Sandybridge Intel i7-2700K & i7-2600K

[JohnG7]

That looks like a good chip! People used to go up to 1.50v for 5ghz. I had my 2500k at 1.45v for years on a dark rock pro cooler. As others have said, an overclockers dream

I cannot remember my settings for the the 2700k but it was not north of 1.45v for 5GHZ and nor is the 4790k now (would need look in bios for exact setting/reading).

 

[JohnG7]

I only jumped to Haswell due to the PCI-E 2.0 limitation on my Rog Mobo (some brands got updated I believe) plus the upgrade itch (I have grown out of it now).

I got a bit of stick for leaving a 5GHZ SB to go to a 4.6GHZ 4770k but the 5GHZ 4790k ended up as good and I am still on it since 2013 as nothing really worth upgrading to till Intel get their finger out.

 

[bulldog147]

I did notice AMD Ryzen was crazy start off with 1.4v as a stock speed. Intel never do this way. Why would AMD Ryzen pushed too far with 1.4v are they try to stop peoples overclocking it?

 

[JohnG7]

Do not know but not same architecture so cannot be compared and 1.4v may be normal for that chip but I CBA with having to mess with a ram calc tool to see if my Memory will work or cause issues etc so I will sit on this till 2022-23.

 

[bulldog147]

Here is my Sandybridge i7-2700K @ 4.5GHz

 

[JazzRookie]

I'm currently running my SB 2700k @4.7Ghz with 1.3661v.
I'll be upgrading shortly but will keep it as I cannot get rid of it. It has served me well for nearly 10 years.

 

[bulldog147]

I'm currently running my SB 2700k @4.7Ghz with 1.3661v.
I'll be upgrading shortly but will keep it as I cannot get rid of it. It has served me well for nearly 10 years.

That's pretty great chip u got there

 

[turbot1984]

Yes wish this sandybridge support PCIE 3.0, USB 3.1, Type C, NVMe slot, 8 cores 16 threads and DDR4 Ram in our dream, because if it does ticked it all, I would stay with Sandybridge for as long it take!

Tough moving on from great cpu's but got to say Skylakes like 6700k were a big upgrade from S Bridge. The same will happen when Intel give us a proper next gen, maybe 12xxx series?

 

[JazzRookie]

That's pretty great chip u got there

Yes you could not go wrong with any SB chip at the time. They were bullet proof.
I upgraded back then from a Core 2 Quad Q6600 @ 3.2Ghz

 

[bulldog147]

My 4.8GHz seem very stable when full load it reduced vcore to 1.344v with max temp no more than 65C on all cores. Ran Cinebench both R20 and R23. Ran prime95 stress for 30 minutes and intel in burn test for 10 minutes all seem stable. Happy overall.

Cpu-z: https://valid.x86.fr/w4f0a9

Results:

Cinebench R23:
Multicore 4432 points
Singlecore 919 points

Cinebench R20:
Multicore 1734 points
Singlecore 355 points

 

[DarkBahamut]

IMHO, I don't think ~1.35 V is really safe for Sandybridge chips. However the word safe needs some context. It depends how long you want it to last.

In the very long term it does degrade. I ran a 2700k at 4.7GHz 1.34 V for around 10 years before moving to a 5950X a couple of months ago. In the last year to 18 months the 2700k was notably degrading. The once flawless chip would start to see a few BSODs here and then. A slight nudge up of vcore would fix them for a few more months, then they'd come back. By the time I pulled the system I had vcore up to 1.37 V to remain stable. Still ok-ish, but it's clear the CPU was past it's better days. If I ever put it back into service I'll almost certainly have to drop the clocks a little to feel comfortable.

So if you want it to last 'forever' I'd stay below 1.3 V, if you just want a couple more years from it, then it's probably fine.

 

[JohnG7]

I would not worry if 1.4v or bellow and how much time does it spend at this voltage as it clocks down to lower clocks in stages and same with voltage, I think intel said 1.37v was getting there (look at the VID).

 

[MartinPrince]

IMHO, I don't think ~1.35 V is really safe for Sandybridge chips. However the word safe needs some context. It depends how long you want it to last...

1.35v is safe as it used to be listed on the official Intel 2600K/2700K spec sheet. The problem is that what many people do (maybe you did as well) is to run constant manual voltage, which the CPU was not designed to do. This is when you start to see degradation after many years. They will enable Speedstep which will drop the speed of the processor when idling but the voltage remains constant and does not drop.

People should always switch to offset voltage after first finding out what the stable voltage is using manual constant voltage.

