AMD PhenomII X3 710(720BE) 3core --> 4core PhenomII X4 920

[jcknouse]

most people are finding this out. Not just the koreans. Many over the world are.

Well for me:

a) if all goes well I'll have a new 2.8GHz Quad-core running on AM2+ that I can just get the AM3 board for later when the SLi-enabled mobos come out
b) i also get to experiment to see if i enable it on the Biostar board, remove it, put it in the ASRock, if it stays ACC Enabled.
c) I get to have fun

I work in software, and it gets monotonous sometimes. I like the variability of hardware and tweaking it and messing with mods. I just don't have time to become a "pro" at it or I'd work on re-learning all my physics from college and my electronics education too, and then join a pro-OCing team.

But, I now have the a) CPU, and b) Mobo...on order from Newegg. Got them as cheap as I could. Gotta file a $10 MIR when i get the mobo.

But, I have OCZ performance DDR2-6400 RAM and my ASRock Mobo is already AM3 Ready. So if this works getting it to work and then moving it to the other mobo, I might just start tweaking my memory and OCing it to see how much I can really push this CPU.

Plus, the 720BE is 95 watt whereas my 9850 is 125W. This should help with heat issues...I hope.

I will let you all know when I get it in, and when I do the BIOS change and if it works.

 

[jcknouse]

any new reports from any credible sources yet?

I'm working on it. I can't just get the stuff overnight, okay?

Edit: Tell you what, werxen...

You tell me what the tools are that I need to independently show you the clocking on each of the cores once/if it unlocks.

I already have prime95 on the PC, I'll use that to show it goes through iterations. I just need something to show what clock speed each core is running at.

Would that be CPU-Z?

Let me know. I will appease you as best I can.

But, I won't go out to buy a video camera. I already spent $270+ on this experiment. That's quite enough.

I'm not Best of Media Group. I don't have $1Ms

 

[chy18]

jcknouse

already korean o.c thought same as u

and he success core over(?) with asrock 780gx

after that he tried to install into another mobo..

which asus m78a-?? (i don't remember the mobo name exactly)

but it wasn't able to enable 1 more core in other mobo..

time after time,day by day more people would try to do it..

i envy u..i also wanna get fun with heca&mobo.. 😉

i hope u'll be able to success to do that..

and if u fail..don't worry about it..

i think just basic heca with tri-core..it also will show good perpormance..

add..if u success..u would get more fun with that 😉

 

[chy18]

but dudu don't forget it needs risk..

it depands on ur luck lol

 

[chy18]

i'm not to sure about it ..

but i think with cpu-z,prime95 it should be no problem..

i really wannt check it 4 jcknouse

unfountenely my favourite community site

is renewal in d/b..

so i couldn't find info 4 u..

sorry dudu..

if i know that i'll let u know 😉

i envy..u r pretty rich..

coz i have just $30..

lol

 

[duzcizgi]

Here's a small speculation of how this has happened:
These chips are black editions, which are sold with unlocked multiplier for OC purposes. So, although they are rated at a certain clock, indeed, they are expected to run at a higher speed.

When we recall what ACC does, it looks more relevant how this can happen:
ACC is moderating the timings and voltage for a given frequency, in order to keep the CPU stable when overclocked. Also, it can change the frequency of the cores independently.

In the AM2+ boards, ACC is in the southbridge and with PII, ACC is carried into the CPU itself.
When you enable ACC in the BIOS options, in fact, the southbridge ACC is taking over the ACC in CPU.

An easier way of disabling a core in ACC is, just putting a small flash memory in the CPU and write there which cores will be running. When ACC on southbridge takes over and not knowing about that memory area in the chip probing all 4 cores and all of them answering "here I am" it's turning them all on.

So, if this chip is expected to be over clocked and let's say rated 2 GHz and one of the cores can OC to only 2.2 GHz, then this core with lower OC overhead can be disabled as if defective, although it could run perfectly under rated speed.

