Searching \ for '[EE] Overclocking and raising voltage on CPUs' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/timers.htm?key=clock
Search entire site for: 'Overclocking and raising voltage on CPUs'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Overclocking and raising voltage on CPUs'
2011\12\11@152822 by V G

picon face
People generally tend to spew a lot of crap on the Internet regarding
overclocking CPUs, so I avoid reading answers on those forums.

Q: From an engineering perspective, if I overclock my core i7 from 3.8GHz
to around 4.4GHz and increase the voltage to support the higher
frequencies, what effect does it have on the lifespan of the CPU (assuming
I cool it adequately to under 55 degrees under full load)? Do the higher
than standard voltages slowly damage the inside of the CPU? I don't really
know anything about the insides of ICs yet..

2011\12\11@164119 by Herbert Graf

picon face
On Sun, 2011-12-11 at 15:28 -0500, V G wrote:
> People generally tend to spew a lot of crap on the Internet regarding
> overclocking CPUs, so I avoid reading answers on those forums.
>
> Q: From an engineering perspective, if I overclock my core i7 from 3.8GHz
> to around 4.4GHz and increase the voltage to support the higher
> frequencies, what effect does it have on the lifespan of the CPU (assuming
> I cool it adequately to under 55 degrees under full load)? Do the higher
> than standard voltages slowly damage the inside of the CPU? I don't really
> know anything about the insides of ICs yet...

The answer is yes, it does reduce the lifetime of the CPU.

There are a few reasons for that, and it varies TREMENDOUSLY on the
particular piece of silcon involved.

That said: CPUs are engineered to last a VERY VERY long time. It's not
surprising to hear a CPU that's 30 years old still running without
issue.
So, you are taking a CPU, designed to run for decades, and probably
affecting it's life by at most a few years. Does that matter?

Chances are, even in the WORST case, your CPU will still outlive, by
far, it's useful life (10 years from now it's very unlikely you'll care
if that particular chip fails).

My experience with overclocking has been only positive. I say go for it.
The worst that happens in my experience is the particular piece of
silicon you get won't overclock as much as others.

TTYL

2011\12\11@165313 by V G

picon face
On Sun, Dec 11, 2011 at 4:41 PM, Herbert Graf <spam_OUThkgrafTakeThisOuTspamgmail.com> wrote:
> The answer is yes, it does reduce the lifetime of the CPU.
>
> There are a few reasons for that, and it varies TREMENDOUSLY on the
> particular piece of silcon involved.

Would you please share some of those reasons?

{Quote hidden}

Thanks! Will do, boss

2011\12\11@172251 by Sean Breheny

face picon face
While I agree that moderate overclocking is probably not going to
cause quick death of the CPU, I would think that the bigger problem is
reliability - there will be some range of frequencies where the
processor works most of the time but will sometimes hang, sometimes
just begin behaving erratically, etc. I would think that it would be
difficult to determine a very reliable operating point yourself. I
would think, too, that the manufacturer, since they test every single
chip they produce, would have rated that chip close to its actual max
frequency for reliable operation because they often "bin" processors
according to how high the test can go in frequency before it starts to
fail.

Sean


On Sun, Dec 11, 2011 at 4:41 PM, Herbert Graf <.....hkgrafKILLspamspam@spam@gmail.com> wrote:
{Quote hidden}

>

2011\12\11@184529 by Peter Johansson

picon face
On Sun, Dec 11, 2011 at 5:22 PM, Sean Breheny <shb7spamKILLspamcornell.edu> wrote:

> While I agree that moderate overclocking is probably not going to
> cause quick death of the CPU, I would think that the bigger problem is
> reliability - there will be some range of frequencies where the
> processor works most of the time but will sometimes hang, sometimes
> just begin behaving erratically, etc. I would think that it would be
> difficult to determine a very reliable operating point yourself.

Most over-clockers are gamers looking for maximum frame rates, and the
ocasional glitch/crash of a game is often an acceptable trade-off for
increased frame rate.

Personally, when I care about the results of my CPUs calculations, I'd
prefer to have as much headroom
as possible.

-p

2011\12\12@052940 by Electron

flavicon
face
At 21.28 2011.12.11, you wrote:
>People generally tend to spew a lot of crap on the Internet regarding
>overclocking CPUs, so I avoid reading answers on those forums.
>
>Q: From an engineering perspective, if I overclock my core i7 from 3.8GHz
>to around 4.4GHz and increase the voltage to support the higher
>frequencies, what effect does it have on the lifespan of the CPU (assuming
>I cool it adequately to under 55 degrees under full load)? Do the higher
>than standard voltages slowly damage the inside of the CPU? I don't really
>know anything about the insides of ICs yet...

Yes, possibly by electromigration, but before it really damages your CPU,
you will think that 10 years have passed since when you thought it was
totally obsolete.

3.8GHz -> 4.4GHz and some more mV shouldn't concern you really. Go for it,
provided you can cool it enough, that shall be your only worry. I have made
much more extreme overclocking (e.g. a Pentium II from 350MHz to 530MHz)
it has been like that for thousands hours and if I turn it on it still works..

2011\12\12@120228 by Herbert Graf

picon face
On Sun, 2011-12-11 at 16:52 -0500, V G wrote:
> On Sun, Dec 11, 2011 at 4:41 PM, Herbert Graf <.....hkgrafKILLspamspam.....gmail.com> wrote:
> > The answer is yes, it does reduce the lifetime of the CPU.
> >
> > There are a few reasons for that, and it varies TREMENDOUSLY on the
> > particular piece of silcon involved.
>
> Would you please share some of those reasons?

It depends on the defects. Each piece of silicon has defects. Some are
in places that cause the CPU not to function. Others are in areas where
it doesn't matter, or can be fused around (i.e. for quad core parts, if
one core has a defect that affects it, to save the die, they fuse off
that core and sell the part as a tri-core).

There are also defects that only affect operation above certain
voltages/speeds, so they sell those parts as lower speed rated parts.

Defects also tend to "migrate" over time, this is what will eventually
kill a piece of silicon. Higher voltages and temps cause these to
effects to happen faster (think of pouring sugar into a glass of water,
the sugar will disolve much faster if that water is hot then if it's
cold).

I'm not expert on this, where I work in the chain is well before a chip
gets made, so I'm sure there are other factors that people here will be
able to mention or clarify.

Thanks, TTYL

More... (looser matching)
- Last day of these posts
- In 2011 , 2012 only
- Today
- New search...