i5 2500K decaying :o - Page 3 - Overclock.net

Forum Jump: 
Reply
 
Thread Tools
post #21 of 56 Old 05-19-2011, 03:58 AM
4.0ghz
 
Falkentyne's Avatar
 
Join Date: Dec 2007
Location: Riverside
Posts: 4,042
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 357
It IS 100% definitely degrading. You really need to be on good water at >1.4v if you don't want any degradation (besides what many, many chips will do from their first initial power on's....I saw a 2500k that loaded windows when brand new (First boot) at 1.29v @ 5 ghz, but wouldn't do it on the second boot...needed 1.31v, then at the end of the day, even THAT didn't work, and it needed 1.33v. It eventually settled at 1.35v and remained there long term). But I think people are getting totally overblown over what degradation is, as the chips can "degrade" at lower voltages slightly, too, as I just explained.

These chips can and DO lose their lowest voltage floors at certain overclocks, after they are run for awhile. How much this happens depends exactly from chip to chip, but it definitely can happen. A minor 5C increase in ambient temp is not going to cause the voltages to need 0.020v more. Both of my 2600k's have degraded slowly (I'm currently deliberately frying the first one now at 5.3 ghz 1.58v); my second one used to only need 1.404v-1.416v loaded for 5 ghz, now it needs around 1.440-1.452v. What caused it? Repeated priming to try to get into the 5 ghz club. And it was NOT rise in ambient temperatures either.

I tested this by going back to 4.5 ghz which used to just need 1.236v load for hours and hours of prime. Now I need 1.260v load for 12 hours prime stable (tested it three times for 36 hours at the 1.260v). That 1.236v BSOD's within minutes now.

The chip never degraded except when I started using it at 5 ghz. All it took was several hours of prime attempts (maybe 12 hours total) at 5 ghz and 1.404v-1.416v load, and a bunch of hours of Bad Company 2 (at around 1.428v load) for it to happen. I'm hoping it has settled at 1.440-1.452v load now. I did leave it idle for 6 hours (had a idle bsod with 1.428v idle, so raised the LLC2 idle voltage to close to 1.440v), then went back and did another 4.5 ghz 1.260v load 12 hour Prime test, which passed fine, but its hard to say if it has settled or not (note that 1.255v BSOD'd in about 9 hours even before the last BC2 barrage and before I did the 3 x 12 hours of 1.260v blends)
I'll find out later when I switch back to that chip, and if I get a idle BSOD at 1.420v set in BIOS with LLC2, at 5 ghz. If I do, then a bit more degradation....

Degradation is real and can happen in a month, easily. Doesn't happen to everyone the same way, but it can happen.

And this degradation is basically the chip needing higher vcore to run at the same clock. The chips usually degrade awhile at a certain vcore, and then level out and stop needing any more vcore increases. But how much a chip will degrade is based directly on the voltages used.
Quote:
Instability is instability, my friend. Like I said, my 2600k ran on 1.26v for two weeks and then couldn't do it anymore. Had to bump it to 1.35v for maximum stability so far.

Yep your chip degraded also, but at these lower voltages, the proper term is more like "breaking in" to level out at a voltage where it will remain stable long term.

I'm deliberately frying my first 2600k just to accelerate this degradation and to see at what voltages the chip will settle down in (at lower overclocks) after it's been subject to massive abuse at 5.3 ghz and almost 1.6v on air. Already seen some more degradation by this; 4.5 ghz @ 1.29v, which used to be stable on this one, now will BSOD in under 30 minutes of prime blend even at 1.303v.

Once my 5.3 ghz won't be able to run games at 1.6v anymore (due to bsod's), I'll back it down to 5.2 ghz, then 5.1, etc, and keep adjusting vcore (higher/lower as long as I don't go above 1.6) to see if or when the chip's vcores settle (at 4.5 and 5 ghz). At the very least, i will have learned something nice.

BTW one person folded 24/7 at 5 ghz 1.52v for three months, and now his chip can't do higher than 4.6 ghz at that voltage.

There are a bunch of other people (in the P67 Gigabyte review thread) who also saw degradation after awhile (like a week or two) of priming tests and other usage, at just 1.3-1.35v. So yes, it happens, folks. But it shouldn't KEEP happening forever unless you're at very high voltages.
Falkentyne is offline  
Sponsored Links
Advertisement
 
post #22 of 56 Old 05-19-2011, 04:06 AM
Semi-Retired OCer
 
PROBN4LYFE's Avatar
 
Join Date: Feb 2005
Location: Northern Virginia
Posts: 10,394
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 603
Quote:
Originally Posted by E-Peen;13559445 
eek.gif

I thought it was 1.46v for 24/7 use, and in due time degradation would follow.
Yeah, but when did we start using safe voltages in overclocking?confused.gif
No serious OC results in staying "inside" manufacturer's perameters. Also your BIOS settings are going to regulate this as well as stated before.

