KILL your i7 980X AT ONCE (and RMA it)...

[Ice009]

So what's the deal here. Are you guys able to get an RMA on your processors? I was going to buy one, but this is putting me off. I'm not a bencher and was only going to use it on Air Cooling so I'm thinking it should be OK?

If I've got this right, don't disable any cores or HT?

Has anyone worked out the actual reason? When people are using only 1 core is that core getting the total voltage you had set for 6 cores to that 1 single core?

[Erklat]

There is no such thing as 'total' voltage. Eg. 1,225V is not divided by the number of cores, but instead every core
gets the same vCore you set in the BIOS.

[Ice009]

There is no such thing as 'total' voltage. Eg. 1,225V is not divided by the number of cores, but instead every core
gets the same vCore you set in the BIOS.

Yeah that is what I thought until other people in the thread confused me. Either way even if you run it on stock volts with only 1 core it still stuffs up right? If that is correct then Intel should RMA those processors.

[BenchZowner]

It's definitely not about voltage, and of course the voltage IS NOT divided by the core count
As for the issue, some CPUs have it ( e.g. die immediately/fast/relatively fast with disabled cores/ht to 1/2 ) and some don't ( some CPUs can simply run 1C/1T for ages ).
Nobody knows why it happens or the source, not even if Intel is investigating this at all.

You can get a RMA for your CPU for sure ( no way to detect it AFAIK ).

[xpatar]

Correct, the voltage is not divided by the core count.
I've run with 1C/1T either @ stock or overclocked (within Intel's specifications safe limits) and my cpu didn't die.
I didn't even see any sign of degradation in the performance of the cpu.
So definately the problem doesn't apply to all Gulftowns.
Soon I'll change mb and continue testing.
I'll post when I'll have any new results.

[Donnie27]

Read the first two pages of this thread before questioning the whether the RMAs were legit or not.

CPU dying on stock by only shutting a couple of cores off sure as hell sounds as legitimate RMA to me.
As does to all guys who got the CPUs which suffer from the same phenomenon. Whether they applied excessive voltage
is irrelevant. They were unable to shut cores off from the beginning, therefore it limited the CPUs functionality, regardless
of whether disabling any of the cores seems reasonable or not.

So what about all the guys who are NOT having any problems, you think they're fibbing?

I did read the first two pages, then the whole thread before I posted. We are NOT talking about processors dying from NORMAL use. These processors were killed so they DON'T really qualify for a RMA. What was going here to start this thread isn't close to what you wrote or what the thread is even about. These are NOT legit reasons for a VAR or etc... to RMA not according to Donnie27 but anyone selling and buying should already know and FULLY understand. Something being abused this way is NOT covered by any VAR. EVERYONE is warned if run products out of spec and overclock, you do so at your OWN RISK, you void the muth%#$&ing warranty

The only legit question is if Single Threaded apps caused a problem or not. That was the only reason I read the whole thread before I posted. If single threaded apps with features (power states-speed step and etc...) turned on is causing problems then there should be a Fix.

The core can't perform correctly if you follow many of the overclocking guides that tell you to disable IMPORTANT features and or settings. These very settings keep these kinds of things from happening, but I bet you think that is irrelevant as well, right?

[sniper_sung]

Now I am scared of enabling C-state of the 980X because it will perhaps close some cores under single thread applications...

[Gautam]

So, anyone kill any chips WITHOUT first running them under extreme conditions?? If someone reports a CPU at stock dying with cores disabled, that's one thing, but all I'm seeing here is people pushing the chips to points where they'd already die, and then disabling cores after the fact and then blaming it on that...

[Erklat]

The core can't perform correctly if you follow many of the overclocking guides that tell you to disable the IMPORTANT features and or settings. These very settings keep these kinds of things from happening, but I bet you think that is irrelevant as well, right?

So you are proposing that disabling c-states, speedstep and the rest of advanced features induce malfunctions like this? Well that's new.
Guess that's why they come disabled by default when I load system optimized settings

So, anyone kill any chips WITHOUT first running them under extreme conditions??

First post. CPU pwnd via telephone on stock

[BenchZowner]

I will do that on first POST with my next new 980X Gautam.

