Post your 2xxxk 4GHz voltages

[Kenmitch]

Also, a better test for stability would be running enough loops of IBT so you can test for at least 12 hours, or half a day. 5 loops is pretty easy for any system. Folding@home is good for stability testing, too, even if it wasn't made with that purpose in mind.

Stability testing is for the most part is over done. I guess if you build a rig to run IBT, Prime, Folding@home 24/7 then you'd have to stress test accordingly. For most people it's extreme and subjects the rig to severe loads that will never be seen in real life.

Not that stress testing isn't important but there is such a thing as over stable at the same time for the intended uses for the rig.

==============================================================================================

4ghz at 1.152v's IBT(8-threads)(10 runs) on the high mem setting for now. The idle vcore is getting down pretty low tho. Had to use a - .115v offset to get the vcore down....Idle's at .872v's

Just using the high memory setting as it allows me to surf the web at the same time. Full load on 8 threads pretty much cripples the system when running max memory which makes it no fun to me. If I planned on running such a whimpy overclock then I'd stress test it more severely.

 

[LOL_Wut_Axel]

12 hours intel burn test? Nah im all set with burning up the vrms on my mother board.

I have a system of testing that 5 loops if itb is enough and i never get any blue screens.

If you can run 5 loops full avx itb and run 8 threads of 32m pi at the same time along with 3dmark 01 your cpu is stable.i also loop cinebench 3 times and if that wll passes the system is stable enough to do everything i need.

Long endurance itb runs strain the board and will crash at some point.iv had my 2600k since launch and its still stable so far.just did 4.6ghz for 1.365 volts and thats pretty spot on with how it clocked in the beggining,maybe a tad more vcore but thsts normal break in increase

At your frequency settings and voltage, sure. I've ran IBT for more than 24 hours on occasions with no ill effects at all.

5 loops of IBT is nowhere near enough to determine system stability. Here's a good test for you: run IBT and smp8 on folding@home, each for a straight day, and then tell me it's stable.

And contrary to what you believe, voltage affects degradation more than temperatures. A 2600K running at full load 80C with 1.35V at 4.5GHz will degrade at a much, much lower pace than a 2600K at 50C with 1.5V at 5GHz.

 

[Rvenger]

I don't even stress test much anymore when I overclock. I just know when my system is stable or not. I will just run 10 passes of IBT standard, run a pass of memtest 86, then use the computer like normal. If its unstable, its bound to come up in a relatively short period of time, especially when gaming.

 

[grkM3]

try running 1.55 volts cpu and 1.66 mem on the board and having the board set to extreme phase on cpu voltage,mem voltage and running the switching at 1100hz and see how hot your board gets.

Im not talking about an easy 4.6 over clock,IM talking 5.1-5.3 overclocks that put extreme strain on the boards to keep the voltage clean and stable at those volts and loads.

there have been vrms catching on fire running past 1.55 volts and IM not about to run 24 hour endurance test on my 380 dollar MB lol.

to put it into how important a higher end board is.I had a mid range gigabyte board and it would hit a wall at around 4.3-4.4 ghz with this cpu,then went up to a mid range p67 pro and it did 4.7-4.9 with this same cpu and right after that I sold that board and stepped up to the p67 MIVE that was the most expensive board at the time and this board is doing 5.2-5.3ghz with the same chip and cooling,also this is the only board that I got to hit 2200mhz ram speed with tight timings using 4 sticks of ram,I could not even get the other boards to post at all running 4 sticks at 2133+ ddrs speeds.

its the higher end vrms and caps on the board that get the system to perform at a higher clock and trust me,if this board was stable at 5ghz its nasa stable at 4.6.

I would challenge you to get this thing to crash lol

the p67 mive has been the best board I have ever owned,it impressed me so much I built another rig with it instead of going to the newer chipsets since I dont need the igpu and will never use the other features.

try my torture testing to see if it will pass.

32m pi is run to test ram stability(run hyper pi and set 8 instances)
latest intel burn test with avx support and at the same time run it while hyper is killing your ram.
next up load up 3dmark 01 and let it rip threw the gpu.

if your system can do that for 10 min its stable.Then I do 3 cinbench runs on its own 3 times and if that passes the system is rock stable for months.

 

[Idontcare]

That seems very strange. Are you sure those numbers are real?

No, oh, dear me, no, they aren't real at all. I totally made all those up. Every one of them. I didn't think that need be stated, otherwise I would have.