As I told, this is just a speculation of what might be under the hood of this "phenomena"

Do you think I am too far off?

 

[chy18]

dudu that's good info.. i didn't know why it works..

after c ur reply i understand(not totally coz i'm not a engineer lol) why it works..

pretty interesting~info..

i'm really appericiate it 😉

sweet dudu..

 

[emesis166]

http://vr-zone.com/articles/board-makers-confirmed-phenom-ii-x3-unlocking/6629.html?doc=6629


game set

 

[rangers]

http://www.fudzilla.com/index.php?option=com_content&task=view&id=12224&Itemid=1


this is a very good peace of business by amd, there going to sell a **** load of CPUs of the back of this news and if you ask me it was intentional, but after they have won some market share of intel they'll fix it (but that's just business)

 

[NewManComing]

werxen, jcknouse // see the vr-zone's articles? U guys shut the hell up.

 

[jcknouse]

werxen, jcknouse // see the vr-zone's articles? U guys shut the hell up.

NewMan,

No need to get hostile or beligerent with me. I was wanting to find out, and I actually was willing to go to the expense of doing it for the rest of you.

Anyways, sorry if you thought I was being a pain. I just wanted to make everyone happy and prove/disprove if it could be unlocked...and whether the chip I got was flawed on the 4th core or not if it did work.

Thanks for being so friendly btw. Glad to see that someone like me who is willing to help prove/dispell myths for you is so appreciated.

 

[jcknouse]

Here's a small speculation of how this has happened:
These chips are black editions, which are sold with unlocked multiplier for OC purposes. So, although they are rated at a certain clock, indeed, they are expected to run at a higher speed.

When we recall what ACC does, it looks more relevant how this can happen:
ACC is moderating the timings and voltage for a given frequency, in order to keep the CPU stable when overclocked. Also, it can change the frequency of the cores independently.

In the AM2+ boards, ACC is in the southbridge and with PII, ACC is carried into the CPU itself.
When you enable ACC in the BIOS options, in fact, the southbridge ACC is taking over the ACC in CPU.

An easier way of disabling a core in ACC is, just putting a small flash memory in the CPU and write there which cores will be running. When ACC on southbridge takes over and not knowing about that memory area in the chip probing all 4 cores and all of them answering "here I am" it's turning them all on.

So, if this chip is expected to be over clocked and let's say rated 2 GHz and one of the cores can OC to only 2.2 GHz, then this core with lower OC overhead can be disabled as if defective, although it could run perfectly under rated speed.

As I told, this is just a speculation of what might be under the hood of this "phenomena"

Do you think I am too far off?

I'm not sure if you're far off or right on target. I'm not much of a hardware specialist. Not sure how the internals of CPUs work since the P5 class of processors from Intel. Once I got out of college, I didn't have time to read 4 magazines a week and surf the net for articles and white papers.

It sounds good tho. If you're right, this makes the AM2+ platform even more valuable because they won't go back and re-rev the spec on it being that it's so established. AM3 might suffer for this because "1.1" of it might be done to disallow the unlocking and make the 790xx class of AM2+ and AM3 priceless to get a hold of.

We'll just have to see tho. Thanks for that info tho. Appreciate it.

 

[B-Unit]

OK chy18, if you dont have one, WHY ARE YOU THE CHAMPION OF THIS!!!

Go away untill you have one to test yourself.

 

[werxen]

werxen, jcknouse // see the vr-zone's articles? U guys shut the hell up.

From what we know, only 0904 batches of Phenom II X3 can be unlocked. Basically, it just uses ACC core to trick the CPU to enable 4th core that has been classed unstable by AMD. However, not all unlocked Phenom II X3 will be stable so you will need some luck here.


i think you need to shut the hell up seeing as how you cannot even read the article.

