Printable View
The thing started degrading in the FIRST two days after first using the chip, every time I kept running it at 1.44 or higher vcore. It's DEFINITELY not the motherboard, that's for sure. Unless you think TWO different UD5 boards degraded. It's the CPU.Quote:
Originally Posted by Skratch
Well when you get your new chip I can try your old one to see if you want.Make sure you dont have some setting that boost the voltage and currant with load in the bios set wrong.With my board I can adjust one step up on the currant limit and it shoots the voltage from 1.4-1.50 once it see load.
Im running an asus board so Im not sure what your bios has for power settings
Since Thorougbred cpu up till SB I never had a chip that degraded, it just died
I just hit tmax of 98c for 5ghz in linx. I did 3 passes and stopped because it was right at 97-98c and Im on air cooling.
Maybe the 2600k degrade faster but I have been abusing the hell out of my 2500k since the day I bought it amd its still clocking the same.
Hyperthreading could make a big difference. Can't really say. I haven't really tried with HT off at lower speeds/voltages, but I should have kept tabs on that.
As I said though, I'm degrading my first cpu on purpose after the second one went from 1.236 to 1.260v at 4.5 ghz and 1.404v to 1.440/1.452v at 5 ghz. I'm trying to test that stuff so my second 2600k (which still runs at lower voltage) lasts longer. (it originally went from 1.404v to 1.428v just from testing prime at 5 ghz, then I stopped priming and only playing bad company 2, and it degraded more (but not as fast).
BTW I should add:
When I was running 5.3 ghz 1.58 vcore (before that degraded to the point where it would BSOD while loading black ops), I was able to game for a couple of days. I couldn't even load XP with HT on at that vcore...it would just reboot about 20 seconds after the loading bar appeared... but with HT off, I was able to game for a few days in XP. I was able to make it into windows 7 with HT on, to the desktop, but it would BSOD with system service exception...
I was able to make it at 5.4 ghz into XP with 2 cores enabled, though :P
In just a matter of how many days did yours degrade falk?
Which chip, demo?
My second 2600k? The second one took months to degrade because I almost -never- ran it at 5 ghz, except a couple of 45 minute prime sessions (since early March) until recently, when I started testing 5 ghz out a lot more. It held steady and firm at 4.5 ghz 1.236v prime blend stable without any degrading, until I started trying to run prime at 5 ghz to get into the 1 hour stable club, and finally managed it (after a bunch of BSOD's and trying to find the lowest voltage possible). (the vcore was 1.404v load). After this, I went back to 4.5 ghz, and saw that 1.236v BSOD'd after 2 hours, which it NEVER did before. I immediately knew why. I went back to 5 ghz and did more tests and saw 1.404v load (1.445v bios LLC1) and 1.450v LLC1) start BSOD'ing. 1.416v load passed (1.455v LLC1), but eventually that started failing too. Then i gamed a bit and tried 1.465v LLC1 and that BSOD'd after about 1 hour and 30 minutes (1.416v load). I tried a quick 30 minute 1.48v LLC1 (1.428v load) which passed, then I stopped priming completely and just started playing bad company 2. I switched to LLC2, because BC2 would only load 4 cores, and the vcore was thus higher. Eventually I saw that (since the vcore would rise under higher load with LLC2), that I would BSOD in the main menu sometimes, and I was using 1.400v LLC2 now, for 1.428v load. Went to 1.405v, was fine for awhile, but eventually, had idle BSOD's on the desktop and in main menu (never inside a map though) after awhile (more degradation at the lower vcore). Eventually went to 1.410v, still BSOD in main menu. 1.415v was fine in game, and on main menu, but had idle BSOD in windows after an initial 5 hour idle session with no BSOD, but had it the next day. Did 1.420v LLC2, which was 1.440 to 1.452v in game, no idle BSOD at all in desktop in 6 hours, but I only did 1 game session at this setting. Went back to 4.5 ghz, saw that I now needed 1.260v load for 12 hour prime. I tried 1.248v load but that BSOD'd after 8 hours. 1.260v passed three times with three 12 hour runs.
Then I swapped chips to my first "worse" one and am now abusing the hell out of it.
Ever heard of a burn in period ? Some CPU's get a little bit more demanding for Vcore after a while... Also what do you expect running +0.3 more vcore than stock, this on air or water with temps due to silly stress testing reaching over 90°C... It's all part of the game...