Now this VCCIO voltage is related to which setting on my MSI BIOS
CPU I/O or DRAM voltage?

PROBN4LYFE is offline  
post #23 of 56 Old 05-19-2011, 04:10 AM
4.0ghz
 
Falkentyne's Avatar
 
Join Date: Dec 2007
Location: Riverside
Posts: 4,042
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 357
These chips can "Degrade" even with low voltages!
The point is, there is a point where they will degrade initially and then settle down. Only at high voltages will the degradation get worse (or just keep repeatedly degrading), unless you can keep the chip very cool.

And the "safe 24/7" settings depend purely on your cooling. Good air is fine up to about 1.4v. Best for 1.4v+ is cold water. Best for 1.5v is subzero. This won't "stop" all degradation (as shown above, some people have had minor degradation even at 1.3v), but will prevent the chip from "Spiral degradation" out of control.
Falkentyne is offline  
Sponsored Links
Advertisement
 
post #24 of 56 Old 05-19-2011, 06:08 AM - Thread Starter
1.3ghz
 
Join Date: Jan 2011
Posts: 283
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 2
Quote:
Originally Posted by Falkentyne;13559849 
These chips can "Degrade" even with low voltages!
The point is, there is a point where they will degrade initially and then settle down. Only at high voltages will the degradation get worse (or just keep repeatedly degrading), unless you can keep the chip very cool.

And the "safe 24/7" settings depend purely on your cooling. Good air is fine up to about 1.4v. Best for 1.4v+ is cold water. Best for 1.5v is subzero. This won't "stop" all degradation (as shown above, some people have had minor degradation even at 1.3v), but will prevent the chip from "Spiral degradation" out of control.

Thats the thing. My chip is under water.. Why I think it's strange. But thanks for your posts, guess it is degradation after all.

LGA 1155 Owners Club


.-[SandyBridge]-.
.5ghz Pl+s Club.

http://valid.canardpc.com/show_oc.php? id=1692162 [@ 5.1 GHz]

http://www.overclock.net/water-cooli...l#post13103058 [Build pics]
IzorkX is offline  
post #25 of 56 Old 05-19-2011, 03:44 PM
 
Join Date: Nov 2009
Location: Detroit
Posts: 3,318
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 200
I love my old power hunger rig more. i was going to get sb , thank god i did not. Yes x58 chips can take 1.45 volts, thats 45nm, the sb chips are smaller and volts in chip degrad it fast @ high volts. I am hopeing its not a cpu problom and more like overvolting. I never whent to 1.4 volts on my chips ok i did on old 920c0 for 4.2 ghz when i 1st got it @ 1.45 volts. but that was prime 20 mins then back to 3.8ghz, now 930 chip @ 4ghz @ 1.275 volts (looked @ bios) the safte volts for me on smaller die cpu would be 1.35 volts regardless what ppl say.

Well op sound like u need to change the sig, to cpu degradeing/ not stable. sorry to hear for your loss. but 2600k @ microcenter is $229 now tho better folder for ya
GfhTattoo is offline  
post #26 of 56 Old 05-19-2011, 03:54 PM
Extreme Cooler
 
Ando's Avatar
 
Join Date: May 2007
Location: Australia
Posts: 943
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 55
Id say it was just settling/wearing in - My chips usually need a bump in voltage after running for a few weeks, After that they stay stable on set voltage. (unless I bench the crap out of it @ 1.6v+)
Ando is offline  
post #27 of 56 Old 05-19-2011, 04:01 PM - Thread Starter
1.3ghz
 
Join Date: Jan 2011
Posts: 283
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 2
Quote:
Originally Posted by GfhTattoo View Post
I love my old power hunger rig more. i was going to get sb , thank god i did not. Yes x58 chips can take 1.45 volts, thats 45nm, the sb chips are smaller and volts in chip degrad it fast @ high volts. I am hopeing its not a cpu problom and more like overvolting. I never whent to 1.4 volts on my chips ok i did on old 920c0 for 4.2 ghz when i 1st got it @ 1.45 volts. but that was prime 20 mins then back to 3.8ghz, now 930 chip @ 4ghz @ 1.275 volts (looked @ bios) the safte volts for me on smaller die cpu would be 1.35 volts regardless what ppl say.