 

[Silverion77]

wtf is "dudu"? Are u referring to the stuff that comes out ur butt?!?!

and i agree with B-Unit. U guys are just citing the same info over and over and yet have not had any experience with such a setup. So what it can be done?! U dont kno if its stable or anything. The Prime95 shots on the first page mean ABSOLUTELY nothing cause it just started. Wait for it to pass 8+hrs or turn Red and BSOD and then well discuss if its "stable" or not.

@chy: 1) Theres an edit button. Almost 3 out of every 5 posts here are urs and u can just edit it. 2) u dont need to double space everything. I do that on my School papers

 

[xx12amanxx]

Look's as if they planned this the whole time!....Damn just one reason not to buy into dual's anymore!

AMD is kicking like a ball strapped bull!

 

[freshspace]

From what we know, only 0904 batches of Phenom II X3 can be unlocked. Basically, it just uses ACC core to trick the CPU to enable 4th core that has been classed unstable by AMD. However, not all unlocked Phenom II X3 will be stable so you will need some luck here.


i think you need to shut the hell up seeing as how you cannot even read the article.

You are an IDIOT! You should know when to concede. I just read the whole thread, and right about now is when you say "You are correct"! You got your answer to your desire for "credible sources"!

 

[jcknouse]

One thing I thought of a bit ago:

A couple guys I remember working in the tech industry (one worked for a firm that built CD-ROM drives, the other worked for a firm that made video cards) back in the day told me something about their Q&A process.

The way both of them worked, they would pull a random, selected number of units out of a lot of items and test them. If x number of y items failed testing, they got rid of the entire lot of items.

Why? It is not cost effective for a company to test individually items that require timely testing, such has processors, CD-ROM/DVD drives, etc., that might take minutes to find some manufacturing flaw in the device.

So, they do something with them or trash them.

In AMD's case, it might be the case that AMD is taking entire lots of X4 CPUs and setting a core to disabled.

I can't be for sure, but I think this is probably the case.

Basic premise: AMD makes 1,000 processors. They pull 20 of them for testing. if more than 1 fails (more than 5%), they see that as a bad lot and send them to the X3 line to be sent through, 1 core disabled, and labelled as X3.

I think this is the likely case. It makes good business sense, as you minimize your cost of QA, your cost of manufacturing, and cost of processing returns of defective items.

But, it's just a guess. We can't know for sure until someone from AMD says so.

Until then, I'm waiting on my hardware to arrive. When I get the board and CPU installed, I'll get my Canon camera and take pics of the install and the CPU and the screens as I do it.

 

[werxen]

You are an IDIOT! You should know when to concede. I just read the whole thread, and right about now is when you say "You are correct"! You got your answer to your desire for "credible sources"!

hey dude shut up seriously you are annoying. 1) this article does not mean anything 2) it says its not stable 3) it says the ones testing were taken from random people testing it.

these claims were taken FROM PEOPLE. its the same people from korea who claim it works yet do not provide any evidence of it being stable aka running prime95 for more than 1 second. get your head out of your ass and open your eyes.

edit: enough time has been put out for a site to claim it works yet nothing has been confirmed. its like taking articles from blogs and claiming them as political facts when they are just opinions of opinions of opinions. shut up dude you know nothing so do not claim you do. i am looking for answers not speculation.

 

[fazers_on_stun]

There is one more possible reason: market demand.
Now that AMD has created a market for cheap tri-core cpus, even if it was because of poor yields and a failed 4th core in the beginning, they might not get enough failed x4s anymore to meet the demand and have chosen to disable perfectly good x4s to sell as a cheaper x3 instead. This has happened with gpus in the past...

While this tri-to-quad-trick seems too easy, I was one of those 'wth has southbridge to do with OCing' -guys when the sb750 and ACC first came out, so I guess anything is possible

Interesting idea about market demand, but couldn't that also mean the quads are overpriced? I mean, if they can still earn a profit by disabling a good core and selling it, then think of how many quads they could sell if they lowered the price to, say, halfway between the current price and that of the tri-core?

