Should I turn to the computer off???

Scott M Sheaf

unread,

Jun 29, 1993, 9:23:55 PM6/29/93

to

I am not sure if this is the right newsgroup but I am
sure there are people here who can help me.

A friend of mine is trying to convince me that
by not turning the computer off when I am done is doing
more harm than good. Now I leave my computer running for
days at a time perhaps using it 5 or 6 times a day.
Does anyone have an answer to this question?
When should you turn the computer off and when should
you leave it running? Does it really cause a lot of
weear and tear to leave a computer on?

Help.

Scott
Please e-mail
sms...@magnus.acs.ohio-state.edu
--

E-mail: sms...@magnus.acs.ohio-state.edu

Zack Evans

unread,

Jun 29, 1993, 11:10:30 PM6/29/93

to

In article <20qpvb$7...@charm.magnus.acs.ohio-state.edu> sms...@magnus.acs.ohio-state.edu (Scott M Sheaf) writes:

>A friend of mine is trying to convince me that
>by not turning the computer off when I am done is doing
>more harm than good.

Well, he's talking rubbish. Continual power cycling is about the fastest way
to cause wear and tear, because of such things as the temperature changes and
transients.

Now I leave my computer running for
>days at a time perhaps using it 5 or 6 times a day.

Well that's OK so long as you can afford the electricity bill :)

>When should you turn the computer off and when should
>you leave it running?

Don't bother turning it off if you are going to use it again in the next couple
of hours, at least, which is what your case sounds like.

Does it really cause a lot of
>weear and tear to leave a computer on?

Not as much as it causes turning it on and off all the time.

>Please e-mail
>sms...@magnus.acs.ohio-state.edu

Done that too.

Zack

--
Zack Evans pyc...@cent1.lancs.ac.uk or zev...@nyx.cs.du.edu (Internet)
pyc...@uk.ac.lancs.cent1 (JANET)

Watch yer bass bins lads, I'm tellin' yer...

Gabriel Velasco

unread,

Jun 30, 1993, 1:16:12 PM6/30/93

to

I've have more systems fail at power-up than at any other time. Just
like light bulbs. I maintained a LAN with 20 IBM PS/2 workstations and
a file server that were up for months at a time with no major hardware
problems. UNIX systems are almost always kept up all the time because
they take so long to boot. The Mac I'm on right now hasn't been turned
off in about 2 months. I think that power spikes that occur during a
power cycle are probably more damaging than leaving the computer on all
the time unless you are doing something to over-burden the computer in
the first place like running it at a higher clock speed than was
intended or something like that.

-=Gabriel=-

Radford Walker

unread,

Jul 2, 1993, 12:04:04 PM7/2/93

to

> I think that power spikes that occur during a
> power cycle are probably more damaging than leaving the computer on all
> the time unless you are doing something to over-burden the computer in

...

Of course leaving the power on is only good for the electronic elements.
The wear on hard disk bearings should be considered too. I operate a bbs
which
runs 24 hours a day. The electronics have never failed, but I go through a
disk about once every 18 months. My other machine is turned on only when I
use it since 1987. No components have ever failed in it.

Does anyone know if there has been any thoughs of "browning up" computers
to
start them. That is turn on the power real low and increase it to
operating
levels. When the power is all the way on the machine could be soft booted.

|----------------------------------------------------------------------------|
| Everything I've said is my own opinion not that of my company or the
echo |
| moderator. Any relation to the truth is pure coincidence. I don't
even try|
| to correct typos and spelling.
|
| Radford Walker - rwa...@atqm.advtech.uswest.com
|
| OS/2 - because 32 bits are terrible things to waste.
|

Cameron Spitzer

unread,

Jul 2, 1993, 5:56:19 PM7/2/93

to

[junking the fouled up attributions in this thread]

>>> I think that power spikes that occur during a
>>> power cycle are probably more damaging than leaving the computer on all

>>> the time...

Modern power supplies are pretty good about not overshooting when
they turn on.
Stuff *that's ready to fail* tends to fail at power up, but generally
that's due to thermal stress, not "power spikes."