[Splave]

im game as well, rma back tomorrow

[xpatar]

After testing with 1C/1T on an ASUS P6T DELUXE V2, I continued testing on a Gigabyte GA-EX58-UD4P (rev. 1.0), bios version F13.
Again in this test I didn't exceed Intel's specifications safe limits, which means a maximum Vcore of 1.375v. With this way I wanted to eliminate the risk of having a dead cpu due to extreme overvolting but also to check if what Intel states as safe limits, are really safe. Current setup under water.


1M SUPER PI 1C/1T @ Default





1M SUPER PI 1C/1T @ 4.53 GHz Vcore 1.35v





32M SUPER PI 1C/1T @ 4.53 GHz Vcore 1.35625v ( a small increase in Vcore to keep it stable)





PRIME 95 (1 hour) 1C/1T @ 4.53 GHz Vcore 1.35625v





After finishing I restarted, set everything @ default,booted into windows and ran Cinebech in order to check if there was any sign of degradation.



Again not any sign of degradation.

So testing with two different boards doesn't make any difference and also the problem doesn't apply to all Gulftowns.
I still have a healthy, working cpu either with one or all cores enabled.

[Donnie27]

So you are proposing that disabling c-states, speedstep and the rest of advanced features induce malfunctions like this? Well that's new.

Guess that's why they come disabled by default when I load system optimized settings

First post. CPU pwnd via telephone on stock

The core can't perform correctly if you follow many of the overclocking guides that tell you to disable the IMPORTANT features and or settings.

Turning off power gating tech and if the Voltage is left high from overclocking Previously, stock MHz speed is MOOT! I'm proposing NOT turning off Speedstep and Other tech that are features most folks paid to have. I think I need to edit the earlier post

I asked was the problem caused at STOCK settings? I mean ALL STOCK settings with those features enabled.

When I used to overclock and I changed a Voltage, I'd change the voltage, save, reboot to enable that voltage then re-setup for stock settings.

I only wondered if the BIOS was set up in stock mode. Then the processor kicked into Turbo with one core running one thread, then it died? If that's the Case Intel needs to step in. NOT because a few of us Geeks created an unreal scenario. Why in the hell would I buy a 6 core Processor to disable 4 or 5 cores?? Oh brother, I'd be soooo worrriieed if I bought one, NOT!

I know someone who killed a Q9550 playing with cores and lowering the voltages(no overclock), at least he claim that, never saw a proof...
This topic is some kind of confirmation that something is wrong in the quad/six core CPU's design...
I have no idea what, but for sure I will never disable any core starting today.
Better then that I will disable that option in BIOS to never be tempted LOL

The key phrase is in BOLD. The folks I disagree with are saying it died, I'm saying THEY KILLED IT. My old 9550 my wife's using hasn't even sniffled yet! I don't know anyone talking about problems with these CPUs while running them not only at stock speed, but many over 4GHz. IMHO, there's too much of the "I need 4.5+" to be happy folks" killing all kinds of Processors. Then there has to be some FUD gorilla marketing tied in as well. If I'm an i980X owner, it doesn't even cross my mind to try this or anything else as LAME! No company should be held responsible for these kinds of *tweaks.

[Ccaution]

So... to sum it up:

It's motherboard independent and it seems as it is cpu relevant only... well, did someone said "subthreshold leakage"

[Erklat]

Turning off power gating tech and if the Voltage is left high from overclocking Previously, stock MHz speed is MOOT! I'm proposing NOT turning off Speedstep and Other tech that are features most folks paid to have. I think I need to edit the earlier post

I asked was the problem caused at STOCK settings? I mean ALL STOCK settings with those features enabled.

So you have two types of CPUs:

1.) the ones who survive disabling the cores both stock and overclocked
2.) the ones who die both on stock and overclocked.

We don't have a single one report of CPUs surviving on stock but then dying
when overclocked. I can hardly be led to believe the overclocking is an issue here that causes them to malfunction.

Both power gating tech and disabling of the cores should be put at users disposal. I might wish to disable the couple of the cores to
reduce power consumption when doing some less CPU intensive tasks,
I don't care. Just because it seems like an unreasonable scenario I don't expect it to die.

