Sandy Bridge maximum vcore for 24/7 use

**CedricFP** · 01-28-2011, 12:02 AM

Originally Posted by ugotd8

What about the engineers that wrote the manual. I don't consider them dolts, as I'm sure no one else does either. They have engineered a marvel of performance/thermal efficiency that will go down as one of the best chips in Intel's history.

Intel has designed this chip to be overclocked. There are thermal limits built in, and none of us are coming close to the 90c throttle stop. They say 1.52 is safe, most are no where near that 24/7.

I for one, am overclocking on temps with this amazing platform.

And if it degrades or burns up, I'll send it back to Intel and tell them my grandmother overclocked it too far. ;-)

I would ague that the engineers who advised those who wrote the manual did not intend the 1.52v limit to be a safe 24/7 limit. It's more likely a threshold above which CPU damage becomes significant - as in, higher than a specific margin / percentage with error taken into account.

I would say that the reps told to follow the manual, in general, are likely not knowledgeable about the implications of long term usage at particular voltages.

Considering that the person asking the rep was asking a specific question - ie, 24/7 usage, safe, etc. - the rep hardly answered his question.

Hence, any dolt can read from a manual and make incorrect inferences.

**vern** · 01-28-2011, 12:14 AM

I think under 1.45V is ok

**JumpingJack** · 01-28-2011, 12:24 AM

Briefly reading over the SB data sheets: http://download.intel.com/design/pro...hts/324641.pdf there is no clear answer to this question. The data sheet is also missing some very, highly useful specifications that have appeared in prior Intel data sheets (namely, absolute Vcc Max). Having read the SB specification, the VccMax reported in the VID table is the pre-determined max VID for any given processor set at the factory:

Each processor is programmed with a maximum valid voltage identification value (VID) that is set at manufacturing and cannot be altered. Individual maximum VID values are calibrated during manufacturing such that two processors at the same frequency may have different settings within the VID range. Note that this differs from the VID employed by the processor during a power management event (Adaptive Thermal Monitor, Enhanced Intel SpeedStep Technology, or Low Power States).

I interpret this to mean that VccMax as specified by the VID (hard coded bits) is the max voltage attained when Turbo is engaged and power saving states change leave idle. I do not see this as the absolute max, which intel will typically specify. For example, in the i7-980X specification: http://download.intel.com/design/pro...hts/323252.pdf see section 2.10...

Table 2-6 specifies absolute maximum and minimum ratings, which lie outside the functional limits of the processor. Only within specified operation limits can functionality and long-term reliability be expected.
....
At conditions exceeding absolute maximum and minimum ratings, neither functionality nor l ong-term reliability can be expected. Moreover, if a device is subjected to these conditions for any length of time then, when returned to conditions within the functional operating condition limits, it will either not function or its reliability will be severely degraded.

In other words, exceeding absolute max conditions will begin a degradation process that is not reversed if returning the CPU to the nominal factory programmed settings. The SB specification does not have such a section in it. For an i7-980X the absolute max is 1.4 V, I would take this to be probably the same region for SB (same process tech).

**wekktor** · 01-28-2011, 12:54 AM

my conversation with an intel guy

You are now chatting with 'Brian'
Brian: Hello. Thank you for using the Intel Customer Support chat service. We are glad to be of service. How can I help you today?
we_kk_tor@hotmail.com: Hi
we_kk_tor@hotmail.com: I want to know what is the max safe voltage for sandy bridge?? no one really seems to know that, thx
Brian: Do you have the model number of one of the Intel products?
we_kk_tor@hotmail.com: you mean i5 2500k?
Brian: I am going to transfer you over to the technical department and they will continue with the process.
Brian: Please hold.
we_kk_tor@hotmail.com: ok thx
Please wait while I transfer the chat to the best suited site operator.
You are now chatting with 'Dago'
Dago: Hello. Thank you for using the Intel Customer Support chat service. We are glad to be of service. How can I help you today?
we_kk_tor@hotmail.com: Hi
we_kk_tor@hotmail.com: I want to know what is the max safe voltage for sandy bridge?? no one really seems to know that, thx
Dago: let me check, the processor specifications, please wait
Dago: this is the processor page, now let me check your inquiry:
Dago: http://ark.intel.com/Product.aspx?id...he%2c+3.30+GHz)
we_kk_tor@hotmail.com: I cant find any volt values?
we_kk_tor@hotmail.com: the intel specs says 0.025v - 1.52v range in volt, but 1.52v just cant be safe for 24/7
we_kk_tor@hotmail.com: some says 1.38v too but I want a official max safe volt from intel, dont you know the max safe volt ?
Dago: I am checking the datasheet, please wait sir
we_kk_tor@hotmail.com: ok thx
Dago: let me confirm, are you looking the max safe voltage for the supported ram or for the processor cores?
we_kk_tor@hotmail.com: not ram, I want to know the vcore for the processor
we_kk_tor@hotmail.com: max safe
Dago: ok fine, I am checking
Dago: thanks for your patience
we_kk_tor@hotmail.com: np
Dago: please check page 78 of the following document:
Dago: http://download.intel.com/design/pro...hts/324641.pdf
we_kk_tor@hotmail.com: ok
we_kk_tor@hotmail.com: yes it is this I mean, MAX = 1.52v, but for 32nm, that just can be a safe volt for 24/7 ?? Shouldnt it be much lower ?
we_kk_tor@hotmail.com: I mean wasnt the previous 45nm max = 1.45v or something like that?
Dago: those are the specifications for this processor, the VID is (0.25) min and (1.52) max, and it is safe for 24*7
we_kk_tor@hotmail.com: ok Thx, so 1.52v for 24/7 is safe
we_kk_tor@hotmail.com: yes or no
Dago: yes
we_kk_tor@hotmail.com: Thx!
Dago: Is there something else I would be able to assist you with?
we_kk_tor@hotmail.com: no sir, this was what I want to know! thx
Dago: Thank you for using chat. We value your suggestions and would appreciate it if you would take a moment to complete our survey. The survey will appear when you click "Close'' to close the chat window. Have a nice day.

