so far flattened out hopefully . . . but then again I haven't been priming just hitting linx lately ,Quote:
Originally Posted by Falkentyne
Printable View
so far flattened out hopefully . . . but then again I haven't been priming just hitting linx lately ,Quote:
Originally Posted by Falkentyne
Will see how long mine lasts but guys you are not degrading your chip in 2 days.Ever since I started overclocking new chips will always break in and as soon as they get more volts they always want more volts.Im running 4 sticks of 2200 ram at 1.66 volts and pushing it and have no problems.
some of you are basically giving the chip stock voltage to hit 4.5 ghz lol
Says the guy hitting 98C on load... :ROTF: ;)Quote:
Originally Posted by Skratch
stock intel coller and closed case LOL and my chip is still doing exactly what it has been for about 4 months now.Quote:
Originally Posted by dr_sharp
I just ran linx with 4gb of ram for 5 passes and I have 4 sticks of 2200mhz ram and 1.66 volts on the memory controller.
it hit 98c for over 10min and guess what? my chip is still running 5ghz haha
I get no blue screens or reboots at all.This thing is rock stable.
It hits 98c on linx and prime and It will never come close to those temps in normal daily use.
99c it will throttle so by intels specs the cpu is still fine,most mac book pros hit 90c with normal use.
Dude - if you're comfortable with those temps, that's all that matters. Don't let me or anyone else deter you from running high clocks.Quote:
Originally Posted by Skratch
However, IMHO, anything above Intel's recommended 74C max temp is suicide whether it be in the short term or long term. If 4 months is long term to you - wonderful :clap:
Those temps Im hiiting is only when doing 100% load like linx with 6GB of ram ect.Quote:
Originally Posted by dr_sharp
My setup never sees over 65c in normal daily use.
I can run intel burn test with 1 gb of ram selected and it wont break 70c
I only brought up the 98c to show that I have no degration and Iv hit those temps while trying to see my max stable overclock.
normal use will never see those temps,heck im at 30c right now
-40c 1.7vCore 1.155vCCIO 5827mhz still max and stable as day one, couple months now :D
Is that a 2500k or a 2600k?Quote:
Originally Posted by Splave
Yeah you're on subzero, so you're definitely not going to get much degradation. So far everyone I've seen who's on extreme cooling and high voltages have lost maybe 50 mhz at most. And 1.5 VCCIO? Wow....why so high? Hypers?
This sounds like a great idea:- forums posts are usually biased toward things that go wrong, and I was starting to think that all 2500/2600 chips were going to degrade.
Would it be helpful to add lot number / date / place of manufacture to the data collected to see if there are any consistencies?
I haven't bought a board yet, had just saved enough when the SATA bug was announced, then the Z68 came along, now these more frequent stories of
ion seep / whatever is causing these degrading clock / voltage problems. Could be that the Ks are a relatively small jump in price because the few bins a non-K will give are all that is sustainable?
This makes things difficult: I really need a new computer :)
Model: i7 2600K
Batch: L040B698
Mobo: Asus P8P67 Pro
4.6 GHz
BCLK: 100.1
multi: 46
vcore (BIOS): 1.275
vcore (load): 1.352 - 1.392
LLC: none (standard)
VRM Current: 100%
VRM Frequency: 380 KHz
VRM Phase Control: Manual: Fast
VRM Duty Control: T.Probe
Been running WCG and other stuff since Janurary. No noticable degradation yet.
Average temps: 65 - 75
Temps spikes during the recent heatwaves: 80 - 84C (sustained for about 5 hours until I woke up and saw it :rofl:)
It'll be running at 4.2 GHz for the summer. My AC can't keep up... :(
I see... do you use prime95 for stability testing? Curious to see if you're 8hr stable and what temps you see, if so.Quote:
Originally Posted by Skratch
Lot number and date of manufacture are irrelevent, as the CPU's aren't even fabbed where they are packaged.
More to the point would be something like default serial VID @ 3.3 or 3.4 ghz, as just about the ONLY identicator of how well a CPU will overclock, is the default VID; usually the lower the VID, the less vcore it will require (this is sort of self explanatory). But oddly enough, the hotter the chip may tend to be.
There ARE some good high VID clockers, but a lot more dogs are high VID.
