Haswell, IVY/SANDY bridge chip degradation questions

Discussion of all forms of processors, from AMD to Intel to VIA.

Moderators: Flying Fox, morphine

Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sat May 03, 2014 4:05 pm

Okay I was wondering if we could get some 1st hand reports from you gerbils out there on CPU chip degradation where you now need more voltage for a overclock that was stable before but not anymore or overclock that were attainable but not anymore etc. Also if more voltage is needed for stock operation.

I know my 2600k has not to my knowledge degraded in any way whatsoever and have not even heard of a Sandy chip degrading. Even Ivy bridge with its much hotter temps I have not heard much of anything with degradation.

But degradation on Haswell chips has caught my attention and I wonder if it is from the built in voltage regulators in the chip itself getting cooked and loosing their ability to regulate correctly.

I would like to hear any 1st hand instances of degradation on Sandy/IVY and Haswell chips any of you gerbils have had 1st hand exp with. I am very curious to know If when I ramp my 2600k to 5ghz + if it is going to kill it or damage it once I get cooling that will let me run at 5ghz under 70c. Since day one my 2600k has never broke 75c and I always keep its temps at a max of around 65c since day one beside overclocking experiments that heated it up to 75+ C but I have yet to throttle my chip.
2600k HT on@4705mhz 8gb Cas9 1600 mem 2x EVGA GTX770 4gb Classified cards in SLI @1320 mhz core and 2003 mhz mem,mounted in CM HAF922 with a TX-850 PSU 2xHTPC's 2xi3 2120 3.3ghz dual core,1xasus LP HD6570 1xHIS hd7750@1150core1325mem,55"PanyVT30
vargis14
Graphmaster Gerbil
 
Posts: 1194
Joined: Fri Aug 20, 2010 6:03 pm
Location: philly suburbs

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 11:24 am

Higher voltages cause increased electromigration but I've never heard of a chip degrade through electromigration. It'll just suddenly go *pop*, irrecoverably so.

You might want to do some reading about NBTI - Negative Bias Temperature Instability; Older transistors generate more heat and take longer to switch. This means that a CPU capably of 5GHz when new may not be stable at 5GHz when old. I didn't think that NBTI would kill a chip, though - more likely your max overclock will decrease over time and you'll be able to regain stability by dialing back the overclock slightly.

Anecdotes I've read say that a chip overclocked to its limits might last less than half the normal lifespan, but since I've seen 15-year-old celerons still running strong despite a 50% overclock all that time, I doubt that halving the lifespan is going to be much of an issue for you. Sadly I have no experience of my own - the only chip deaths I've seen have been related to motherboard failures, power surges or cracked dies.
<insert large, flashing, epileptic-fit-inducing signature (based on the latest internet-meme) here>
Chrispy_
Gerbil Jedi
Gold subscriber
 
 
Posts: 1884
Joined: Fri Apr 09, 2004 3:49 pm

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 11:49 am

Thanks for your input Chrispy
This is the reason I asked the TR community. But it seems like no one wishes to give me any input or have had any cases of their chips degrading. But I find that highly unlikely with the size of the TR community.
2600k HT on@4705mhz 8gb Cas9 1600 mem 2x EVGA GTX770 4gb Classified cards in SLI @1320 mhz core and 2003 mhz mem,mounted in CM HAF922 with a TX-850 PSU 2xHTPC's 2xi3 2120 3.3ghz dual core,1xasus LP HD6570 1xHIS hd7750@1150core1325mem,55"PanyVT30
vargis14
Graphmaster Gerbil
 
Posts: 1194
Joined: Fri Aug 20, 2010 6:03 pm
Location: philly suburbs

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 12:15 pm

It could be that such cases are in your head, rather than being real :). I've certainly not had any evidence of my 2500K @ 4.5, or the 2500Ks owned by 3 of my friends "degrading" over time, and they're all overclocked but reasonably well cooled. Likewise I know of a 2.13ghz Core2 Duo running at 3.2ghz since the start of 2007 without requiring any extra voltage or tampering with.
i5-2500k @ 4.5ghz | ASUS P8Z68-V | nVidia GTX780 | 32gb DDR3| 2x SSDs 2x HDD | Antec Solo II | Dell U2713HM
Mentawl
Gerbil XP
Silver subscriber
 
 
Posts: 449
Joined: Sun Dec 26, 2004 5:21 pm
Location: UK

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 12:24 pm

Or maybe we just don't overclock cuz there is no need to. My PC running at base is more than enough for all the compiling I do.
A fact is a simple statement that everyone believes. It is innocent, unless found guilty. A hypothesis is a novel suggestion that no one wants to believe. It is guilty, until found effective.
Edward Teller
SomeOtherGeek
Gerbil First Class
Gold subscriber
 
