Sandy Bridge 2500k Just Died

DBT85 · 10 Jan 2011 at 08:07

I ordered some OCZ gold 1.65 memory with my SB build but I've got some 1.5v Patriot stuff lying around (was from an inlaws PC that didn't like it) that i'll be using at first i think!

Another thing to consider is that it seems that at least some of the machgines OCUK are selling are using Corsair XMS3 4GB 2x2gb ram which is 1.6-1.7 rated and they have had no issues that we are aware of. It's still entirely possible that these are just manufacturing defects.

Garido · 10 Jan 2011 at 08:34

Are you sure it's not a faulty or finicky board as opposed to the chip? I used to have a Q6600 many moons ago. Would lock up on the first build. After 2 boards kept locking up. Thought it was the chip but as it turns out the store had a bad batch of boards (confirmed by the store).

EDIT: Store was not OCUK. I live across the ocean.

flopper · 10 Jan 2011 at 08:47

ton of people have run these chips at 1.6vcore, its not like they all suddenly died.
people also bricked the 980x really easy.

It can be so many things that cause this.

redrooster303 · 10 Jan 2011 at 08:54

Maybe when you start sticking the volts through the chip and it starts warming up more than usual thats when the higher memory voltages start making their mark.
When you think about it the prefered voltage for these chips is 1.55v-1.6v and people are using 1.65-1.7v kits from there older 1156-1366 builds. Now thats like using 1.75-1.8v ram on the 1156/1366 chips and then putting a massive overclock through them.
G-skill Eco for my 2600k I think, its only 1.35v, saying that though the ram volts might have nothing to do with it, though it does seem funny the IMC's are going.

EDIT: its gonna take some time for someone to pin down whats doing it I suppose, until then I'm going easy on mine lol.

Purgatory · 10 Jan 2011 at 08:56

flopper said:
ton of people have run these chips at 1.6vcore, its not like they all suddenly died.
people also bricked the 980x really easy.

It can be so many things that cause this.

1.6vcore

on a 32nm chip ??? That is insane and will damage them.

Max I would advise anyone to go on the vcore on a 2500 or 2600 sandy bridge is 1.35vcore any more and you are just asking for it to fry... Also at this moment I don't advise anyone to overclock thier chips untill motherboard companies get their BIOS working 100% ... Reading some reviews I keep finding that the only stable motherboards are the Intel boards for overclocking so far... It's just all too new and the companies that make the boards need to tune their BIOS and remove dangerous limits or atleast highlight when you are heading to dangerous limits..

***DON'T be running the vcore past 1.35v*** you might aswell kiss the cpu goodbye... Wait for the OCuk system builders to give you some overclocking tips and what settings they are using to get a stable safe overclock. I'm sure Nath or one of the guys will be nice enough soon to tell you what they are finding works and works in a safe manner without damaging the cpu.

Also I advise don't touch the BCLK for now ... Leave it at 100 and careful of motherboards that have the AUTO option, set it manually to 100 if you plan to overclock because some motherboard companies like to make their boards look fast and in AUTO they will raise it slightly.

SimonMaltby · 10 Jan 2011 at 09:02

Purgatory said:
Wait for the OCuk system builders to give you some overclocking tips and what settings they are using to get a stable safe overclock. I'm sure Nath or one of the guys will be nice enough soon to tell you what they are finding works and works in a safe manner without damaging the cpu.

Ah, you have so much faith

I think you may find that the OCUK system builders are watching the Forums and relying on the overclocking community for guidance LOL.

koolpc · 10 Jan 2011 at 09:07

Strokes my lovely AMD quad...................spits on Sandy!! lol

deusexlumina · 10 Jan 2011 at 09:09

Maybe they are just not tolerant to the heat/voltage that has been going through thiese chips. These are 32nm parts, and are therefore much more senistive to voltage than 45nm. The fact both people reported processors working at lower frequencies shows it may well be degradation.

Purgatory · 10 Jan 2011 at 09:13

SimonMaltby said:
Ah, you have so much faith
I think you may find that the OCUK system builders are watching the Forums and relying on the overclocking community for guidance LOL.

Simon it's better they fry the stuff there and figure out what works then a customer RMAing stuff back to them once a week... and then coming on the forums saying another dead chip....

NathWraith · 10 Jan 2011 at 09:13

SimonMaltby said:
Ah, you have so much faith
I think you may find that the OCUK system builders are watching the Forums and relying on the overclocking community for guidance LOL.

Ha, the cheek!

