Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.hardware > #2454 > unrolled thread

Is it true that it is a bad idea to put a swap/pagefile on a SSD?

Back to article view | Back to comp.os.linux.hardware

Contents

  Is it true that it is a bad idea to put a swap/pagefile on a SSD? Ant <ant@zimage.comANT> - 2014-06-28 11:25 -0700
    Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-06-28 15:11 -0400
      Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? floyd@apaflo.com (Floyd L. Davidson) - 2014-06-28 11:25 -0800
        Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? "Rod Speed" <rod.speed.aaa@gmail.com> - 2014-06-29 05:38 +1000
        Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-06-28 15:55 -0400
          Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? floyd@apaflo.com (Floyd L. Davidson) - 2014-06-28 12:59 -0800
            Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-06-28 18:02 -0400
              Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Larry Blanchard <lblanch@fastmail.fm> - 2014-06-28 23:55 +0000
                Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? floyd@apaflo.com (Floyd L. Davidson) - 2014-06-28 17:20 -0800
                  Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2014-06-28 20:26 -0600
                    Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? floyd@apaflo.com (Floyd L. Davidson) - 2014-06-28 21:17 -0800
              Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? floyd@apaflo.com (Floyd L. Davidson) - 2014-06-28 17:11 -0800
                Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-06-28 23:39 -0400
                  Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? floyd@apaflo.com (Floyd L. Davidson) - 2014-06-28 21:29 -0800
                    Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-06-29 02:21 -0400
          Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Larry Blanchard <lblanch@fastmail.fm> - 2014-06-28 23:52 +0000
            Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? floyd@apaflo.com (Floyd L. Davidson) - 2014-06-28 17:03 -0800
          Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? "David W. Hodgins" <dwhodgins@nomail.afraid.org> - 2014-06-29 15:23 -0400
            Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Ant <ant@zimage.comANT> - 2014-06-29 15:12 -0700
              Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2014-06-29 17:36 -0600
              Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Mike Tomlinson <mike@jasper.org.uk> - 2014-07-01 10:37 +0100
                Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Jolly Roger <jollyroger@pobox.com> - 2014-07-01 16:14 +0000
                Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Ant <ant@zimage.comANT> - 2014-07-02 23:13 -0700
          Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Yousuf Khan <bbbl67@spammenot.yahoo.com> - 2014-07-03 02:13 -0400
    Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Arno <me@privacy.net> - 2014-06-28 22:08 +0000
      Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-06-28 19:42 -0400
        Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Mike Tomlinson <mike@jasper.org.uk> - 2014-07-01 10:41 +0100
          Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Lusotec <nomail@nomail.not> - 2014-07-01 11:31 +0100
    Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Kevin McMurtrie <mcmurtrie@pixelmemory.us> - 2014-06-28 23:31 -0700
      Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? "Rod Speed" <rod.speed.aaa@gmail.com> - 2014-06-30 06:30 +1000
    Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Andreas Rutishauser <andreas@macandreas.ch> - 2014-06-29 08:39 +0100
      Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Ant <ant@zimage.comANT> - 2014-06-29 12:01 -0700
        Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-06-29 16:53 -0400
          Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Ant <ant@zimage.comANT> - 2014-06-29 15:13 -0700
            Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-06-29 21:30 -0400
      Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? hal-usenet@ip-64-139-1-69.sjc.megapath.net (Hal Murray) - 2014-07-01 02:54 -0500
        Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Poutnik <Poutnik@privacy.net> - 2014-07-01 10:39 +0200
        Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? "Rod Speed" <rod.speed.aaa@gmail.com> - 2014-07-02 09:49 +1000
        Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Andreas Rutishauser <andreas@macandreas.ch> - 2014-07-02 08:18 +0100
          Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Ant <ant@zimage.comANT> - 2014-07-02 23:16 -0700
    Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Michael Black <et472@ncf.ca> - 2014-06-29 12:36 -0400
      Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-06-29 14:06 -0400
    Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Yousuf Khan <bbbl67@spammenot.yahoo.com> - 2014-07-03 02:16 -0400
      Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Ant <ant@zimage.comANT> - 2014-07-02 23:19 -0700
      Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Your Name <YourName@YourISP.com> - 2014-07-03 18:55 +1200
      Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? "Charles T. Smith" <cts.private.yahoo@gmail.com> - 2014-07-03 12:46 +0000
        Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? nospam <nospam@nospam.invalid> - 2014-07-03 14:31 -0400
          Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? The Natural Philosopher <tnp@invalid.invalid> - 2014-07-03 19:57 +0100
            Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? "whacker james" <wj852@gmail.com> - 2014-07-04 09:02 +1000
              Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? The Natural Philosopher <tnp@invalid.invalid> - 2014-07-04 00:23 +0100
                Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? Måns Rullgård <mans@mansr.com> - 2014-07-04 00:29 +0100
                  Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? The Natural Philosopher <tnp@invalid.invalid> - 2014-07-04 00:47 +0100
                Re: Is it true that it is a bad idea to put a swap/pagefile on a SSD? "whacker james" <wj852@gmail.com> - 2014-07-04 17:54 +1000