 
Posts: 166
Joined: Fri Jun 06, 2008 7:49 pm
Location: Outside somewhere

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 1:15 pm

SomeOtherGeek wrote:Or maybe we just don't overclock because there is no need to do so.
This. Games are mostly still GPU-bound. The other day, I noticed reduced framerates during a Guild Wars 2 mega-server world boss event. I thought that perhaps rendering hundreds of attacking players at one time was taxing my CPU and that maybe I should overclock it from its default. Not so. The CPU wasn't hitting 20% load. In fact, the Core i7-4770K was loafing along at around 2.2 GHz most of the time. Meanwhile, my Radeon R9-290 was running full tilt. I bumped up the overclock (and the cooling) on the GPU and left the CPU alone.
i7-4770K, H70, Gryphon Z87, 16 GiB, R9-290, SSD, 2 HD, Blu-ray, SB ZX, TJ08-E, SS-660XP², 3007WFP+2001FP, RK-9000BR, MX518
JustAnEngineer
Gerbil God
Gold subscriber
 
 
Posts: 15440
Joined: Sat Jan 26, 2002 7:00 pm
Location: The Heart of Dixie

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 1:38 pm

JustAnEngineer wrote:
SomeOtherGeek wrote:Or maybe we just don't overclock because there is no need to do so.
This. Games are mostly still GPU-bound. The other day, I noticed reduced framerates during a Guild Wars 2 mega-server world boss event. I thought that perhaps rendering hundreds of attacking players at one time was taxing my CPU and that maybe I should overclock it from its default. Not so. The CPU wasn't hitting 20% load. In fact, the Core i7-4770K was loafing along at around 2.2 GHz most of the time. Meanwhile, my Radeon R9-290 was running full tilt. I bumped up the overclock (and the cooling) on the GPU and left the CPU alone.


For gaming certainly doesn't make much sense, but for heavy video encoding it makes all sense. It will always depend on what your are doing.
Intel i5 3570K - Corsair A50 - Asus P8Z77-V - Corsair 2x4Gb - Asus 280X Direct CU2 - 2x Corsair F90 - Samsung Spinpoint F3 1Tb - Antec One Hundred - Corsair CS750M - ASUS MX239H AH-IPS - Logitech LS21 - Logitech K520 Combo - Win 8.1
Nokia Lumia 920
Jon1984
Gerbil XP
Silver subscriber
 
 
Posts: 314
Joined: Wed Mar 23, 2011 3:07 pm
Location: Portugal

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 2:25 pm

I recently had a theory that 'chip degradation' - as in lower stable max overclocks - may not necessarily be due to the CPU itself, although chips that have experienced long-term high borderline throttle temps or excess voltage will degrade for the reason Chrispy_ said. My theory is that there could be degradation in motherboard components or other non-CPU components that cause a 'ragged edge' overclock to become unstable. That might apply less to Haswell with integrated VRMs, but could still apply because the motherboard still has to do some stuff, but for non-integrated VRM CPUs it could certainly be the case.

Anyway, yes there are clear cases of 'degradation' - whether it's the CPU itself or something else, it does happen.
MadManOriginal
Graphmaster Gerbil
 
Posts: 1431
Joined: Wed Jan 30, 2002 7:00 pm
Location: In my head...

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 4:02 pm

My P35-DS3R with blackened/darkened PCB around the CPU VRM is still running a Q6600 @ 3.0GHz on stock voltage. :) Years ago when pushing it really hard I didn't realize the power involved and how it would affect the heat sink-less MOSFETs. I actually ended up melting the 4 pin CPU aux power jack and had to replace it. Seems that the resistance of the connector and the other external heat didn't get along.

I guess Gigabyte did in fact use some durable MOSFETs.
swaaye
Gerbil First Class
 
Posts: 174
Joined: Mon Apr 21, 2003 4:45 pm

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 6:44 pm

I'd like to second what MMO put forth. This isn't evidence because it's not data, but here's my story...

1. Overclock Phenom II/ 955 to ~ 3.7 GHz, run fairly intensive floating point calculations in matlab (runtime ~ 1-2 days). You could push the processor to 3.9 GHz without a problem, but I usually back off a little.
2. Stop running intensive floating point calculations in matlab. Leave processor overclocked. No stability problem, but, otoh, the calculations are the only real stress I'd ever put on the FPUs.
3. After a couple of years, go back to do the same calculations. CPU temps had crept up a bit in the meantime (or so I'd decided, based on what happened next). Not stable.
4. For other reasons, I bought a new motherboard. I didn't buy new RAM. I did re-seat the heatsink, and replaced the fan. CPU temps went down something like 2-3 degrees, which is nothing, as far as I'm concerned. The same calculation worked again at 3.7GHz, and even 3.8 GHz.
d0ugmelvin
Gerbil
Silver subscriber
 
 
Posts: 10
Joined: Tue Mar 18, 2008 8:42 am

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 7:40 pm

MadManOriginal wrote:I recently had a theory that 'chip degradation' - as in lower stable max overclocks - may not necessarily be due to the CPU itself, although chips that have experienced long-term high borderline throttle temps or excess voltage will degrade for the reason Chrispy_ said. My theory is that there could be degradation in motherboard components or other non-CPU components that cause a 'ragged edge' overclock to become unstable. That might apply less to Haswell with integrated VRMs, but could still apply because the motherboard still has to do some stuff, but for non-integrated VRM CPUs it could certainly be the case.