**Leeghoofd** · 01-28-2011, 01:02 AM

I'm keeping my 2600K at it's sweetspot. Which is 4.5 at 1.28Vcore, for 4.6 I need 1.33ish and temps go up pretty steep (all on high end watercooling).

For benching preps I use 1.5Vcore for 5.2. It's 3D stable then...

I honestly feel daily usage won't rapidly degrade the chip (except folding 24/7), but all this uber stability testing does more harm then good...

**C-N** · 01-28-2011, 01:13 AM

Subscribed.... Interesting read

**Sam_oslo** · 01-28-2011, 01:19 AM

Originally Posted by Leeghoofd

I'm keeping my 2600K at it's sweetspot. Which is 4.5 at 1.28Vcore, for 4.6 I need 1.33ish and temps go up pretty steep (all on high end watercooling).

For benching preps I use 1.5Vcore for 5.2. It's 3D stable then...

I honestly feel daily usage won't rapidly degrade the chip (except folding 24/7), but all this uber stability testing does more harm then good...

I just got my SB, and still testing, but I've observed a jump/bump in vCore around 4.5GHz too. I had a few short 10-min Prime-runs, and these may need minor adjustments, but I got these so far: EDIT(putting in exact vCore)
4.5GHz = 1.272v
4.6GHz = 1.296v
4.7GHz = 1.336v

4.5 to 4.6 needs 0.024v increase, while 4.6 to 4.7 needs 0.040v, that's a bump/jump. Next jump will probably get steeper, maybe 0.08v, but I didn't have time to test 4.8Ghz yet.

Generally, vCore requirement (and the heat-signature too actually) may vary depending on you CPU-Batch#, but 24/7 OC is not about pushing it to the limits and edges. Even if you can get away with "safe"-voltages, you would usually need excessive vCore (and other voltages too) to get there.

24/7 OC is all about the sweet-spot, and it's a good idea to settle down when the vCore makes a sudden jump/bump. But it's always fun trying to push for benching testing tho.

**Falkentyne** · 01-28-2011, 02:25 AM

Originally Posted by Leeghoofd

I'm keeping my 2600K at it's sweetspot. Which is 4.5 at 1.28Vcore, for 4.6 I need 1.33ish and temps go up pretty steep (all on high end watercooling).

For benching preps I use 1.5Vcore for 5.2. It's 3D stable then...

I honestly feel daily usage won't rapidly degrade the chip (except folding 24/7), but all this uber stability testing does more harm then good...

Lee,
You were right dead on back during the core 2 days, when you talked about degradation. The Yorkies would experience a little bit of degradation even at lower voltages (like 1.36), sometimes as much as 0.04, and then would stabilize. Which is exactly what you said.
But repeated abuse and linpack and small FFT stress testing only helped make them degrade more, and faster...

I feel that all this 12 hour linpack/prime testing at high voltages does nothing but help degrade the chips...they're going through abuse they really werent meant to withstand ...

**C-N** · 01-28-2011, 02:51 AM

^^^ agree with that

**WintersSC** · 01-28-2011, 04:17 AM

Id say 1.375, or thats what ive heard so far.

**overcrash** · 01-28-2011, 04:21 AM

My personal opinion is 1.38v for Top Air cooling and 1.42v for top water cooling. Anything over 1.45v will degrade the chip over time. Running 5 GHz @ 1.5v because you can, doesn't mean you should.