Page 1 of 3  [1] 2 3  Next page →

#2454 — Is it true that it is a bad idea to put a swap/pagefile on a SSD?

Because that wears out the SSD quickly, and it is better to put it on a 
slow HDD. If so, then how do computer (especially laptop/notebook) users 
with only a SSD (not hybrids with HDDs) do this? I assume Windows, Mac 
OS X, and Linux follow this rule?

Currently, I have a Debian stable box that uses an 115 GB Corsair Force 
Series F115 SSD since 11/24/2001. It ran almost 24/7. I do have a very 
old PATA/IDE HDD too, but it is mostly for storage. I'd think it would 
be too slow for virtual memory (VM) if it was on it. However, my Debian 
box (2 GB of RAM) doesn't use it much according to top with a VM (512 MB 
of RAM for an updated Windows XP Pro. SP3):

$ top - 11:13:23 up 23 days,  4:39,  3 users,  load average: 0.21, 0.11, 
0.08
Tasks: 174 total,   1 running, 172 sleeping,   1 stopped,   0 zombie
%Cpu(s):  1.1 us,  1.2 sy,  0.0 ni, 97.7 id,  0.0 wa,  0.0 hi,  0.0 si, 
  0.0 st
KiB Mem:   2060516 total,  1976072 used,    84444 free,   223096 buffers
KiB Swap:  3905532 total,    99548 used,  3805984 free,   439456 cached

   PID USER      PR  NI  VIRT  RES  SHR S  %CPU %MEM    TIME+  COMMAND
  4907 ant       20   0 1477m 655m 593m S   0.3 32.6   9:21.44 VirtualBox
  3231 ant       20   0  326m 135m 4880 S   0.0  6.7   2:06.69 ruby
  4631 ant       20   0 2955m 104m  36m S   1.3  5.2  15:42.50 
plasma-desktop
  4621 ant       20   0 2689m  73m  28m S   1.3  3.6  20:39.79 kwin
10202 ant       20   0  389m  41m  27m S   0.0  2.1   0:05.09 kscreenlocker
  5290 root      20   0  178m  41m  17m S   1.7  2.0 187:13.77 Xorg
  4835 ant       20   0  589m  39m  20m S   0.0  1.9   0:10.53 VirtualBox
  4665 ant       20   0  785m  36m  10m S   0.0  1.8   0:07.64 krunner
  4683 ant       20   0  531m  26m 8876 S   0.0  1.3   0:00.78 kmix
  4601 ant       20   0  354m  15m  13m S   0.0  0.8   0:00.38 kdeinit4
  4628 ant       20   0  619m  14m 6644 S   0.0  0.7   0:01.01 knotify4
  4604 ant       20   0  703m  14m 8100 S   0.0  0.7   0:06.77 kded4
  4619 ant       20   0  531m  11m 6188 S   0.0  0.6   0:00.37 ksmserver
  4708 ant       20   0  283m  10m 6148 S   0.0  0.5   0:00.25 klipper
  4680 ant       20   0  364m  10m 6012 S   0.0  0.5   0:00.16 
polkit-kde-auth
12819 root      20   0 65312  10m 1732 S   0.0  0.5   1:26.60 python
...