Anyway, yes there are clear cases of 'degradation' - whether it's the CPU itself or something else, it does happen.


M M Original,
The reason I asked was since Haswell has tiny integrated VRMs that are getting much hotter then big 8, 10, 12 + phase motherboard VRM designs been really the only chips I have really heard anything of degrading with overclocks and I think it would make degrading apply more to Haswell because of those tiny integrated VRM's. I for one have not heard of any degrading on Sandy/Ivy chips that use the big 8, 10, 12 + phase motherboard VRM designs. That is why I asked about Sandy/Ivy bridge and Haswell. Ivy bridge chips run pretty much as hot as Haswell but have not heard of any degrading of Ivy...only Haswell with it built in VRM design that is great for notebooks and laptops and other low power high performance chips. But it has no place on the desktop where it only saves a few watts of power.
it make me wonder if Haswell-E will have built in voltage regulators?? I guess we will find out soon enough.

As for overclocking for me, I do do a decent amount of encoding. Overclocking gaming wise BF3 multiplayer it definitely helps. Along with pushing 2 fairly heavy duty 770s in SLI depending on the game.
But the main reason for needing my overclock is The Smooth Video Program or SVP player addon for MPC-HC and pot player etc that converts 24, 25 30 FPS videos to your monitors refresh rate on the fly up to 120 even 144fps I think. But is 60fps in my case. It can bring my 2600k at 4635mhz to its knees running a 1920-800 or 1920-1080 video file when you turn up the anti artifacting masking features and other video improvement tweaks. You might say/ask why do you need all those tweaks and it is a fair question. I need those tweaks cause I am on a huge 55" tv that we are sitting 4-6ft from depending on if I pull the tv off the wall 3 ft with its wall mount. But the improvement in video quality is amazing along with the smoothness from going from 24 to 60 fps is pretty dramatic. On a smaller monitor/TV you do not notice a lot of stuff because the pixels are so much smaller unless you are 12" away from the screen.
2600k HT on@4705mhz 8gb Cas9 1600 mem 2x EVGA GTX770 4gb Classified cards in SLI @1320 mhz core and 2003 mhz mem,mounted in CM HAF922 with a TX-850 PSU 2xHTPC's 2xi3 2120 3.3ghz dual core,1xasus LP HD6570 1xHIS hd7750@1150core1325mem,55"PanyVT30
vargis14
Graphmaster Gerbil
 
Posts: 1194
Joined: Fri Aug 20, 2010 6:03 pm
Location: philly suburbs

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 8:13 pm

vargis14 wrote:But it seems like no one wishes to give me any input or have had any cases of their chips degrading. But I find that highly unlikely with the size of the TR community.


Hey, you hurt my feeling, man!

I have not had any CPUs degrade in the manner you describe, no matter how excessively overclocked. My clocks are stable at day #1 and stable at day #3650, or they were never really stable at all.

That said, I've watched a full generation of M4D 1337 CL0X0R K1NG$ come to fruition after my heyday, and half of them could not tell you a stable machine from a dodgy one. THIRTY MINUTES DOES NOT A STRESS TEST MAKE!
Siglessness is boring.
Image - M4800-Eight1
Image - Vargr-Z97
Forge
Lord High Gerbil
 
Posts: 8038
Joined: Wed Dec 26, 2001 7:00 pm
Location: SouthEast PA

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Sun May 04, 2014 10:10 pm

Well all I can say is the rig in my sig has been running OCed for over 15 years with nary a problem. YMMV :D
Image
Abit BP6 2X 300A@450 mhz, 1.5 GB PC133 ECC, Renditon Verite V2200, SB AWE32, Adaptec 2940U2W, 4.3GB Quantum Atlas 10K, Plextor Ultraplex
Starfalcon
Gerbilus Supremus
 
Posts: 11712
Joined: Mon Oct 14, 2002 10:43 am
Location: Evergreen Park, IL

Re: Haswell, IVY/SANDY bridge chip degradation questions

Postposted on Mon May 05, 2014 9:38 am

Starfalcon wrote:Well all I can say is the rig in my sig has been running OCed for over 15 years with nary a problem. YMMV :D


Lol, that's the same as the Celeron I was talking about earlier :P
I built it to play Half-Life and sold it cheap to a friend's brother, who passed it on to his dad who still has it. Last Christmas I used it (still stable at 100MHz FSB) to instruct their family on the merits of Jim'll Paint it!
<insert large, flashing, epileptic-fit-inducing signature (based on the latest internet-meme) here>
Chrispy_
Gerbil Jedi
Gold subscriber
 
 
Posts: 1884
Joined: Fri Apr 09, 2004 3:49 pm


Return to Processors

Who is online

Users browsing this forum: No registered users and 2 guests