1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Lifespan of CPUs, GPUs. Varies or not?

Discussion in 'Overclocking & Cooling' started by itsakjt, Jan 21, 2013.

  1. itsakjt

    itsakjt

    Joined:
    Oct 9, 2010
    Messages:
    1,171 (0.84/day)
    Thanks Received:
    372
    Location:
    Kolkata, India
    I have seen that say a CPU when overclocked and it exceeds its stability point at stock voltage(e.g. A 3.0 GHz CPU max stable at 3.7 GHz at stock volts) malfunctions the way an aged(dying) CPU does. Does this mean that a CPU having more overclocking headroom(space between the failure start frequency and stock frequency) has a greater lifespan(MTBF) than the same CPU having a less overclocking headroom at stock voltage considering both operate at stock freq and voltages?
  2. cadaveca

    cadaveca My name is Dave

    Joined:
    Apr 10, 2006
    Messages:
    13,770 (4.54/day)
    Thanks Received:
    6,845
    Location:
    Edmonton, Alberta
    Technically, yes, a better chip will last longer.

    At stock, current chips should last 3x the warranty period, or about 10 years or more. I think a board would fail first, might take the CPU with it.:p MOSFETs/Capacitors are the most likely thing to fail, IMHO.
  3. itsakjt

    itsakjt

    Joined:
    Oct 9, 2010
    Messages:
    1,171 (0.84/day)
    Thanks Received:
    372
    Location:
    Kolkata, India
    Yeah. :) Or maybe the PSU taking everything with it. :p
  4. cadaveca

    cadaveca My name is Dave

    Joined:
    Apr 10, 2006
    Messages:
    13,770 (4.54/day)
    Thanks Received:
    6,845
    Location:
    Edmonton, Alberta
    Exactly. These IVB chips are pretty tough, I must say, 1.15V stock, and I push 1.4 V on air no problem. That's like a what...25% voltage boost? Power draw goes from ~~70 W to 120W...near double...and only a 25% boost in clocks.

    :roll:
    itsakjt says thanks.
  5. W1zzard

    W1zzard Administrator Staff Member

    Joined:
    May 14, 2004
    Messages:
    14,645 (3.93/day)
    Thanks Received:
    11,383
    in theory the lifespan will be shorter, i have never seen any evidence of it in real life.

    silicon processors will be obsolete long before they fail from old age
    syeef, newtekie1, de.das.dude and 2 others say thanks.
  6. itsakjt

    itsakjt

    Joined:
    Oct 9, 2010
    Messages:
    1,171 (0.84/day)
    Thanks Received:
    372
    Location:
    Kolkata, India
    LOL yes. :laugh: Actually the thing is the maximum overclocking efficiency is there when you are at stock volts only but push clocks at the max. That way you are actually getting more speed without using more power. :)

    For me, it was even better. I clocked my Phenom II X4 955 at 3.7 GHz using 1.375V(undervolted :p 1.4V stock) and never had a problem. However just for a mere 200 MHz more that is 3.9 GHz I had to set the voltage to 1.45V or 1.378 V LLC on as in my specs. For another 100 MHz, I need greater than 1.5 V even(less when LLC on).
  7. itsakjt

    itsakjt

    Joined:
    Oct 9, 2010
    Messages:
    1,171 (0.84/day)
    Thanks Received:
    372
    Location:
    Kolkata, India
    @W1zzard-That's true. But I know some people here in Kolkata, India who still uses 478 pin Pentium 4. :p
    Last edited: Jan 21, 2013
  8. brandonwh64

    brandonwh64 Addicted to Bacon and StarCrunches!!!

    Joined:
    Sep 6, 2009
    Messages:
    18,465 (10.33/day)
    Thanks Received:
    6,009
    Location:
    Chatsworth, GA
    I still have a pentium 4 that is on a motherboard with 4 busted caps (yes stuff oozing out the top) but it still works 24/7 as my home router.
    itsakjt says thanks.
    Crunching for Team TPU
  9. de.das.dude

    de.das.dude Pro Indian Modder

    Joined:
    Jun 13, 2010
    Messages:
    7,519 (4.99/day)
    Thanks Received:
    1,948
    i have seen some scaringly old processors! they just dont die! i know an old sempron and motherboard in a cousins PC thats 9 years old!
  10. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    39,533 (13.48/day)
    Thanks Received:
    13,929
    I have a 80386 in my alignment machine at work still kicking.
    AlienIsGOD says thanks.
  11. Frick

    Frick Fishfaced Nincompoop

    Joined:
    Feb 27, 2006
    Messages:
    10,427 (3.39/day)
    Thanks Received:
    2,113
    9 years is NOTHING for a CPU, especially not when it's a PC. They idle most of the time.
  12. McSteel

    McSteel

    Joined:
    Nov 19, 2012
    Messages:
    565 (0.91/day)
    Thanks Received:
    273
    CPU lifetime is calculated at continuous full load, at a stock voltage range and at a certain temperature (varies by CPU type). Lower any of the three, and you get quite an extension of that predicted life. CPU warranty period is the absolute worst-case scenario of all parameters. As it has been said, in a typical usage scenario, a CPU will be long obsolete before it dies.

    On the other hand, LN2 and > 50% overvolting + ~100% overclocking will usually bring the lifetime down quite considerably. I've seen actual examples of this, though it was just certain parts of the CPU failing (just one of the cores, or the L3 cache), never seen an entire CPU outright stop working. It is worth noting that going sub-zero is as bad as going over 50°C, because you increase the electrical conductivity of the P-N junctions, which makes electromigration danage spike to problematic levels.
    itsakjt says thanks.
  13. james888

    james888

    Joined:
    Jun 27, 2011
    Messages:
    4,281 (3.79/day)
    Thanks Received:
    1,423
    I thought the lifespan had to do with electromagnetic mitigation that increased with temps, so that if you had lower temps it would increase the lifespan. I do not have a source for this, I just remember from something I read quite some time ago.
    Crunching for Team TPU

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page