Thank you in advance. :)
-- 
"Fall in those single lines like army ants..." --unknown
    /\___/\         Ant(Dude) @ http://antfarm.ma.cx (Personal Web Site)
   / /\ /\ \                Ant's Quality Foraged Links: http://aqfl.net
  | |o   o| |
     \ _ /        If crediting, then use Ant nickname and AQFL URL/link.
      ( )         If e-mailing, then axe ANT from its address if needed.
Ant is currently not listening to any songs on this computer.

[toc] | [next] | [standalone]

#2455

In article <obKdnU708pAllTLOnZ2dnUVZ_rKdnZ2d@earthlink.com>, Ant
<ant@zimage.comANT> wrote:

> Because that wears out the SSD quickly, and it is better to put it on a 
> slow HDD. If so, then how do computer (especially laptop/notebook) users 
> with only a SSD (not hybrids with HDDs) do this? I assume Windows, Mac 
> OS X, and Linux follow this rule?

ssd doesn't wear out that quickly and you'll probably replace the
computer before the ssd wears out, even with swap on it. 

some ssds even have 10 year warranties. good luck finding a hard drive
with a warranty anywhere near that long.

[toc] | [prev] | [next] | [standalone]

#2456

nospam <nospam@nospam.invalid> wrote:
>In article <obKdnU708pAllTLOnZ2dnUVZ_rKdnZ2d@earthlink.com>, Ant
><ant@zimage.comANT> wrote:
>
>> Because that wears out the SSD quickly, and it is better to put it on a
>> slow HDD. If so, then how do computer (especially laptop/notebook) users
>> with only a SSD (not hybrids with HDDs) do this? I assume Windows, Mac
>> OS X, and Linux follow this rule?
>
>ssd doesn't wear out that quickly and you'll probably replace the
>computer before the ssd wears out, even with swap on it.

"Even" with swap?  Swapping is something that should be
relatively rare.  It's there to prevent a major disaster
on the odd occasion when there isn't enough memory.  The
oddness of those occasions should result in minisule
wear on the swap device.

In effect, it is not sound policy to use an expensive
fast device for swap, as the benefit is just exceedingly
small.

>some ssds even have 10 year warranties. good luck finding a hard drive
>with a warranty anywhere near that long.

True, which just adds to reasons for doing the right thing
in the first place.

-- 
Floyd L. Davidson                         http://www.apaflo.com/
Ukpeagvik (Barrow, Alaska)                      floyd@apaflo.com

[toc] | [prev] | [next] | [standalone]

#2457

"Floyd L. Davidson" <floyd@apaflo.com> wrote in message 
news:874mz496r6.fld@barrow.com...
> nospam <nospam@nospam.invalid> wrote:
>>In article <obKdnU708pAllTLOnZ2dnUVZ_rKdnZ2d@earthlink.com>, Ant
>><ant@zimage.comANT> wrote:
>>
>>> Because that wears out the SSD quickly, and it is better to put it on a
>>> slow HDD. If so, then how do computer (especially laptop/notebook) users
>>> with only a SSD (not hybrids with HDDs) do this? I assume Windows, Mac
>>> OS X, and Linux follow this rule?
>>
>>ssd doesn't wear out that quickly and you'll probably replace the
>>computer before the ssd wears out, even with swap on it.

> "Even" with swap?  Swapping is something that should
> be relatively rare.  It's there to prevent a major disaster
> on the odd occasion when there isn't enough memory.

That's not true of plenty of Win systems.

> The oddness of those occasions should result in minisule
> wear on the swap device.

Ditto.

> In effect, it is not sound policy to use an expensive fast
> device for swap, as the benefit is just exceedingly small.

But he was talking about the situation where the is nothing else.

>> some ssds even have 10 year warranties. good luck finding
>> a hard drive with a warranty anywhere near that long.

> True, which just adds to reasons for
> doing the right thing in the first place.

Easier said than done at times, like with his system. 

[toc] | [prev] | [next] | [standalone]

#2458

In article <874mz496r6.fld@barrow.com>, Floyd L. Davidson
<floyd@apaflo.com> wrote:

> >
> >> Because that wears out the SSD quickly, and it is better to put it on a
> >> slow HDD. If so, then how do computer (especially laptop/notebook) users
> >> with only a SSD (not hybrids with HDDs) do this? I assume Windows, Mac
> >> OS X, and Linux follow this rule?
> >
> >ssd doesn't wear out that quickly and you'll probably replace the
> >computer before the ssd wears out, even with swap on it.
> 
> "Even" with swap?  Swapping is something that should be
> relatively rare.

nonsense. swapping happens as a matter of course on any desktop/laptop
system because physical memory is always limited.

