GA-MA785 is it broken?

[scamper2]

after attempting to unlock my x2 550 to a quad core, my computer restarted but now it just has a black screen when turned on.

 

[bilbat]

OK - as this is about the fifth post about the same thing - core 'unlocking', I feel obligated to finally post something that someone may read in advance of screwing around with it... Unless you value your time at absolutely nothing, if you want a four core CPU, pay for a four core CPU - you'll be far ahead in the long run. The reason for these processors in the first place, is it allows AMD to improve their 'yield', i.e., they get more 'saleable' processors per wafer, albeit they sell them more cheaply. There was a reason, during testing after 'the slice', that cores were disabled - they don't meet spec! At Intel, they go on the floor; at AMD, they go in a box, and get shipped to some 'hopeful', who guesses he might 'rehabilitate' something that's 'broken' in the first place! You may be able to 'turn them on', and you may or may not have stability problems ever after - why mess with it? It's like marrying a drunk, in the hope of 'fixing' their problems!

I would never buy an AMD setup in the first place, even for the most mundane task... IMHO, Intel makes better, more 'tightly' integrated chips/chipsets, with infinitely better overclockability - but, that aside, my main reason is documentation. If you need to know something about Intel system components, there's literally gigabytes of information, easily and publicly available; I recently set about to learn about the i7/x58, and dl'd thousands of pages of documentation - which makes the task finding what's important to your needs - but it's all there! On the other hand, I recently undertook to find out what the actual hardware situation is for the ganged/unganged memory question for AMDs, as well as the actual reason for the DIMM count limits on high-speed memory - searched and scrounged for two whole days, read nearly three-hundred pages of 'semi-randomly' distributed documentation - very little up-to-date - and never learned anything really relevant!

 

[scamper2]

OK haha I fixed it (reset the CMOS). I have no problem with AMD selling quad cores as duo's because it does give you a CHANCE that you can unlock them (many people are successful at it) I was just unlucky. "infinitely better overclockability" give me an example of a CPU that is better than an x2 550 at overclocking for the same price? (and by "overclocking" i dont mean how much it can be overclocked, i mean how much MORE it can be overclocked than the 550.)

 

[bilbat]

because it does give you a CHANCE that you can unlock them

I guess, at heart, I'm just not a gambler - I'm more of a "belt, suspenders, and a safety pin" type!

 

[JofaMang]

3 out of 4 friends have Prime95 stable quad cores from 550s, the 4th wasn't stable. I have unlocked the core 4th core on this CPU (x3 710) on another mobo, but didn't have time to prime95 test it. I did have time to play some GTAIV though, and it ran it very well for the 30mins I was stress testing it(x4@3.2ghz). Only anecdotal, true, but follows the trend. I wouldn't rip on AMD for this situation, nor use it as a launching pad for an intel devotee pledge of allegience. Let people who play, have their play. Different manufacturing strategies between companies is fair to expect, and in this case, beneficial to the enthusiast who is willing to give it a shot.

How is this any different than cranking an i7 920 well beyond 975 clock speeds? Unlocked potential, squirreled away by the manufacturer to meet production requirements.

 

[bilbat]

Not meant as a 'rip'; simply a (cogent, I hope) observation, to help people avoid obvious problems - and obviously, everyone is certainly welcome to go to hell in their own chosen handbasket! This is simply one of those posts that I see over and over again, like flashing with @BIOS, or OCing with EasyTune. They are situations that I liken to deliberately shooting yourself in the foot, and then complaining about having to change the bandage twice a week!

I get large numbers of questions here about problems/peculiarities with AMD equipment that I cannot find 'firm' answers for, as the documentation simply seems to not exist. Anecdotal information is all well and good, but it too often turns out to be apocryphal - there is nothing like canonical info: "incrementing register 0x0002FE results in a .0025 volt per increment increase at pin 17dd..." I espouse thorough planning before undertaking a build: read all the manuals and docs, find all the 'gotchas', before you buy the stuff - it falls in with the definition of 'too late'; it's never 'too late' until it's 'too late', and then, well - it's too late! I try to make as much of the info I have, or know, available here in 'searchable' form, so it can be considered by anyone researching a GB build, and I would feel remiss to not comment on these recurrent problems that crop up over and over and over again...

 

[JofaMang]

Fair enough, but you have to admit, your attention to detail goes far and beyond even what the top 1% of tech savvy home builders can muster, or even understand (myself included, tbh). To even the above average builder, technical details of that level are meaningless, and brand loyalty is built through their independent experiences with the products, much the same as yours seems to be developed through your access to documentation. I am not inferring that you are a fanboy, don't get me wrong, I respect your explanation, but the vast majority of us do not go that in depth when choosing our PC configurations, or how to modify them.

