round saying test it. As long as it's in his computer, he can make it look like whatever he wants.

---

I'm not too sure what you mean, but ok...

ok. The guy above said that it is possible to change what the CPU claims to be, i.e. a 3500+ could say it is a 3000+ in CPU-Z, by changing the registry?
That is true, yes?

Quote:

---

Originally Posted by don_vercetti

ok. The guy above said that it is possible to change what the CPU claims to be, i.e. a 3500+ could say it is a 3000+ in CPU-Z, by changing the registry?
That is true, yes?

---

No ;) Take a look at my cpu-z submission. While the cpu-spec name can be altered (FX25), you can't change the hard coded cpu-name (FX57)

oh rite lol i didn't notice that fx-25 bit. puts a downer on that theory, thanks for clearing it up. Personally, i think it is genuine, there must be some reason for it (ES maybe?)

Quote:

---

Originally Posted by IvanAndreevich

i found nemo
I remember that thread. Wasn't it a Gigabyte bios glitch?

---

yes ivan, i'm pretty sure it was

oh give it a rest. AMD makes millions of chips per year and omg like 1/1000000 chips cant have accidently got higher multis.. :slap:

i reckon the adv of a multiplier of 11x cos u prob wont b able the reach the highest fsb u can achieve on 9x anyways. gd luck. pls update on results. :)

Hey, guys! Look at this http://www.hardforum.com/showthread....2&page=1&pp=20 THE SAME STORY And even dated on 06-09-2005 :)

any overclock with 11 multi ?

I think, that one way to find out what core that CPU has is to remove the heatspreader and then to measure the DIE square.

I think this is not fake. This CPU has very interesting cpu id - F30h. It seems to me that we have remarked Athlon64 X2(maybe one of two core was damaged).

I think it's just something that guys from AMD do in porpouse. They knew this would become a known thing, so they made some 3000+ to be unlocked up to 11x, as a 3500+, just to see what would happen.

Quote:

---

Originally Posted by PoL

I think it's just something that guys from AMD do in porpouse. They knew this would become a known thing, so they made some 3000+ to be unlocked up to 11x, as a 3500+, just to see what would happen.

---

Would it not be more plausable to say that it was a 3500+ that didn't
meet spec and they accidently left the multi at default (11) ???

i think maybe is an ES

they should make this mistake more frecuently

@seldomsean

I don't think so, cuz you gotta think how exact all those processes are, so there is no place for such a mistake. If that could happen to one chip, that would happen to all chips, mainly to those with the same stepping. :rolleyes:

Quote:

---

Originally Posted by PoL

@seldomsean

I don't think so, cuz you gotta think how exact all those processes are, so there is no place for such a mistake. If that could happen to one chip, that would happen to all chips, mainly to those with the same stepping. :rolleyes:

---

I agree with you to there..But were all human and DO make mistakes ;)


I am also thinking this might (Could be) correct...

Quote:

---

I think this is not fake. This CPU has very interesting cpu id - F30h. It seems to me that we have remarked Athlon64 X2(maybe one of two core was damaged).

---

That could be also... but don't really know... Cuz there are not Venice X2... :rolleyes:

or a 3500+ clawhammer with 512k L2?


i have same proc., except mine says cpaw. not unlocked to 11 multi. must find mopre fpaw's.

and who's got that gigabyte bios?

Quote:

---

Originally Posted by PoL

That could be also... but don't really know... Cuz there are not Venice X2... :rolleyes:

---

No but they are both 939's Correct?

Noone really knows what AMD does with the "Failed" proc's
except mark them down to a lower grade proc...

Ohh and BTW here ya go..( :rolleyes: ) lol

Yea, that's right, but I guess software reads the CPU code by their revision, that's why it's gotta be a Venice, cuz if it were a winchester it would be D0, or if it was a X2 it would be E6...
I also think it's something we'll never find out... :P

Quote:

---

Originally Posted by PoL

Yea, that's right, but I guess software reads the CPU code by their revision, that's why it's gotta be a Venice, cuz if it were a winchester it would be D0, or if it was a X2 it would be E6...
I also think it's something we'll never find out... :P

---

lol ;) I agree with what your saying..