> It's there to prevent a major disaster
> on the odd occasion when there isn't enough memory. 

odd occasion?? there is never enough physical memory unless you don't
do much of anything with the computer.

> The
> oddness of those occasions should result in minisule
> wear on the swap device.

an ssd can easily handle swap without any issue. there is no need to
worry about ssd longevity anymore. ssds are *extremely* reliable and
smoking fast and will almost certainly outlast the computer.

> In effect, it is not sound policy to use an expensive
> fast device for swap, as the benefit is just exceedingly
> small.

nonsense.

putting swap on ssd is a *huge* boost to overall system speed. the
difference is *dramatic*. 

and ssds are not that expensive anymore either, plus they're completely
silent and not susceptible to shock or sudden motion (not that that's a
big issue with shock sensors but it's not zero either). 

> >some ssds even have 10 year warranties. good luck finding a hard drive
> >with a warranty anywhere near that long.
> 
> True, which just adds to reasons for doing the right thing
> in the first place.

with rare exception, the right thing is use ssd for the system, apps &
user data, especially with a computer designed *for* ssd.

a hard drive is well suited for data that is less commonly accessed
and/or doesn't need speed and/or the cost would be prohibitive due to
its size, such as a media library. 

however, the key here is that worrying about 'burning out' an ssd due
to excessive use is a myth. it's actually the other way around. a hard
drive will wear out sooner because it has moving parts.

[toc] | [prev] | [next] | [standalone]

#2459

nospam <nospam@nospam.invalid> wrote:
>In article <874mz496r6.fld@barrow.com>, Floyd L. Davidson
><floyd@apaflo.com> wrote:
>
>> >
>> >> Because that wears out the SSD quickly, and it is better to put it on a
>> >> slow HDD. If so, then how do computer (especially laptop/notebook) users
>> >> with only a SSD (not hybrids with HDDs) do this? I assume Windows, Mac
>> >> OS X, and Linux follow this rule?
>> >
>> >ssd doesn't wear out that quickly and you'll probably replace the
>> >computer before the ssd wears out, even with swap on it.
>>
>> "Even" with swap?  Swapping is something that should be
>> relatively rare.
>
>nonsense. swapping happens as a matter of course on any desktop/laptop
>system because physical memory is always limited.

Only with a poorly designed system.

The only time swap is written to on any of my systems is
shortly after a reboot, as essentially unused memory is
swapped out to provide more space in RAM.

>> It's there to prevent a major disaster
>> on the odd occasion when there isn't enough memory.
>
>odd occasion?? there is never enough physical memory unless you don't
>do much of anything with the computer.

There are lots of systems that have enough RAM installed
to simply not need any installed swap space.  That RAM
is typically used as buffering for disk reads, so the
more there is the faster the system appears to be.  But
one effect is that there is virtually always enough RAM
for program execution (generally with many GB more
available at any given moment).

Computers that "don't do much of anything" don't need or
use that type of effective design.  But those with real
loads do.

>> The
>> oddness of those occasions should result in minisule
>> wear on the swap device.
>
>an ssd can easily handle swap without any issue. there is no need to
>worry about ssd longevity anymore. ssds are *extremely* reliable and
>smoking fast and will almost certainly outlast the computer.

That is true enough, but has zero significance.  Putting
money into an SSD to use as swap space is hilarously
lacking in understanding.

>> In effect, it is not sound policy to use an expensive
>> fast device for swap, as the benefit is just exceedingly
>> small.
>
>nonsense.
>
>putting swap on ssd is a *huge* boost to overall system speed. the
>difference is *dramatic*.

Not if the system is properly designed.

>and ssds are not that expensive anymore either, plus they're completely
>silent and not susceptible to shock or sudden motion (not that that's a
>big issue with shock sensors but it's not zero either).

Again, quite true but of no significance.