I didn't run my better (second) 2600k at 0.3v vcore over stock. I only primed for a short time at about 0.16v over stock. I'm *Deliberately* running my first one like that on purpose for the sole purpose of degrading it. (and the second one just some normal gaming at about 0.2v over stock). Where was I complaining? I don't care about the first chip, only the second.
And I didn't start this thread. He asked for feedback, thus I'm giving feedback. I did buy the first cpu and it had a high VID, so I feel like seeing how many volts it can take. Is that a bad thing?
have lost 30-40Mhz on 2 x 2600K (both could do above 5.6Ghz) DryIce at 1.65V bios set with Level 2 Vdroop on these 2 chips from the time started testing them.
On air, they are being run at stock, :D
Seems that extreme cooling allows the chips to be run safely at higher volts, greatly minimizing this effect.
CPU: 2500k @ 4.5ghz
Vcore(bios): around 1.33v
Vcore(DMM):
VTT: 1.2v
vDIMM: 1.65v
Mobo: Asrock P67 Extreme4 B2
Running fine for 3-4 months.
I wont say that, atleast in my experience as mentioned above both my chips lost 30-40 Mhz in say 6-7 or maybe max 10hours of benching above 1.65V with Dry Ice. Ofcourse if I would have run Prime 95 for that duration maybe the degradation would have been worst.Quote:
Originally Posted by Falkentyne
Quote:
Originally Posted by Leeghoofd
+1 I fully agree.
Every CPU I have had since the inital launch of 45nm (e8400) has needed a bit more voltage after burn-in. I just factor this phenomenon in now.
CPU: 2600k @ 4.8ghz 24/7
Vcore(bios): around 1.48v in windows 1.46v
Vcore(DMM):1.48
VTT: 1.2v
vDIMM: 1.68v
Mobo: Gigabyte P67A UD7
Ive slammed my chip with lots of 12 hour prime , she used to do 4.8 at 1.430 vcore , now she needs more juice , also think changing my ram from 1600mhz to 2133mhz stressed my cpu a bit more
Which LLC option do you use The Real 7? Is that 1.48v dmm reading at load? I do not trust software reading on my UD7. I always assume that bios set ~ real vcore. I still need to check with a dmm to verify Sin's readings with his board.
How much more juice, real7?
Would running the mem at 2133Mhz (vDIMM 1.520v) some how accelerate the degradation?
Also those who posted so far are using VTT at 1.15 ~ 1.200v, how do I exactly determine when I need to up the VTT and when to up or lower PLL voltages?
I use LL1 , I use to not trust the software reading but I tested hard with idle and loads , and i can tell you without a doubt the reading are on point in software , look couple pages down in the ud7 forum I have pics with the software and the dmm at same time :up:Quote:
Originally Posted by chuchnit
well my chip used to do 1.430vcore in bios and used to pass at 5.0 easy , now shes needs 1.480v - 14.90v to remain stable at 4.8 , for 5.0 i need something like 1.550v now . . . is this settling or degration ,I would say a bit of bothQuote:
Originally Posted by Falkentyne
I've been crunching for almost 1 month now at 1.5v (for 4.88) but the temps are typically 58 deg C. What is the consensus on "safe from degredation" temps? or put another way can we look at this from a temp point of view, regardless of voltage?
:|
wow
i think that another important information about it it's the stepping
however i hope that these cpus are only unlucky
well I don't know about accelerate the degradation , but for me the rule of thumb has always folowe, the more aggressive the ram , the more vtt and vcore you need to keep stable , I can run 1600mhz tight timing at stock vtt and a bit lower vcore , so in theory with slower ram I could probably easily get a better overclock with stresses the proc lessQuote:
Originally Posted by Demo
how do I exactly determine when I need to up the VTT and when to up or lower PLL voltages? Great question dude following the bsod guide doesn't seem to work always cause 101 and 124 could either be vtt or vcore :confused:
:shocked:Quote:
Originally Posted by the_real_7
Is the chip still needing more vcore or has it flattened out?
If you run at 4.8-5.0, are your voltages still needing to be increased?
I had 2 2600K with sign of degradation...
2 things was common in both CPUs:
- Both was at 1.5v Vcore, VTT around 1.3v, PLL Auto, vDIMM 1.5v
- And Bad clockers.
One with P8P67 WS Revolution (B2) and the other MIVE (B3).