>>Of course leaving the power on is only good for the electronic elements.

Actually, there are wearout mechanisms at work in solid state electronics
just operating at constant temperature. Metal ions migrate along heated
glass. Ions un-implant themselves from diffusion wells. Imperfect solder
crystalizes. Imperfect die-wire bonds corrode. Resistor values
drift. Etc. Some of these mechanisms are driven by chemistry and would
happen almost as fast in unpowered gear. But ion migration is driven by
electric fields which are only present in powered up stuff.
(A DRAM designer from NEC was kind enough to explain this to me once,
with microscope photos and all.)

>>The wear on hard disk bearings should be considered too.

Yes. Computers take a life expectancy hit when they're powered up,
but they wear out as they run, too.

>Does anyone know if there has been any thoughs of "browning up" computers
>to
>start them. That is turn on the power real low and increase it to
>operating
>levels. When the power is all the way on the machine could be soft booted.

Gradual power-on makes sense in CRTs, but the market doesn't like
waiting for stuff to warm up.
It makes less sense for solid state electronics, which can wear out
faster at reduced voltage than at full on.
To ameliorate the thermal shock problem you'd have to ramp the juice
up over several minutes. Problem is the heat production is a very
nonlinear function of input voltage.
Another problem is the added power supply complexity would reduce
reliability. You might lose more than you'd gain.

Sleep mode is a different story. Part of the "green" PC effort at EPA
will be to bring laptop-style power management to the desktop.
This is already beginning to happen, as PCMCIA becomes available on
desktop-sized machines.

Cameron

Philip Hurley

unread,

Jul 7, 1993, 11:05:49 AM7/7/93

to

In article <28...@deadmin.ucsd.edu> gvel...@mechanics.ucsd.edu (Gabriel
Velasco) writes:

I manage a PRIME 9755 minicomputer. One day I came to work to discover that
all the print queues had stopped printing. There are six and every single
one would not produce output. Job after job could be queued up but nothing
would come out. We tried restarting them but they still refused to print.

Finally, we called the problem in and while they were talking to me on the
phone, they happened to ask about the date on the queue log files. When they
found out that the date was about three months old, they became very
incredulous. "When was the last time you booted your machine?" they asked.
"Oh, I don't know for sure, about 3 or 4 months," was my innocent reply.
"Don't you boot your machine once a month?" they gasped. "No, we only boot
if there is a need and we have not had a need to reboot."
"Well," came the explanation, "the spool queue has in it a hard coded
maximum number of days (or hours) that it will run. The way to fix your
problem is to reboot your machine so that counter will be reset."

Philip Hurley "It is absurd to say that you are
T.A.E.X. especially advancing freedom when you only
Computer Technology use free thought to destroy free will."
(409) 845-9689 -- Chesterton
p-hu...@tamu.edu I, too, disclaim.

Danny R. Faught

unread,

Jul 7, 1993, 12:31:23 PM7/7/93

to

Sorry, we were on a roll and I just couldn't resist.
--
Danny Faught -- Convex rookie -- MPP OS Test Development
"Everything is deeply intertwingled." (Ted Nelson, _Computer Lib_)

ehh...@staff.tc.umn.edu

unread,

Jul 8, 1993, 2:43:52 PM7/8/93

to

Often my brain fails, and my eyes light up. And then I hear a message:

"cerebral cortex dumped"

Then I hear a creaking noise, then memory failure.

So since we are on a roll, I was wondering if I should be rebooting my
brain every month?

Please, serious replies only.

--
o/ \ / \ / / \o email: chix...@student.tc.umn.edu
/# ##o # o## #\ - or -
/ \ / \ /o\ / |\ / \ c...@mermaid.micro.umn.edu

Danny R. Faught

unread,

Jul 8, 1993, 3:44:14 PM7/8/93

to

In article <C9v03...@news.cis.umn.edu> ehh...@staff.tc.umn.edu () writes:
>Often my brain fails, and my eyes light up. And then I hear a message:
>
>"cerebral cortex dumped"

Hmmm. Your hearing must be directly in the hardware. Or maybe you're
reading the raw audio device at this point.