The other contributory factor is heat. So in tandem with the constant 1.35v run at relatively high temps, (in part due to the constant manual voltage) it will result in degradation after several years.

I had a 2600K running at 5Ghz at 1.38v for ~10 years. It is in a friends system still running at that speed with no problems but this is using offset voltage so the voltage drops significantly when fairly idle, plus it is under a Noctua D14 with the thermal paste changed at least every couple years so it has been kept pretty cool.

 

[DarkBahamut]

1.35v is safe as it used to be listed on the official Intel 2600K/2700K spec sheet. The problem is that what many people do (maybe you did as well) is to run constant manual voltage, which the CPU was not designed to do. This is when you start to see degradation after many years. They will enable Speedstep which will drop the speed of the processor when idling but the voltage remains constant and does not drop.

People should always switch to offset voltage after first finding out what the stable voltage is using manual constant voltage.

The other contributory factor is heat. So in tandem with the constant 1.35v run at relatively high temps, (in part due to the constant manual voltage) it will result in degradation after several years.

I had a 2600K running at 5Ghz at 1.38v for ~10 years. It is in a friends system still running at that speed with no problems but this is using offset voltage so the voltage drops significantly when fairly idle, plus it is under a Noctua D14 with the thermal paste changed at least every couple years so it has been kept pretty cool.

It's been a while since I've seen it, but from what I remember of the Intel data sheet it just simply lists the VID range programmed/available. It doesn't actually say any 'safe' values. It's very common for opp points to be coded in which are not stable/not safe/not to be used, you see it in most CPU code.

It why most people used to say 1.5 V was safe too. Because the VID values go up to ~1.52 V. Loads of threads claiming >1.5 V was fair game if you look back. I wonder how many of those CPUs are in good order today.

If your 2600k hasn't suffered any then that's great, but mine certainly did. 5 Ghz @ 1.38 V is definitely a gold sample chip for Sandy, most were at 1.45 V for that sort of clocks. Possible the good silicon quality contributes to lower degrading over time.

 

[bulldog147]

IMHO, I don't think ~1.35 V is really safe for Sandybridge chips. However the word safe needs some context. It depends how long you want it to last.

In the very long term it does degrade. I ran a 2700k at 4.7GHz 1.34 V for around 10 years before moving to a 5950X a couple of months ago. In the last year to 18 months the 2700k was notably degrading. The once flawless chip would start to see a few BSODs here and then. A slight nudge up of vcore would fix them for a few more months, then they'd come back. By the time I pulled the system I had vcore up to 1.37 V to remain stable. Still ok-ish, but it's clear the CPU was past it's better days. If I ever put it back into service I'll almost certainly have to drop the clocks a little to feel comfortable.

So if you want it to last 'forever' I'd stay below 1.3 V, if you just want a couple more years from it, then it's probably fine.

Nonsense! Intel stated: max safe vcore for sandybridge chips of 1.38v, technical documents from Intel state a max vcore of 1.425v is absolutely safe

 

[Alexrose1uk]

My old 2700K is still alive and well with my father in law, running 4.8GHz 24/7 from memory. He is interested in doing an upgrade for higher IPC and to get onto a modern platform (as he does online bet taking (the opposite of betting), so a bit like stock market, some of the software favours high IPC). It's only now, now that system is ~10 years old, that he is considering upgrading. Absolutely legendary shelf life really.

 

[MartinPrince]

...It why most people used to say 1.5 V was safe too. Because the VID values go up to ~1.52 V. Loads of threads claiming >1.5 V was fair game if you look back. I wonder how many of those CPUs are in good order today....

Most people certainly never said 1.5V was safe! The general consensus from the wise/experienced (from 8pack et al) was always stay below 1.4v for longevity. My 2600K was indeed a golden sample but that has nothing to do with it not degrading. I also had a golden 3930K that did 5Ghz but that needed ~1.47v for stability and after a few years it struggled to do 4.6Ghz at any voltage, which was really of no surprise to me. Plus the heat was getting into the mid 80's and ideally with SB/IB you want to be in the very low 70's.

The other reason I'm pretty sure it is the constant manual voltage that really helps to degrade CPU's is because I've done and overclocked around 10 Sandridge CPU's for colleagues/family etc over the years and the only one that degraded was the one that I forgot to set back to offset voltage and left at manual for several years. It wasn't until he complained about crashes that I checked his bios and realised my mistake!

You also have the issue of high transient spikes when using offset with high levels of LLC. While it is fine to use high levels of LLC with constant voltage, when using Offset (or Dynamic Vcore) then the lower the LLC the better. This, I think was my mistake.

You didn't say whether you used constant manual voltage?