Well op sound like u need to change the sig, to cpu degradeing/ not stable. sorry to hear for your loss. but 2600k @ microcenter is $229 now tho better folder for ya
Indeed.. Unfortunately I live in Denmark, so I don't have the pleasure of a microcenter. Prices here is crazy. A 2600K costs 400USD ex delivery the cheapest place.. Wish I lived in US :=)

For now I will stay at my 4.5GHz 1.28V overclock, which seems to be stable

LGA 1155 Owners Club


.-[SandyBridge]-.
.5ghz Pl+s Club.

http://valid.canardpc.com/show_oc.php? id=1692162 [@ 5.1 GHz]

http://www.overclock.net/water-cooli...l#post13103058 [Build pics]
IzorkX is offline  
post #28 of 56 Old 05-19-2011, 06:53 PM
News Fiend
 
Scorpion667's Avatar
 
Join Date: Dec 2007
Location: Toronto, Canada
Posts: 3,242
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 120
Quote:
Originally Posted by pioneerisloud;13559418 
Okay, so you ran Prime for 12 hours. What settings? I can pass for 12 hours just fine with Small FFT's or blended just fine with 1.39v still. It requires a "Custom" blend forcing almost all my RAM to make mine BSOD with a 124 error. GTA IV seems to cause that error quite quickly too.

I'm interested in what settings you use to custom blend!

Mine can pass blend at 1.43v for 8 hours, yet requires 1.45v for full stability. That's why I ran AVX Linpack although that was only possible for me in the winter months with sub zero ambients. Blend temps I can handle 72c max.

Scorpion667 is offline  
post #29 of 56 Old 05-19-2011, 07:54 PM
Retired Staff
 
pioneerisloud's Avatar
 
Join Date: Jun 2007
Location: Yakima, WA
Posts: 32,714
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 1387
Quote:
Originally Posted by Scorpion667;13567647 
I'm interested in what settings you use to custom blend!

Mine can pass blend at 1.43v for 8 hours, yet requires 1.45v for full stability. That's why I ran AVX Linpack although that was only possible for me in the winter months with sub zero ambients. Blend temps I can handle 72c max.

What I personally do....

Hit "Blend" on Prime95, but do NOT start.
Hit "Custom" (this forces blend settings, but you can change them manually).
Change the Memory to use in MB to 7168 for 7GB or 3072 for 3GB. I use all but 1GB of my RAM. More, is of course better.

Funny Quotes (Click to show)
Quote:
Originally Posted by Syrillian 
...I just ran in and... started ripping off the clothes I had on.
Quote:
Originally Posted by steelbom 
A GTX 580 is like a bear. AA or "small bunnies" don't really hurt it.
Quote:
Originally Posted by Coolio831 
Syrillian - Heroes get remembered, but legends never die!
How To: Install Car Audio to a PC Power Supply
¡¡¡ʍʇɟ qn1ɔ uoıʇɐıɔǝɹddɐ 939 ʇǝʞɔos ǝɥʇ
♫The♪Official♪Car♪Stereo♪Club♫
Pioneerisloud's 92 Prelude Si Build Log!!!!

3 Million+ Folding at Home points  Chimp Challenge 3 x Participant  Overclock.net 10 Year Member 

pioneerisloud is offline  
post #30 of 56 Old 05-20-2011, 10:54 AM
4.0ghz
 
compudaze's Avatar
 
Join Date: May 2010
Location: United States
Posts: 1,269
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 149
Anyone ever stop to think about the context at which Intel proclaims Sandy Bridge's max VID to be 1.52V? IIRC disabling LLC and keeping leaving Vdroop in tact is the spec at which Intel designed these chips to be run. Anyone ever set their CPU voltage to 1.52V and disabled LLC? I haven't yet, but I have at 1.4V and I noticed a Vdroop of roughly 0.12V. If I assume the same Vdroop at 1.52V, that would give me a load voltage of roughly 1.4V. IMO keeping load voltages at or below 1.4V (with LLC or not) is the safest bet for longevity.

Back on topic... OP: have you tried letting the computer idle or even stay off for one day at 5GHz 1.424V? Only reason I ask is that I was 12H prime blend stable at 4.8GHz 1.36V, but then I did a 5.4GHz 1.575V 4H bench session. After wards, I went back to 4.8GHz 1.36V to find prime 95 blend crashed within ~2 mins. I let the PC sit overnight and the next day. Powered it up ~20 hours later and it was 12H prime95 blend stable again at 4.8GHz 1.36V.
compudaze is offline  
Reply

Quick Reply

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off