>Then I hear a creaking noise, then memory failure.

What a feature. I wish my kernel creaked whenever it failed.

>So since we are on a roll, I was wondering if I should be rebooting my
>brain every month?

I usually turn mine off every night, though it's probably true that
starting it in the morning provides a lot of wear and tear.
Otherwise, just do an occasional garbage collection to clear the hung
ptys.

>Please, serious replies only.

I have a strange urge to crosspost to alt.fan.hofstadter.

Nah.

--
Danny Faught -- Convex rookie -- MPP OS Test Development

Because a mind is a terrible thing...
..to keep running 24 hours a day

Paul Tomblin

unread,

Jul 8, 1993, 5:06:52 PM7/8/93

to

ehh...@staff.tc.umn.edu () writes:

>Often my brain fails, and my eyes light up. And then I hear a message:

>"cerebral cortex dumped"

>Then I hear a creaking noise, then memory failure.

That's nothing to worry about - just love.

>So since we are on a roll, I was wondering if I should be rebooting my
>brain every month?

You should reboot between relationships. And purge the non-volatile memorys
as well.

--
Paul Tomblin - formerly {pt{omblin},news}@{geovision.}gvc.com
"Anybody who has spent time in Algonquin Park will likely be disappointed if
they go to heaven" - Algonquin, the Park and its People.

Bruce Ikenaga

unread,

Jul 8, 1993, 8:15:48 PM7/8/93

to

In article <C9v6n...@freenet.carleton.ca> ab...@freenet.carleton.ca (Paul Tomblin) writes:
>
>>Often my brain fails, and my eyes light up. And then I hear a message:
>
>>"cerebral cortex dumped"
>
>>Then I hear a creaking noise, then memory failure.
>
>That's nothing to worry about - just love.
>
>>So since we are on a roll, I was wondering if I should be rebooting my
>>brain every month?
>
>You should reboot between relationships. And purge the non-volatile memorys
>as well.

But flush your cache before rebooting. If you forget, do
"chkbrain /f" in case you have some lost clusters.
----------
Bruce Ikenaga
US mail: Dept. of Math, CWRU, Cleveland, Ohio 44106
Email: b...@po.cwru.edu

Terry Kennedy, Operations Mgr.

unread,

Jul 9, 1993, 4:35:24 AM7/9/93

to

In article <p-hurle...@tamu.edu>, p-hu...@tamu.edu (Philip Hurley) writes:
> "Well," came the explanation, "the spool queue has in it a hard coded
> maximum number of days (or hours) that it will run. The way to fix your
> problem is to reboot your machine so that counter will be reset."

This seems to be a big problem with a lot of science fiction 8-). Authors
write a story where some unattended computer is found to be still operating
dozens (or hundreds, or thousands) of years later. Never mind subtle things
like uncorrectable memory errors, they'll fail from log files overflowing the
storage device, or the console running out of paper, etc...

There was a thread here some time ago about the longest system uptime, and
another one about the oldest file. It would be interesting to see what sort
of design / programming changes would be needed to develop a system which
would truly "run for ages".

On a similar note, you'd think that the "fault tolerant"/"redundant"/what-
ever systems would amass substantially better uptimes. Many of these systems
allow any component to be serviced without shutting down the system, but I
don't see significant (2 orders of magnitude or better) uptime reports from
these boxes. Possibly that's due to software issues, as it's harder to have
dual operating systems, independently coded and maintained, perform the same
task in lockstep. Without this, a software fault will shut down the system.
I remember making a comment to a DEC engineer about the VAXft system OS:
"Walking hand-in-hand, both sides of the processor step off a cliff". Oddly
enough, at the DECUS Symposium where I said that, the registration system (a
VAXft) did exactly that...