 

[onelivinlarge]

4.4 ghz at 1.278v

 

[Kenmitch]

Figure I gotta be close to the minimum vcore required for a 4ghz overclock on my 2700k with HT enabled. Was using IBT high mem settings for 10 runs which passed at my current settings. Decided to kick the testing over to prime95 for some long term stability testing.

I figure with this vcore and temps I won't burn alot of power anyways! Been running about an hour and twenty minutes so far guess I'll let it run for the rest of the day to see how long before a core errors out. Looks to be bsod proof at least

 

[LOL_Wut_Axel]

try running 1.55 volts cpu and 1.66 mem on the board and having the board set to extreme phase on cpu voltage,mem voltage and running the switching at 1100hz and see how hot your board gets.

Im not talking about an easy 4.6 over clock,IM talking 5.1-5.3 overclocks that put extreme strain on the boards to keep the voltage clean and stable at those volts and loads.

there have been vrms catching on fire running past 1.55 volts and IM not about to run 24 hour endurance test on my 380 dollar MB lol.

to put it into how important a higher end board is.I had a mid range gigabyte board and it would hit a wall at around 4.3-4.4 ghz with this cpu,then went up to a mid range p67 pro and it did 4.7-4.9 with this same cpu and right after that I sold that board and stepped up to the p67 MIVE that was the most expensive board at the time and this board is doing 5.2-5.3ghz with the same chip and cooling,also this is the only board that I got to hit 2200mhz ram speed with tight timings using 4 sticks of ram,I could not even get the other boards to post at all running 4 sticks at 2133+ ddrs speeds.

its the higher end vrms and caps on the board that get the system to perform at a higher clock and trust me,if this board was stable at 5ghz its nasa stable at 4.6.

I would challenge you to get this thing to crash lol

the p67 mive has been the best board I have ever owned,it impressed me so much I built another rig with it instead of going to the newer chipsets since I dont need the igpu and will never use the other features.

try my torture testing to see if it will pass.

32m pi is run to test ram stability(run hyper pi and set 8 instances)
latest intel burn test with avx support and at the same time run it while hyper is killing your ram.
next up load up 3dmark 01 and let it rip threw the gpu.

if your system can do that for 10 min its stable.Then I do 3 cinbench runs on its own 3 times and if that passes the system is rock stable for months.

No, it's not.

I'm telling you to do it because it's you that insists that what you're running at is stable, when I know for a fact it'll probably blue screen well before you get to 24 hours. Now you're backing off and saying that will burn your VRMs. Here's an interesting thing you may have not thought about: if your system can't run at full load for extended periods of time without burning up, you have problems you need to look into.

But hey, why stop at 10 min of testing? That's too much. One minute ought to be enough to determine complete stability.

 

[LOL_Wut_Axel]

No, oh, dear me, no, they aren't real at all. I totally made all those up. Every one of them. I didn't think that need be stated, otherwise I would have.

Wow, great counterargument. Now would you mind explaining why it's totally out of line with what everyone else is getting?

 

[grkM3]

here is mine with avx enabled ibt 1.192@4.0ghz 8 threads

and LOL my system is stable,what you dont get is that benching and 24/7 settings are totally different and I dont need to run my chip at 5.3ghz for 24 hours to be sure its stable to your testings.a long indurance run like that tests a lot more than just your cpu.

let me ask you,if it blue screens at 23 hours,how do you know that wasnt caused from your ram,or a voltage spike/drop on your psu?

we are talking about the cpu being stable not the whole system and for my 24/7 use it not bluescreening like you think.

you can run intel burn test for 24hours but running 8 instances of 32m hyper pi can fail in 5 min,give it a shot and see if your 24 hour overclock has stable ram settings.

 

[Kenmitch]

here is mine with avx enabled ibt 1.192@4.0ghz 8 threads

That's as low as you can go? Your chip sucks! Just kidding

I can't see mine being stable at lower than my above screenshot. I might try later tonight to see tho. During IBT w/AVX at the same settings as my above screenshot my vcore fluctuates from 1.144v - 1.152v's.

 

[bullbearish]

4.4Ghz at 1.336v

 

[grkM3]

That's as low as you can go? Your chip sucks! Just kidding

I can't see mine being stable at lower than my above screenshot. I might try later tonight to see tho. During IBT w/AVX at the same settings as my above screenshot my vcore fluctuates from 1.144v - 1.152v's.