And especially this...

Quote:

---

I also think it's something we'll never find out...

---

:banana:

-Sean

Quote:

---

Originally Posted by leviathan18

i think maybe is an ES

they should make this mistake more frecuently

---

It's not an ES - they don't look like that... ;)

Quote:

---

Originally Posted by cadaveca

or a 3500+ clawhammer with 512k L2?

---

Your comment made me think for a bit and you may be right, sortof :p: I looked over the cpuz submission again and I'm begining to think that this may be a 754 pin Clawhammer 3000+ with a default speed @ 10x200. It would have to be running on a board similar to the glitched Gigabyte board that someone had reported in a past thread. That board had locked his multplier higher than should have been possible for his cpu. BTW, IIRC, a Clawhammer 3000+'s cpuid is something like 0xF4A and that pic of the cpu may not be the same one shown in the cpu-z submission....

Quote:

---

That could be also... but don't really know... Cuz there are not Venice X2...

---

An official cpuid of Toledo - 00020F32h. And now look at this photo http://overclockers.ru/images/news/2.../toledo_01.jpg - this is an ES of Toledo(http://www.hwupgrade.it/articoli/1193/index.html).

If you want to learn more about your CPU try to remove the heatspeader and then to measure the DIE square :-) It must be ~200 sqr. mm.

Quote:

---

It's not an ES - they don't look like that...

---

Yep, it's not an ES, cause It has BrandID = 4.

Most likely it Toledo (200*11) which has already been marked at a factory, but here have found out some troubles with the second core and they put another CPUID registers into CPU to identifie on another way and they have interrupt marks. Especially it looks plausibly in a view of that the cost price manufacture is high, but it is necessary to reduce of the price in August. What shold they do? - Raise rentability of X2 production. How? Easy: to sell objectionable garbage, as cheaper processor.
(c) overclockers.ru

PS: surprise for Lexagon ;)

That's no hoax. It's my CPU and here is a validation link:
http://valid.x86-secret.com/show_oc.php?id=29156
(and pics in my homepage: http://koti.mbnet.fi/aop13/athlon64 )
I bought it 6/27/2005 from finnish online-store called ATK-Agentti. We have tested it in Asus A8N-SLI and Epox 9NDA3+ mobos (epox is my, asus is my friends). Multipliers 5-11x work and 9x is default.

edit: When multiplier is 9x CPU-Z says that the CPU is Venice but when I increase multiplier CPU-Z doesn't regonize the CPU.

edit2: Took me over a week to answer because I was on trip to other side of my country.

Quote:

---

Originally Posted by Ref

There are few interesting things.
1. CPU-Z is not able to recognize core revision, code name etc.
2. 11x multi unlocked while CPU name is 3000+ - means that they are not connected to each other
3. different size letters and numbers

Question:
What is the default multi... 9x or 11x ?

There has to be a reason for this ... these are not fake or remaked procesors (IHS has not been removed) ... and it happend more than once ... but did it in the same location or not? does anyone know?

---

1. It does when I use 9x multi (default).
2. CPU name stays the same if I increase multi.
3. Yes.

Answer: default multi is 9x.

I'm the owner of the CPU in Sampsas first post - ask me if you have anything to ask about that CPU.

Quote:

---

Originally Posted by s7e9h3n

It's not an ES - they don't look like that... ;)


Your comment made me think for a bit and you may be right, sortof :p: I looked over the cpuz submission again and I'm begining to think that this may be a 754 pin Clawhammer 3000+ with a default speed @ 10x200. It would have to be running on a board similar to the glitched Gigabyte board that someone had reported in a past thread. That board had locked his multplier higher than should have been possible for his cpu. BTW, IIRC, a Clawhammer 3000+'s cpuid is something like 0xF4A and that pic of the cpu may not be the same one shown in the cpu-z submission....

---

no way clawhammers 3000+ have 9x multi and 1mb l2, check my avatar
also, model on his cpu is 3 my claw is 4 so it's def not a claw