Same goes for the power gating tech. Seems Intel doesn't have any grudge against users disabling it for years now.

[OBR]

There is few DEAD chips over Gigabyte GOOC 2010 EU final competition ... lot of degradation, few DEAD CPUs.

[shadewither]

There is few DEAD chips over Gigabyte GOOC 2010 EU final competition ... lot of degradation, few DEAD CPUs.

Who seems like the murderer? MB? Volt(Current)?

[BenchZowner]

Tried this on 2 980X's on 2 mobos.

Sofos1990's 3013A540:
eVGA X58 SLI Classified E760 = would work and boot into windows at both 1c1t & 2c2t just fine.
Rampage III Extreme = would work and boot into windows at both 1c1t & 2c2t just fine.

my 3020A457:
eVGA X58 SLI Classified E760 = would work and boot into windows at 2c2t just fine, would POST but hang before getting into the BIOS or at the "launching operating system" screen at 1c1t.
Rampage III Extreme = would work and boot into windows at 2c2t just fine, would POST but hang before getting into the BIOS or at the "launching operating system" screen at 1c1t.

None of them showed any signs of degradation, or died during those tests.

[Donnie27]

So you have two types of CPUs:

1.) the ones who survive disabling the cores both stock and overclocked
2.) the ones who die both on stock and overclocked.

We don't have a single one report of CPUs surviving on stock but then dying
when overclocked. I can hardly be led to believe the overclocking is an issue here that causes them to malfunction.

Both power gating tech and disabling of the cores should be put at users disposal. I might wish to disable the couple of the cores to
reduce power consumption when doing some less CPU intensive tasks,
I don't care. Just because it seems like an unreasonable scenario I don't expect it to die.

Same goes for the power gating tech. Seems Intel doesn't have any grudge against users disabling it for years now.

Again I'm not sure of what combination KILLED the processor. The key word is KILLED here. The processor didn't die is all I was saying. There is no malfunction when you're creating the problem

No, but he does~!

Tried this on 2 980X's on 2 mobos.

Sofos1990's 3013A540:
eVGA X58 SLI Classified E760 = would work and boot into windows at both 1c1t & 2c2t just fine.
Rampage III Extreme = would work and boot into windows at both 1c1t & 2c2t just fine.

my 3020A457:
eVGA X58 SLI Classified E760 = would work and boot into windows at 2c2t just fine, would POST but hang before getting into the BIOS or at the "launching operating system" screen at 1c1t.
Rampage III Extreme = would work and boot into windows at 2c2t just fine, would POST but hang before getting into the BIOS or at the "launching operating system" screen at 1c1t.

None of them showed any signs of degradation, or died during those tests.

why you tried not only cores coupling ? Crazy bug

Hehehehe! Weak attempt at FUD LOL!

[BenchZowner]

Again I'm not sure of what combination KILLED the processor. The key word is KILLED here. The processor didn't die is all I was saying. There is no malfunction when you're creating the problem

No, but he does~!





Hehehehe! Weak attempt at FUD LOL!

It's not FUD man.
I've seen one getting killed just by saving & exiting the BIOS with 1c1t.

My chip won't boot/work at 1c1t, so there's indeed something wrong with some chips at least.

[Bun-Bun]

I have found this thread incredibly amusing (though sucks to those who lost some processors!)

But I have to say something to Donnie27...

Regardless of whether a chip has been OC'ed or not there is an apparent problem with at least "some" of these chips and/or MB combinations. In the first post one died that was never overclocked and was on air cooling. The ability to disable certain features and cores is something that is availible in most to all MB's and is in the Intel spec. No where does it state from Intel that disabling cores or features is out of the norm or warranty voiding.

If a chip was never OC'ed and never taken out of spec and simply had cores disabled (which is the case of the first post over the phone)... and then died? Definite defect. And who knows what else could be wrong?

If you are going to go as far as saying that changing a setting in bios (disabling a feature for example) is something that is not covered under warranty... then the Intel warranty is null and void under every condition in which the system was not built and configured by an authroized Intel tech.

That thought is simply ludicris.