From a thread over on Roborat's blog, some of the process engineers theorize that it's less of a question of defect density (nonworking cores or L3), but more that a certain percentage of the quads won't bin into the TDP envelope. I guess AMD could have sold these at a lower speed grade however, if they wanted to.

Used to be that when you wanted to cut out a defective part of a chip, you blew fuses in the power traces and reconnected the inputs to ground. But you could effectively accomplish the same thing with a lookup table and pass transistors. So if you can get access to the table you can throw the switches back.

BTW, what is "dudu" - Korean for "dude" I hope

 

[jcknouse]

BTW, what is "dudu" - Korean for "dude" I hope

Me too, otherwise I would hate to know what I have been called today :lol:

Oh well, I've been called worse...by my own mother!!! :lol:

 

[blackpanther26]

hey dude shut up seriously you are annoying. 1) this article does not mean anything 2) it says its not stable 3) it says the ones testing were taken from random people testing it.

these claims were taken FROM PEOPLE. its the same people from korea who claim it works yet do not provide any evidence of it being stable aka running prime95 for more than 1 second. get your head out of your ass and open your eyes.

edit: enough time has been put out for a site to claim it works yet nothing has been confirmed. its like taking articles from blogs and claiming them as political facts when they are just opinions of opinions of opinions. shut up dude you know nothing so do not claim you do. i am looking for answers not speculation.

acually it dose not just say it is not stable it says it could be stable. There is a difference

 

[B-Unit]

One thing I thought of a bit ago:

A couple guys I remember working in the tech industry (one worked for a firm that built CD-ROM drives, the other worked for a firm that made video cards) back in the day told me something about their Q&A process.

The way both of them worked, they would pull a random, selected number of units out of a lot of items and test them. If x number of y items failed testing, they got rid of the entire lot of items.

Why? It is not cost effective for a company to test individually items that require timely testing, such has processors, CD-ROM/DVD drives, etc., that might take minutes to find some manufacturing flaw in the device.

So, they do something with them or trash them.

In AMD's case, it might be the case that AMD is taking entire lots of X4 CPUs and setting a core to disabled.

I can't be for sure, but I think this is probably the case.

Basic premise: AMD makes 1,000 processors. They pull 20 of them for testing. if more than 1 fails (more than 5%), they see that as a bad lot and send them to the X3 line to be sent through, 1 core disabled, and labelled as X3.

I think this is the likely case. It makes good business sense, as you minimize your cost of QA, your cost of manufacturing, and cost of processing returns of defective items.

But, it's just a guess. We can't know for sure until someone from AMD says so.

Until then, I'm waiting on my hardware to arrive. When I get the board and CPU installed, I'll get my Canon camera and take pics of the install and the CPU and the screens as I do it.

The theory sounds nice, but inorder to make it an X3 you have to identify the bad core to disable. I find it hard to belive in their current finacial situation that if a perfectly functioning X4 comes thru they go ahead and castrate it.

 

[roofus]

+1 B-Unit

 

[freshspace]

hey dude shut up seriously you are annoying. 1) this article does not mean anything 2) it says its not stable 3) it says the ones testing were taken from random people testing it.

these claims were taken FROM PEOPLE. its the same people from korea who claim it works yet do not provide any evidence of it being stable aka running prime95 for more than 1 second. get your head out of your ass and open your eyes.

edit: enough time has been put out for a site to claim it works yet nothing has been confirmed. its like taking articles from blogs and claiming them as political facts when they are just opinions of opinions of opinions. shut up dude you know nothing so do not claim you do. i am looking for answers not speculation.

You sound like someone who suffers from manic depression, "shut up, shut up...." Listen kid, if you have nothing to contribute, don't write anything at all. If your way to getting your answer is to oppose others who actually contribute, then maybe this forum is not for you.