If you want the vcore to not move at all you need to set cpu manually,then set the load line 100% and the over current to 100% and it should stay to what you set it no matter what the load is.

here is my setting and please dont use these without learning your board.

and do not set over current to 160 or your vcore will spike with load.and lol when I set my vrms to 1100khz and extreme phase I get a warning pop up that says to watch out for the vrms thermal conditions and warns me in every adjustment i set with extreme phase,there have been boards that got on fire but its cool,your asus board wont put out half the current and power the mivs do so you never really will see how hot the board gets.

and you can even diable the vrm temp protection if you want to fry the board going for a max 57x ln2 run,you can also disable max cpu current limit like where I set mine to 160,you can set it to disable and the board will give the cpu all the current it is demanding for xxx overclock.

that how the first few people were burning out sandys when they came out at launch,a few wrong settings on this board and the cpu will get 1.75 volts when it loads up

 

[Acanthus]

My chip wont go down to 4.0GHz. It wont take that dishonor.

 

[bryanW1995]

You guys are whining about 1.2v?

It takes my 1090T 1.45v to hit 4GHz and momma does it get hot D:

Oh the other hand, I think my friend's 2500K does 4.3 stable at 1.25v. Not sure if I should take him seriously...never seen the figures myself.

That's totally believable, it only takes me 1.34 to be stable at 4.54 fir 300 min linx and 24/7/365 seti with optimized apps and zero errors (for two months now). I also test stability by playing older but still somewhat demanding games like nwn2 or civ iv while keeping SETI running. That is probably the best long term stability test I know. I didn't even bother checking stability at 4.0.

 

[Rvenger]

When I get home, I am going to check my voltage at 4ghz. I will post it up here later.

 

[Idontcare]

Wow, great counterargument. Now would you mind explaining why it's totally out of line with what everyone else is getting?

Try clicking the link provided in the post of mine which you quoted.

As for explaining the differences, you are comparing prime95 power usage to Linx and wondering why there is a difference in power-consumption?

Not to mention the differences in video cards, ram, and ram usage during stress testing?

A bit of an apples-to-oranges issue there I'd expect.

 

[LOL_Wut_Axel]

Try clicking the link provided in the post of mine which you quoted.

As for explaining the differences, you are comparing prime95 power usage to Linx and wondering why there is a difference in power-consumption?

Not to mention the differences in video cards, ram, and ram usage during stress testing?

A bit of an apples-to-oranges issue there I'd expect.

So what graphics card where you using on these seemingly dubious tests? The RAM and their usage won't make a difference anywhere near as big, so it's irrelevant.

Also, Prime95 and IntelBurnTest are both designed to stress the CPU's components as much as they can. They test the CPU and the cache to their max. There's absolutely no way the differences in power consumption running both will end up being 100W or more.

To me it seems like your graphs are very exaggerated, unless you have something like 3x GTX 580 which even while being idle would add up to something significant.

 

[Charlie98]

I've got mine set at 4.1GHz at 1.260 vCore in LinX with full (8GB) memory. My temps are right at 70-72C with a 212+ push/pull. It started getting unstable beyond 4.2GHz, so I called it a day, and it did NOT like any BCLK adjustments.

 

[Rvenger]

4ghz at 1.15v. I was expecting much lower since I am at 1.18v right now. Strange....

 

[Joseph F]

Dude, 4.4GHz at 1.18v is beastly.
I'd probably pay $250 for a chip that was known to do that.

 

[Kenmitch]

4ghz at 1.15v. I was expecting much lower since I am at 1.18v right now. Strange....

WTF? Doesn't seem logical to me....Strange

Not sure how much longer I'm gonna run prime95 for. Seems like a waste of energy trying.

2700K ULV

Crazy low temps. Most likely about idle temps with the stock cooler for alot of people

 

[Joseph F]

WTF? Doesn't seem logical to me....Strange

Not sure how much longer I'm gonna run prime95 for. Seems like a waste of energy trying.

2700K ULV

Crazy low temps. Most likely about idle temps with the stock cooler for alot of people

Can you run HWiNFO64 and show the power consumption figures it gives?

 

[Kenmitch]

Can you run HWiNFO64 and show the power consumption figures it gives?

Not sure what part you wanted. I've never used the app is it any good? Seems like it shows alot of stuff....Although I'm sure my memory temps are not correct. Prime95 is still running currently during this shot.

 

[Joseph F]

Not sure what part you wanted. I've never used the app is it any good? Seems like it shows alot of stuff....Although I'm sure my memory temps are not correct. Prime95 is still running currently during this shot.

Your CPU package power is what I was asking for. 83.478W is very nice.
My 2500k can use up to 90W at 4GHz/1.2V.