[Donnie27]

I have found this thread incredibly amusing (though sucks to those who lost some processors!)

But I have to say something to Donnie27...

If you are going to go as far as saying that changing a setting in bios (disabling a feature for example) is something that is not covered under warranty... then the Intel warranty is null and void under every condition in which the system was not built and configured by an authroized Intel tech.

That thought is simply ludicris.

Link to Intel saying running 1 core is alright please?

It's not FUD man.
I've seen one getting killed just by saving & exiting the BIOS with 1c1t.

My chip won't boot/work at 1c1t, so there's indeed something wrong with some chips at least.

I didn't say what you said was FUD. This is;

why you tried not only cores coupling ? Crazy bug

I should have used a separator or something, sorry my bad!

Again, the system is not changing the settings, USERS ARE! I know 10 folks using this processor, 6 have them overclocked. Not one has died but none of the users are disabling cores. Now if any of those folks not doing this had any of the processors Die, I'd join in on this.

The ability to disable certain features and cores is something that is availible in most to all MB's and is in the Intel spec. No where does it state from Intel that disabling cores or features is out of the norm or warranty voiding.

So if I gave my processor too much voltage (available), and too high of a multiplier (unlocked version that 980X is) (available) and killed it, then that's Intel's fault as well? OTOH, maybe Intel should warn folks not to do this"* so I'm NOT totally opposed to what you're saying. **This = disable all but one core.

1 Warning: altering clock frequency and/or voltage may (i) reduce system stability and useful life of the system and processor; (ii) cause the processor and other system components to fail; (iii) cause reductions in system performance; (iv) cause additional damage; and (v) affect system data integrity. Intel has not tested, and does not warranty, the operation of the processor beyond its specifications

Maybe motherboard BIOS makers should add a warning in RED when these problem causing settings are used or remove them. All I'm saying is that the action that caused this problem isn't likely to affect 99.998% of folks using this processor. They are not dieing, they are being killed=p

Here's what started some of this;

As does to all guys who got the CPUs which suffer from the same phenomenon. Whether they applied excessive voltage is irrelevant.
and
CPU dying on stock by only shutting a couple of cores off

He's not talking about shutting down all but one core. He states that even shutting down two cores as well. There is not a Phenomenon because 1/999,980th of the one million or so 980x sold will have someone disabling all but one Core IMHO. Now if this was something the OS or an app was causing, then I'd go along with you guys!

[Bun-Bun]

Doesn't matter if it is software or a bios change that caused it. It is still a defect.

By your logic going into bios to apply proper voltages and other various settings because AUTO incorrectly does so (esspecially for RAM) is warranty voiding.

I am sorry Donnie but you are making yourself look like a fool.

Disabling cores is not taking a processor out of spec. Hyperthreading is a feature known to cause issues in certain situations and can be disabled (some OEM's disable it by default or hard lock it off in their BIOS). Are you going to tell me that if my processor died with HT off that I killed it? Disabling cores is no different. Esspecially when Intel's own features do this as a power save feature.

If I disable unused integrated devices (fireware, LAN, etc) and something goes wrong; I killed it then as well?

Link to Intel saying running 1 core is alright please?

I am not here to replace google. Various documentation states this is an option and no where does it state it is warranty voiding nor out of "spec".

Really everyone should be asking you to provide a link to Intel saying that disabling cores is bad and warranty voiding.

[Glow9]

Curious I know they ran 1 in this test they seemed okay unless they didn't really. Maybe it's certain chips http://www.bit-tech.net/hardware/cpu...rcraft-2-use/1

[Toolius]

Ok.. was running stock @ 1.14 volts... and then i overclocked to running 25x160 at only 1.20v Cpu and 1.25v Vtt , VDIMM 1.62 on UD3R and Cpu was fine..
Moved it to UD7 and i cant load windows no more... BSOD saying Unrecoverable Hardware Error. Then Tried it on a RIIE as well.. Same BSOD. This thing just died !!
Moved back to my 920's and both systems function perfectly so i know its not any busted hardware on either system.
Batch 3005F584.. Now im sad
Never tried disabling any cores or anything. The highest this chip has seen is 1.27v Vcore and 1.25v VTT