>> >some ssds even have 10 year warranties. good luck finding a hard drive
>> >with a warranty anywhere near that long.
>>
>> True, which just adds to reasons for doing the right thing
>> in the first place.
>
>with rare exception, the right thing is use ssd for the system, apps &
>user data, especially with a computer designed *for* ssd.

Exactly.  Not for swap space.

>a hard drive is well suited for data that is less commonly accessed
>and/or doesn't need speed and/or the cost would be prohibitive due to
>its size, such as a media library.

"less commonly accessed" describes swap space.

>however, the key here is that worrying about 'burning out' an ssd due
>to excessive use is a myth. it's actually the other way around. a hard
>drive will wear out sooner because it has moving parts.

Which isn't of significance for swap space.

-- 
Floyd L. Davidson                         http://www.apaflo.com/
Ukpeagvik (Barrow, Alaska)                      floyd@apaflo.com

[toc] | [prev] | [next] | [standalone]

#2460

In article <87vbrk7nsm.fld@barrow.com>, Floyd L. Davidson
<floyd@apaflo.com> wrote:

> >> >> Because that wears out the SSD quickly, and it is better to put it on a
> >> >> slow HDD. If so, then how do computer (especially laptop/notebook) users
> >> >> with only a SSD (not hybrids with HDDs) do this? I assume Windows, Mac
> >> >> OS X, and Linux follow this rule?
> >> >
> >> >ssd doesn't wear out that quickly and you'll probably replace the
> >> >computer before the ssd wears out, even with swap on it.
> >>
> >> "Even" with swap?  Swapping is something that should be
> >> relatively rare.
> >
> >nonsense. swapping happens as a matter of course on any desktop/laptop
> >system because physical memory is always limited.
> 
> Only with a poorly designed system.

nonsense.

> The only time swap is written to on any of my systems is
> shortly after a reboot, as essentially unused memory is
> swapped out to provide more space in RAM.

then your system is not a typical system.

> >> It's there to prevent a major disaster
> >> on the odd occasion when there isn't enough memory.
> >
> >odd occasion?? there is never enough physical memory unless you don't
> >do much of anything with the computer.
> 
> There are lots of systems that have enough RAM installed
> to simply not need any installed swap space. 

and there are lots that aren't. so what?

nearly every computer in use today could benefit from more memory but
there's diminishing returns by adding more.

> That RAM
> is typically used as buffering for disk reads, so the
> more there is the faster the system appears to be.  

memory can be used for a lot of things, not just disk caching.

> But
> one effect is that there is virtually always enough RAM
> for program execution (generally with many GB more
> available at any given moment).

nonsense.

*one* app might fit into physical memory, but *all* apps won't fit and
switching between them *will* hit swap.

and 'virtually always enough' are weasel words. 

> Computers that "don't do much of anything" don't need or
> use that type of effective design.  But those with real
> loads do.

what matters is the computers that people actually have and what they
do on them, not how you think it should be.

> >> The
> >> oddness of those occasions should result in minisule
> >> wear on the swap device.
> >
> >an ssd can easily handle swap without any issue. there is no need to
> >worry about ssd longevity anymore. ssds are *extremely* reliable and
> >smoking fast and will almost certainly outlast the computer.
> 
> That is true enough, but has zero significance. 

it has every significance since the original question was whether
replacing a hard drive with an ssd would 'wear out' due to swap, which
will be on the ssd if the drive is replaced. and the answer is it
won't.

> Putting
> money into an SSD to use as swap space is hilarously
> lacking in understanding.

nobody said anything about putting money into ssd *just* for swap.
where in the world did you come up with that ??

> >> In effect, it is not sound policy to use an expensive
> >> fast device for swap, as the benefit is just exceedingly
> >> small.
> >
> >nonsense.
> >
> >putting swap on ssd is a *huge* boost to overall system speed. the
> >difference is *dramatic*.
> 
> Not if the system is properly designed.

wrong. ssd is one of the easiest ways to drastically improve
performance, especially if memory is maxed out which it often is,
unless the computer is older and a sata 3 drive saturates it.

> >and ssds are not that expensive anymore either, plus they're completely
> >silent and not susceptible to shock or sudden motion (not that that's a
> >big issue with shock sensors but it's not zero either).
> 
> Again, quite true but of no significance.

it's significant in that it increases reliability as well as the
overall user experience in using the computer.