BTW, if anyone cares, I managed to compensate for SOME of the degradation on my lower VID 2600k @ 5 ghz (and at prime blend @4.5 ghz at 1.236-1.248v load) by raising the PLL voltage to 1.86v. This helped with some of the idle/light load 0x124's I was getting at 5 ghz, with "creeping" degradation (meaning starting to get idle bsod's after several days of playing BC2 at 5 ghz), as well as slowing how fast they occured in prime at 4.5 ghz at 1.236-1.248 load. (1.260v load is prime blend 12h stable). By raising the pll voltage, I was able to stay on the desktop at 5 ghz (1.410v BIOS LLC2, 0.01v lower than before) for 1 hour idle, and in the main menu in bad company 2 for another hour before I got bored and closed it. (previously, it would BSOD 0x124 in the menu but not on a server).
I Guess the point was that I was getting 0x124's on this chip, while my heavily degraded one *always* 101's in prime blend without enough vcore (Now you guys see why I experiment).
1.260v is like my VID for 3.4Ghz on my 2600K, pretty darn nice chip!
But Im a bit confused on my VID sometimes after a CMOS reset it shows 1.240 or 1.255
Haha, I know why.
When you reset the CMOS, the turbo multiplier is enabled for 4 core usage. So the multiplier is +1 higher than the default. Notice that whenever you reset the CMOS, the 2600k posts at 3500 mhz?
When you go into your BIOS and disable the turbo options, you're back at normal (either 33x100 for 2500k or x34 for 2600k), then the VID shows properly.
BTW I restabilized 5 ghz again, partially. Dropped my vcore by 0.01v and raised the PLL to 1.86v, and I was able to stay in the bad company 2 main menu for an hour without a 0x124 error popping up, and in the desktop for another hour. I'll try one more test with leaving bad company 2 idle, then I'll go back to my worse 2600k and degrade it some more some some volts. No idea if it will pass again or not, though.
See if it can do 24hrs idle :D
BTW, would 1.392 be a safe ildle voltage? Loads at 1.380v
Still going strong:up:Quote:
Originally Posted by Splave
Same here:D
Imo, VID, batch etc. doesn't make any difference in term of Mhz. wall.Quote:
Originally Posted by Falkentyne
SB K chip that can go 57X+ is out there in retail channel, but it won't carry any specific mark...unless Intel decided to bin it:D
Not interested in degrading my cpu leaving at 5 ghz for no reason. I'm done doing stress testing...I want to game.
BTW it passed 1.5 hours being idle in bad company 2 menu with pll at 1.86v (Default PLL used to BSOD there within 15-30 minutes).
You sure about that claim ? we have easily lost over 100Mhz on some AMD CPUs subzero... It's not always just about cold...Quote:
Originally Posted by Falkentyne
Also when pushing SB, there's no Mhz left to squeeze out as it will refuse to post windows, detect the HDs,...
Splave uses 1.15Vcio :p
Like said before some CPUs have a burn in period. If you only need a small voltage increment to keep the same clocks stable, so be it... is it really degredation? dunno for sure...
Failing a stability test can have multiple reasons and is not necesarrily only the CPU too blame... mostly it's the moterboards circuitry or new biosses that influence the vcore setting...
Losing 100 mhz is not the same as losing 300 mhz if you were on air at the same settings, Lee :) 0.03v rise is a lot different than having a 0.08v rise. And maybe i made my claim come out badly; all I was saying was, you're going to have less degradation on subzero than you would on air, at high voltages.
Wonder if its anything to do with vccio? All you guys reporting degradation running vccio 1.2v+?
Mines never needed to be over 1.16v when you get the memory subtimings right, even up to 2250mhz. Maybe ive just got a good IMC, who knows. Still itd be interesting to plot vccio against degraded chips.
Nope, 1.05v vccio only. And you will never get 0x101 from vtt problems. usually degradation affects the cpu cores directly and causes 0x101. 0x124's can be the result of IMC or the clock stability/PLL or other reasons.
Cpu 2600k @5418 (52x104,2) for benching only.
Vcore (Bios) 1,59
Vtt 1,12
vDIMM 1,693
Cpu on watercooling, 54 celcius max
Mobo ASUS P8P67 PRO
After 50 days cpu needs 1,6 for 5318.
Pretty big degradation
2600k, bought on release day, 24-7 5Ghz 1.418v. Has been frozen about a dozen time @ 1.6v. When it was new I got 5768Mhz, now it does 5758Mhz so I lost 10Mhz in total.
its 1.15 :rofl:Quote:
Originally Posted by Falkentyne