**Mathemagician** · 01-28-2011, 04:48 AM

Thank you all for sharing. This is actually a practical matter for me rather than a theoretical one. I am not a gamer. I overclock work PCs because much of the work that I do is very CPU-intensive. Every MHz that I can squeeze out of a box translates to time saved, and over the service life of these PCs this time saved can be worth a good deal more than $330 for a new chip. (Roughly speaking, a 10% gain in CPU speed turns a 10 hr procedure into a 9 hr procedure. Two such procedures a week adds up to 100 hrs/year/box.) After that, they become email/web/goofing off family PCs and can be clocked conservatively. The effective service life of these PCs tends to be 1-2 years. So, what I'm really trying to do is get a good sense of the relationship between vcore+temp and rate of degradation, and find the sweet spot along that curve for my application. This process has been complicated because it's just been far too easy to find opinions (some of which are highly valuable when based on the extensive experience many of you have) and far too difficult to find hard facts in this area. Hopefully this will help frame the discussion a bit more effectively. Thanks again for your input.

**zalbard** · 01-28-2011, 05:51 AM

It was 1.4V for both other 32nm chips (Westmere: Gulftown and Clarkdale). I'd also stick to that. There were cases of people killing SB at 1.6V in a couple of days...
And temperature wise, I'd recommend staying below 70C at all times.
I obviously don't know for sure, but seems pretty accurate to me.

**Hondacity** · 01-28-2011, 06:58 AM

wow "some" of you guys are just blurting out numbers...

sheeesh

an idiot/kid/idiot(again) would just stick to 1.45v...then prime95 for 1-4 weeks..and confirm your fhucking recommendation and kill their chip.....or it will survive...or it will degrade.

**bro20000** · 01-28-2011, 07:03 AM

1.36v 5Ghz for me maybe higher when all my new hardware is here!

**aytacaksel** · 01-28-2011, 07:56 AM

nm and voltage? i just dont understand..

Pentium® 4 Processor Extreme Edition supporting HT Technology 3.73 GHz
90nm
VID Voltage Range: 1.200V-1.400V

Intel® Pentium® Processor Extreme Edition 955
65nm
VID Voltage Range: 1.200V-1.3375V

Intel® Core™2 Quad Processor Q9650
45nm
VID Voltage Range: 0.8500V-1.3625V

Intel® Core™2 Duo Processor E4300
65nm
VID Voltage Range: 0.8500V-1.5V

**Eldonko** · 01-28-2011, 07:59 AM

This is what Im doing for voltage:
1.4v for 24/7 and stick to 100 bclk
1.52v for 3d and CPU intensive benches, bclk oc ok
1.62v for 2d, PI and pifast only, bclk oc ok

Also, there is no need for excessive stress testing, that is just asking for trouble with these chips. Run LinX or similar for an hour, chances are you will be fine with that.

I have a 5.7+ Ghz chip and no degradation at all thus far with the above.

**wekktor** · 01-28-2011, 08:06 AM

intel says 1.52v is safe for 24/7, I dont understand why people claims other voltages then?

**Eldonko** · 01-28-2011, 08:08 AM

Where does Intel say that?

**wekktor** · 01-28-2011, 08:10 AM

here, and I have already posted that:

my conversation with an intel guy

You are now chatting with 'Brian'
Brian: Hello. Thank you for using the Intel Customer Support chat service. We are glad to be of service. How can I help you today?
we_kk_tor@hotmail.com: Hi
we_kk_tor@hotmail.com: I want to know what is the max safe voltage for sandy bridge?? no one really seems to know that, thx
Brian: Do you have the model number of one of the Intel products?
we_kk_tor@hotmail.com: you mean i5 2500k?
Brian: I am going to transfer you over to the technical department and they will continue with the process.
Brian: Please hold.
we_kk_tor@hotmail.com: ok thx
Please wait while I transfer the chat to the best suited site operator.
You are now chatting with 'Dago'
Dago: Hello. Thank you for using the Intel Customer Support chat service. We are glad to be of service. How can I help you today?
we_kk_tor@hotmail.com: Hi
we_kk_tor@hotmail.com: I want to know what is the max safe voltage for sandy bridge?? no one really seems to know that, thx
Dago: let me check, the processor specifications, please wait
Dago: this is the processor page, now let me check your inquiry:
Dago: http://ark.intel.com/Product.aspx?id...he%2c+3.30+GHz)
we_kk_tor@hotmail.com: I cant find any volt values?
we_kk_tor@hotmail.com: the intel specs says 0.025v - 1.52v range in volt, but 1.52v just cant be safe for 24/7
we_kk_tor@hotmail.com: some says 1.38v too but I want a official max safe volt from intel, dont you know the max safe volt ?
Dago: I am checking the datasheet, please wait sir
we_kk_tor@hotmail.com: ok thx
Dago: let me confirm, are you looking the max safe voltage for the supported ram or for the processor cores?
we_kk_tor@hotmail.com: not ram, I want to know the vcore for the processor
we_kk_tor@hotmail.com: max safe
Dago: ok fine, I am checking
Dago: thanks for your patience
we_kk_tor@hotmail.com: np
Dago: please check page 78 of the following document:
Dago: http://download.intel.com/design/pro...hts/324641.pdf
we_kk_tor@hotmail.com: ok
we_kk_tor@hotmail.com: yes it is this I mean, MAX = 1.52v, but for 32nm, that just can be a safe volt for 24/7 ?? Shouldnt it be much lower ?
we_kk_tor@hotmail.com: I mean wasnt the previous 45nm max = 1.45v or something like that?
Dago: those are the specifications for this processor, the VID is (0.25) min and (1.52) max, and it is safe for 24*7
we_kk_tor@hotmail.com: ok Thx, so 1.52v for 24/7 is safe
we_kk_tor@hotmail.com: yes or no
Dago: yes
we_kk_tor@hotmail.com: Thx!
Dago: Is there something else I would be able to assist you with?
we_kk_tor@hotmail.com: no sir, this was what I want to know! thx
Dago: Thank you for using chat. We value your suggestions and would appreciate it if you would take a moment to complete our survey. The survey will appear when you click "Close'' to close the chat window. Have a nice day.

**aytacaksel** · 01-28-2011, 08:11 AM

all cpu's diffirent than others. a cpu can select any vid from a vid range for thermal properties. what about if a cpu select 1.52v for run? then you can run 1.52v 7/24 at stock clocks. overclock is realy diffirent bussnies but i dont think 1.52 volt going damage any cpu component.

and intel live support no idea what we are talking about..

Allan: Hello. Thank you for using the Intel Customer Support chat service. We are glad to be of service. How can I help you today?

aytacaksel@gmail.com: hi

aytacaksel@gmail.com: i have a question

Allan: yes

aytacaksel@gmail.com: i know maybe alot ppl asking for same thing but i wanna ask again

aytacaksel@gmail.com: sandy bridge max vcore :P

aytacaksel@gmail.com: is it realy safe running 1.52 volt 7/24 if cooling isnt problem

Allan: what is your processor model?

aytacaksel@gmail.com: i5 2500k

Allan: Please wait for 2 minutes and I will check your concern. Thank you.

aytacaksel@gmail.com: thank you

Allan: I remind you that Vcore information may not be available as this can be used for overclocking...

Allan: please hold

aytacaksel@gmail.com: yea i realy dont care my cpu can burn but why i buyed 2500k? :P

aytacaksel@gmail.com: just wanna learn my limits

Allan: actually, the information is not available...

Allan: Intel does not provide that information

aytacaksel@gmail.com: oh

aytacaksel@gmail.com: okay thank you then

Allan: you're welcome.

Allan: Thank you for using chat. We value your suggestions and would appreciate it if you would take a moment to complete our survey. The survey will appear when you click "Close'' to close the chat window. Have a nice day.

Chat InformationChat session has been terminated by the site operator.

**Eldonko** · 01-28-2011, 08:24 AM

Uh yeah I dont think Im going to risk my golden chip on a conversation with some guy that doesnt sound like he knows what he is talking about..

**One_Hertz** · 01-28-2011, 09:14 AM

The question is... will you cry if the chip dies? If yes, then run 1.35v without internal PLL increase. If not, then run 1.45 with PLL increase... 1.5 is a little silly for 24/7 as there are multiple cases of this killing/degrading chips in just days. 1.45 is not known yet and might be safe. Not much is known about the increase internal PLL checkbox we are all enabling...

**deaffob** · 01-28-2011, 09:16 AM

Also guys, if you want even safer and greener(lol) way, use offset setting. Even if your load is 1.4, it will automatically scale down to below 1.0 at rest.

**zoob** · 01-28-2011, 10:51 AM

Originally Posted by Eldonko

Also, there is no need for excessive stress testing, that is just asking for trouble with these chips. Run LinX or similar for an hour, chances are you will be fine with that.

Would Folding@Home 24/7 be considered a form of "stress testing"? I find it pointless to overclock a CPU to 4.8+ only to let it sit idle 99% of the time.

Thread: Sandy Bridge maximum vcore for 24/7 use

Thread Tools

Search Thread

Rate This Thread

Display

Bookmarks

Bookmarks

Posting Permissions