> >> >some ssds even have 10 year warranties. good luck finding a hard drive
> >> >with a warranty anywhere near that long.
> >>
> >> True, which just adds to reasons for doing the right thing
> >> in the first place.
> >
> >with rare exception, the right thing is use ssd for the system, apps &
> >user data, especially with a computer designed *for* ssd.
> 
> Exactly.  Not for swap space.

swap is included because it's on the same drive in almost every case. 

> >a hard drive is well suited for data that is less commonly accessed
> >and/or doesn't need speed and/or the cost would be prohibitive due to
> >its size, such as a media library.
> 
> "less commonly accessed" describes swap space.

not even remotely close to true.

> >however, the key here is that worrying about 'burning out' an ssd due
> >to excessive use is a myth. it's actually the other way around. a hard
> >drive will wear out sooner because it has moving parts.
> 
> Which isn't of significance for swap space.

of course it is, because when a hard drive fails, it takes out
everything on it, including swap.

once again, the original question was whether ssds would wear out
because swap is on them and the answer is no, even if swap is hit hard.

[toc] | [prev] | [next] | [standalone]

#2464

On Sat, 28 Jun 2014 18:02:44 -0400, nospam wrote:

> what matters is the computers that people actually have and what they do
> on them, not how you think it should be.

pot vs kettle ????

[toc] | [prev] | [next] | [standalone]

#2467

Larry Blanchard <lblanch@fastmail.fm> wrote:
>On Sat, 28 Jun 2014 18:02:44 -0400, nospam wrote:
>
>> what matters is the computers that people actually have and what they do
>> on them, not how you think it should be.
>
>pot vs kettle ????

Not really, because nospam is just plain wrong.

He responded as above to this statement:

  "Computers that "don't do much of anything" don't need or
   use that type of effective design.  But those with real
   loads do."

There is no pot or kettle in that.  You and I posted a pair of
examples that shows exactly how opposite from reality his
concept is.  In fact there are many large busy computer systems
that are purposely configured to not require ever using swap
space.  The reason is simple:  running out of swap space is
unconscionably slow, and to be avoided.

-- 
Floyd L. Davidson                         http://www.apaflo.com/
Ukpeagvik (Barrow, Alaska)                      floyd@apaflo.com

[toc] | [prev] | [next] | [standalone]

#2468

floyd@apaflo.com (Floyd L. Davidson) writes:

> Larry Blanchard <lblanch@fastmail.fm> wrote:
>>On Sat, 28 Jun 2014 18:02:44 -0400, nospam wrote:
>>
>>> what matters is the computers that people actually have and what they do
>>> on them, not how you think it should be.
>>
>>pot vs kettle ????
>
> Not really, because nospam is just plain wrong.
>
> He responded as above to this statement:
>
>   "Computers that "don't do much of anything" don't need or
>    use that type of effective design.  But those with real
>    loads do."
>
> There is no pot or kettle in that.  You and I posted a pair of
> examples that shows exactly how opposite from reality his
> concept is.  In fact there are many large busy computer systems
> that are purposely configured to not require ever using swap
> space.  The reason is simple:  running out of swap space is
> unconscionably slow, and to be avoided.

From the posting style, is it another Rod Speed sockpuppet?  That would
explain everything.

[toc] | [prev] | [next] | [standalone]

#2470

Joe Pfeiffer <pfeiffer@cs.nmsu.edu> wrote:
>floyd@apaflo.com (Floyd L. Davidson) writes:
>
>> Larry Blanchard <lblanch@fastmail.fm> wrote:
>>>On Sat, 28 Jun 2014 18:02:44 -0400, nospam wrote:
>>>
>>>> what matters is the computers that people actually have and what they do
>>>> on them, not how you think it should be.
>>>
>>>pot vs kettle ????
>>
>> Not really, because nospam is just plain wrong.
>>
>> He responded as above to this statement:
>>
>>   "Computers that "don't do much of anything" don't need or
>>    use that type of effective design.  But those with real
>>    loads do."
>>
>> There is no pot or kettle in that.  You and I posted a pair of
>> examples that shows exactly how opposite from reality his
>> concept is.  In fact there are many large busy computer systems
>> that are purposely configured to not require ever using swap
>> space.  The reason is simple:  running out of swap space is
>> unconscionably slow, and to be avoided.
>
>From the posting style, is it another Rod Speed sockpuppet?  That would
>explain everything.

