How safe is undervolting?

All about them.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
=assassin=
Posts: 243
Joined: Thu Aug 25, 2005 2:46 am
Location: Blackpool, England, UK
Contact:

How safe is undervolting?

Post by =assassin= » Fri Sep 23, 2005 8:40 am

I know over-volting can be bad - increased heat and voltages above what a processor is designed for. But is undervolting 100% safe - i.e. are there any side effects from testing the limit of the processor at a low voltage and at the same frequency, or is it, as i assume, perfectly safe?

I'm just curious...

threevok
Posts: 83
Joined: Fri Nov 12, 2004 2:12 pm
Location: Thornton, Colorado USA

Post by threevok » Fri Sep 23, 2005 9:04 am

You want to make sure you keep the voltage positive or you might unprocess some work you have already done.

Other than that I've never heard of undervolting doing damage on anything other than an inductive (motor) load. Run Prime95 to check stability.

Tibors
Patron of SPCR
Posts: 2674
Joined: Sun Jul 04, 2004 6:07 am
Location: Houten, The Netherlands, Europe

Post by Tibors » Fri Sep 23, 2005 9:05 am

If you mean: Is it safe for my processor and other hardware?
Then the answer is: Yes.

If you mean: Is it safe for my data?
Then the answer is: Like with overclocking, you have to thoroughly test your new settings, else you might have problems.

=assassin=
Posts: 243
Joined: Thu Aug 25, 2005 2:46 am
Location: Blackpool, England, UK
Contact:

Post by =assassin= » Mon Sep 26, 2005 3:56 am

I'm confused: why would undervolting the CPU affect data storage?

qviri
Posts: 2465
Joined: Tue May 24, 2005 8:22 pm
Location: Berlin
Contact:

Post by qviri » Mon Sep 26, 2005 4:01 am

Undervolting may create instability, which may result in crashes, which may corrupt data.

=assassin=
Posts: 243
Joined: Thu Aug 25, 2005 2:46 am
Location: Blackpool, England, UK
Contact:

Post by =assassin= » Mon Sep 26, 2005 4:17 am

lol... oh I see, I wasn't thinking at all!!

Well, it will be on a new build (when I get it) so there's no important data/files to worry about - if I have to reformat because of corruption that's Ok in the testing phase.

I feel like a proper noob now.

How much would an average Venice A64 3000+ easily undervolt to at stock clocks (i.e. not pushing the limit but what they can all likely take as minimum) would at least 1.3 volts be guaranteed?

frostedflakes
Posts: 1608
Joined: Tue Jan 04, 2005 4:02 pm
Location: United States

Post by frostedflakes » Mon Sep 26, 2005 7:52 am

Much lower. More like 1.050V to 1.075V would be my guess.

But to be safe, you'd probably want to start off around 1.150V and lower it in 0.025V increments. As others suggested, be sure to test stability for at least a few hours before lowering the voltage another increment. Once you find what you believe to be your lowest stable undervolt, you may want to let Prime95 (or whatever utility you're using to test CPU stability) run for 24+ hours to make sure it's completely stable. :)

NeilBlanchard
Moderator
Posts: 7681
Joined: Mon Dec 09, 2002 7:11 pm
Location: Maynard, MA, Eaarth
Contact:

Post by NeilBlanchard » Mon Sep 26, 2005 7:59 am

Greetings,

Just think of undervolting as overclocking in reverse: you are keeping the same speed and running with less power. So, things run cooler, to be sure (which is the reverse of overclocking), but you can have the the same challenges of finding the right voltage for stability.

=assassin=
Posts: 243
Joined: Thu Aug 25, 2005 2:46 am
Location: Blackpool, England, UK
Contact:

Post by =assassin= » Tue Sep 27, 2005 2:51 am

Excellent - thanks for the additional information; I'm amazed I can start that low with the voltage - that would be fantastic!!

Post Reply