As for the unlocking cores specifically, I do agree in one facet: don't buy a dual core depending on it being a stable unlocked quad. I do disagree with your assertation that ALL duals are rejected quads. Simple economics: cheaper products sell at higher volumes. Now I'll pull some numbers out of my ass for the sake of giving an example: if 30% of the chips are good quads, but only 15% of total sales are quads, do you honestly expect AMD to sit on that other 15% per batch and NOT sell them in one form or another? Add into the theory, that manufacturing processes improve over time, and the % of solid quads will increase over time, and you can see how, if AMD is not taking physical actions to disable the other cores (which they aren't) that the number of safe unlocks will only increase over time.

This is a boutique function, and I am sure the decision to not etch the extra cores is both production cost conscientious, as well as a pretty successful, if accidental, grass roots PR campaign.

 

[bilbat]

As for the unlocking cores specifically, I do agree in one facet: don't buy a dual core depending on it being a stable unlocked quad. I do disagree with your assertation that ALL duals are rejected quads. Simple economics: cheaper products sell at higher volumes. Now I'll pull some numbers out of my ass for the sake of giving an example: if 30% of the chips are good quads, but only 15% of total sales are quads, do you honestly expect AMD to sit on that other 15% per batch and NOT sell them in one form or another? Add into the theory, that manufacturing processes improve over time, and the % of solid quads will increase over time, and you can see how, if AMD is not taking physical actions to disable the other cores (which they aren't) that the number of safe unlocks will only increase over time.

This is a boutique function, and I am sure the decision to not etch the extra cores is both production cost conscientious, as well as a pretty successful, if accidental, grass roots PR campaign.

You know - I never thought of any of that - and it makes a huge amount of sense, in every direction. Dammit, as an engineer, I hate it when the marketing guys actually know something - that's why the marketing guys never know anything - when an engineer is asked (ordered, most likely, he'd never do it of his own free will!) to talk to the marketing guys, it's pretty much the equivalent of being told to wear a 'crown of thorns' for an afternoon...

Fair enough, but you have to admit, your attention to detail goes far and beyond even what the top 1% of tech savvy home builders can muster, or even understand (myself included, tbh). To even the above average builder, technical details of that level are meaningless, and brand loyalty is built through their independent experiences with the products, much the same as yours seems to be developed through your access to documentation. I am not inferring that you are a fanboy, don't get me wrong, I respect your explanation, but the vast majority of us do not go that in depth when choosing our PC configurations, or how to modify them.

I know - that's why, due to my experience and proclivities, I kind of consider it my function here - to sort through issues from the 'technical end' (when possible), and present them as clearly as I can. Frankly, a lot of what I do is to simply read the manuals that most users seem to discard along with the shipping box - I have more than eighty on file here! Before my last build, for example, I determined that to run RAID1 on WD 1Tb drives, I had to pop for the RE3s, as they charge more to turn off a firmware feature ('deep recovery cycling') that 'breaks' RAIDs. No one really needs to understand why, just to know, in advance, that a Caviar Black won't work! But the key term is before - that's when you wanna know! You might take a look at this memory 'write up' as an example:
http/www.tomshardware.com/forum/261306-30-your-thought-gigabyte-ep45-ud3p-mobo
I've posted it a number of times:
http/www.tomshardware.com/forum/261471-30-1066-motherboard
http/www.tomshardware.com/forum/260696-30-ep45-ud3r-opinion-needed
http/www.tomshardware.com/forum/260042-30-choose-correct-memory
sometimes with minor modifications, to address people's immediate concerns...

 

[JofaMang]

Once upon a Time, In a Land far far away, I was a marketing guy running off a Journalism Degree. We each have our forte.

 

[bilbat]

:lol: :lol: :lol:
...like putting a snake and a mongoose in a bag!

 

[JofaMang]

PSST: We hate talking to the engineering staff, too. Always too much recipe, not enough Cake, lol.

 

[bilbat]

Always too much recipe, not enough Cake

That's not a problem, that's our strategy: we just 'talk calculus' for a while, and throw in a few 'slew rates' and 'avalanche breakdowns', hoping the whole audience will just drift away, and allow us to sneak back to our beloved cubicle!

 

[Mongox]

Well, I'm happy I spent the extra $10 for a AMD 550 BE - I got a pushable CPU that might have quad cores. And mine is stable as can be - heck, decision is how much overclocking to do on it.

I've used AMD now for at least 15 yrs and been happy across the board.

 

[SystemPat]

Well I went and looked for the freebee upgrade but didn't get it. My 550 BE would not unlock and come up stable, even using the -12% adjustments on the 3-4 cores. But I was never counting on it. However it is now running at 3.6 ghz, up from 3.1. On the stock cooler with no voltage adjustment. And the temperature hardly moved at all. So I'm happy with it.

 

[Mongox]

I'm testing now between 3.6 & 3.7 and waiting for my new RAM. My CPU actually runs a couple of C cooler with the 4 cores un-locked, go figure.

But my temps with stock cooler would hit the low 70s running Prime95 and I didn't feel comfortable with that - even without any OC involved. Got a inexpensive cooler and doesn't go over 62C now when OC'd.