I would strongly recommend NOT going over 1.4 vcore (I mean over 1.4 vcore In windows) so in reality you want less than 1.375 in the BIOS. But be careful when changing the Vcore as the QPI scales with it so if you leave it on AUTO it will over compensate too much, so you have to be careful with the volts.

Especially when you are gunning for a 4.6GHz clock or higher!

Skyrocket · 10 Jan 2011 at 09:29

lollyhayes said:
Looks like my 2500k is dead... 4 days of use, now it doesn't even make it into windows without locking up...
Tried rolling back to stock - locks up on windows load up. Windows repair locks up.
Tried to use another HD and install windows on that. Locks up...
Looks like these might be duff, another guy with a d2 stepping 2500k had his die...
Any ideas on a way to test with only around 30sec before lockup when started?
Trying to underclock now...

how much voltage were you running to begin with?

I doubt that Memory VID is killing the chips as Sandy Bridge has probably the same mem controller as the 1156 CPU's. which are rated to 1.65.

delpy8 · 10 Jan 2011 at 09:34

Am I safe to use the corsair vengeance 1.5v 1600c9 with the
2500k and p67 deluxe asus

Durzel · 10 Jan 2011 at 09:37

This thread is getting linked from various places, but is there actually other reports of this? Only thread I can find on "2500k died" or variations is this one.

Also "was running a Vcore within the spec, well within the spec at 4.8ghz" sounds a bit clandestine. What vCore were you actually running?

SimonMaltby · 10 Jan 2011 at 09:41

delpy8 said:
Am I safe to use the corsair vengeance 1.5v 1600c9 with the
2500k and p67 deluxe asus

The Asus manual for the P8P67 states
"According to Intel CPU spec, DIMMs with voltage requirement over 1.65v may damage the CPU permanently. We recommend you install the DIMMs with the voltage requirement below 1.65v"

So why on earth can the Dimm voltage in the bios go up to 2.2v ?????
As I said earlier, I'm really worried that the Motherboard manufacturers have such 'wild' voltage ranges in the bios, with little warning or protection against serious overvolting.

I can understand it in the pure enthusiast boards like the Maximus, but the P8P67 is a general supply board and should really be well protected.

Purgatory · 10 Jan 2011 at 09:46

delpy8 said:
Am I safe to use the corsair vengeance 1.5v 1600c9 with the
2500k and p67 deluxe asus

yup

Garido · 10 Jan 2011 at 09:51

In all honesty, if a quad or a dual core got over 4.0Ghz I'd be stoked and leave it at that. Good on to those early adopters who pay a premium and explore the boundaries. I did that in 2007 with the AMD Spider platform and I told myself never again.

robj20 · 10 Jan 2011 at 09:52

Bit gutted really i was hoping to get Mushkin memory but the ones i want are all 1.65V not keen on running it so close to the wire. Think im going to hold of a few weeks to see how all this pans out.

DBT85 · 10 Jan 2011 at 10:01

It baffles me that if the memory needs to run at 1.5v why there arn't more kits at 1.5. Most that anyone is actually going to want are 1.6-1.7.

Vertigo · 10 Jan 2011 at 10:04

SimonMaltby said:
The Asus manual for the P8P67 states
"According to Intel CPU spec, DIMMs with voltage requirement over 1.65v may damage the CPU permanently. We recommend you install the DIMMs with the voltage requirement below 1.65v"

Am I going to be OK using standard Corsair XMS stuff then?

OcUK shows the voltage for this memory as 1.6-1.7v but I've seen it listed as 1.65v elsewhere. 1.5v rated stuff seems to cost a load more.

Durzel · 10 Jan 2011 at 10:09

SimonMaltby said:
The Asus manual for the P8P67 states
"According to Intel CPU spec, DIMMs with voltage requirement over 1.65v may damage the CPU permanently. We recommend you install the DIMMs with the voltage requirement below 1.65v"

So why on earth can the Dimm voltage in the bios go up to 2.2v ?????
As I said earlier, I'm really worried that the Motherboard manufacturers have such 'wild' voltage ranges in the bios, with little warning or protection against serious overvolting.

I can understand it in the pure enthusiast boards like the Maximus, but the P8P67 is a general supply board and should really be well protected.

Yeah I find this weird too.

If as you say there are no safeguards in place and you can (for example) select 1.9v for the CPU, then that will fairly obvious nuke it the moment you turn the PC on. Same with DIMM voltage too I guess.

I never actually realised that vDIMM voltage played any part in CPU lifespan, is this something new to Sandybridge or have I been living in a cave?