Terry Kennedy Operations Manager, Academic Computing
te...@spcvxa.bitnet St. Peter's College, Jersey City, NJ USA
te...@spcvxa.spc.edu +1 201 915 9381

Danny R. Faught

unread,

Jul 9, 1993, 9:43:36 AM7/9/93

to

In article <bxi.370....@po.cwru.edu> b...@po.cwru.edu (Bruce Ikenaga) writes:
>In article <C9v6n...@freenet.carleton.ca> ab...@freenet.carleton.ca (Paul Tomblin) writes:
>>You should reboot between relationships. And purge the non-volatile memorys
>>as well.
>
> But flush your cache before rebooting. If you forget, do
>"chkbrain /f" in case you have some lost clusters.

Or if you have a *real* brain, use fsck /dev/brain ;-)

Lennart Regebro

unread,

Jul 9, 1993, 11:34:00 AM7/9/93

to

In article <1993Jul8.1...@news.eng.convex.com> fau...@convex.com (Danny R. Faught) writes:
>In article <C9v03...@news.cis.umn.edu> ehh...@staff.tc.umn.edu () writes:
>>Often my brain fails, and my eyes light up. And then I hear a message:
>>
>>"cerebral cortex dumped"
>
>Hmmm. Your hearing must be directly in the hardware. Or maybe you're
>reading the raw audio device at this point.

Hearing is mostly hardware, actually. Lots of hardwired stuff to figure out
from where a sound comes and splitting the soundwaves into seperate
sounds. But interpretation of language is more of software I think, so how
he manages that, I don't know. Maybe he switches from english to binary...

>>So since we are on a roll, I was wondering if I should be rebooting my
>>brain every month?
>
>I usually turn mine off every night, though it's probably true that
>starting it in the morning provides a lot of wear and tear.
>Otherwise, just do an occasional garbage collection to clear the hung
>ptys.

I usually do it by watching TV. Works very good. Especially by watching
programs like 'bikini beach party'. They are so incredibly moronic that
I get a divide by zero on my kernel. Luckily I have trapped this interrupt
so it goes directly to the bootstrap. Otherwise I would continue to stare
at the TV until the alarm clock buzzes.

--
Lennart Regebro, Stacken Computer Club reg...@stacken.kth.se
Any Opinion expressed above is (c) Rent-An-Opinion(tm). It is not an Opinion
of either Lennart Regebro or the Stacken Computer Club.
Now you also can get an Opinion. Call Welcome To Reality(tm) +1 (800) NO-CLUES.

Carl S. Leichter

unread,

Jul 9, 1993, 5:09:13 PM7/9/93

to

In article H...@news.cis.umn.edu, ehh...@staff.tc.umn.edu () writes:
> In article <1993Jul7.1...@news.eng.convex.com> fau...@convex.com (Danny R. Faught) writes:
> >Sorry, we were on a roll and I just couldn't resist.
> >--
> >Danny Faught -- Convex rookie -- MPP OS Test Development
> >"Everything is deeply intertwingled." (Ted Nelson, _Computer Lib_)
>
> Often my brain fails, and my eyes light up. And then I hear a message:
>
> "cerebral cortex dumped"
>
> Then I hear a creaking noise, then memory failure.
>
> So since we are on a roll, I was wondering if I should be rebooting my
> brain every month?
>
> Please, serious replies only.
>
>

I have had this problem too. I have found it is not necessary to go through the fuss and mess of a cold cerebrum reboot. Periodic immersion in a weak solvent (e.g. ethanol) will restore the cortex to an unloaded state.

****************************************************************************
These are just my musings and do not reflect the position of LANL,
the University of California,
the State of California,
the Department of Energy,
the Federal government
or any person living or dead.................... :{
Carl S. Leichter
c...@holmes.lanl.gov
*****************************************************************************

Scott M Sheaf

Zack Evans

Gabriel Velasco

Radford Walker

Cameron Spitzer

Philip Hurley

Danny R. Faught

ehh...@staff.tc.umn.edu

Danny R. Faught

Paul Tomblin

Bruce Ikenaga

Terry Kennedy, Operations Mgr.

Danny R. Faught

Lennart Regebro

Carl S. Leichter

Carl S. Leichter

iv...@cc.usu.edu

M. Edward Borasky

Thayne Forbes

Patrik Dhaeseleer

*-- Sunbird --*

-- Sunbird --