They certainly appear to be peas from the same pod!

-- 
Floyd L. Davidson                         http://www.apaflo.com/
Ukpeagvik (Barrow, Alaska)                      floyd@apaflo.com

[toc] | [prev] | [next] | [standalone]

#2466

nospam <nospam@nospam.invalid> wrote:
>In article <87vbrk7nsm.fld@barrow.com>, Floyd L. Davidson
><floyd@apaflo.com> wrote:
>
>> >> >> Because that wears out the SSD quickly, and it is better to put it on a
>> >> >> slow HDD. If so, then how do computer (especially laptop/notebook) users
>> >> >> with only a SSD (not hybrids with HDDs) do this? I assume Windows, Mac
>> >> >> OS X, and Linux follow this rule?
>> >> >
>> >> >ssd doesn't wear out that quickly and you'll probably replace the
>> >> >computer before the ssd wears out, even with swap on it.
>> >>
>> >> "Even" with swap?  Swapping is something that should be
>> >> relatively rare.
>> >
>> >nonsense. swapping happens as a matter of course on any desktop/laptop
>> >system because physical memory is always limited.
>>
>> Only with a poorly designed system.
>
>nonsense.

More of *your* endless nonsense on a topic you know little
about.

>> The only time swap is written to on any of my systems is
>> shortly after a reboot, as essentially unused memory is
>> swapped out to provide more space in RAM.
>
>then your system is not a typical system.
>
>> >> It's there to prevent a major disaster
>> >> on the odd occasion when there isn't enough memory.
>> >
>> >odd occasion?? there is never enough physical memory unless you don't
>> >do much of anything with the computer.
>>
>> There are lots of systems that have enough RAM installed
>> to simply not need any installed swap space.
>
>and there are lots that aren't. so what?

You said it "never" happens.  Clearly it happens very often!

>nearly every computer in use today could benefit from more memory but
>there's diminishing returns by adding more.

The diminishing return is not the amount of RAM used for program
execution though.  Which is to say, swap space commonly has
nothing to do with how much RAM is provisioned, nor the other
way around.

>> That RAM
>> is typically used as buffering for disk reads, so the
>> more there is the faster the system appears to be.
>
>memory can be used for a lot of things, not just disk caching.

So?

>> But
>> one effect is that there is virtually always enough RAM
>> for program execution (generally with many GB more
>> available at any given moment).
>
>nonsense.
>
>*one* app might fit into physical memory, but *all* apps won't fit and
>switching between them *will* hit swap.
>
>and 'virtually always enough' are weasel words.

I've got a box here that typically shows a load average well into
double digits when I'm editing images.  It has never swapped out
a running program.

[remaining garbage snipped]

-- 
Floyd L. Davidson                         http://www.apaflo.com/
Ukpeagvik (Barrow, Alaska)                      floyd@apaflo.com

[toc] | [prev] | [next] | [standalone]

#2469

In article <87fvio7c61.fld@barrow.com>, Floyd L. Davidson
<floyd@apaflo.com> wrote:

> >> >> >> Because that wears out the SSD quickly, and it is better to put it on a
> >> >> >> slow HDD. If so, then how do computer (especially laptop/notebook)
> >> >> >> users with only a SSD (not hybrids with HDDs) do this? I assume Windows,
> >> >> >> Mac OS X, and Linux follow this rule?
> >> >> >
> >> >> >ssd doesn't wear out that quickly and you'll probably replace the
> >> >> >computer before the ssd wears out, even with swap on it.
> >> >>
> >> >> "Even" with swap?  Swapping is something that should be
> >> >> relatively rare.
> >> >
> >> >nonsense. swapping happens as a matter of course on any desktop/laptop
> >> >system because physical memory is always limited.
> >>
> >> Only with a poorly designed system.
> >
> >nonsense.
> 
> More of *your* endless nonsense on a topic you know little
> about.

quite the opposite. you say that swap is rare, yet it's very definitely
*not* rare. you say that ssd won't make a difference yet it makes a
*huge* difference in performance, partly due to swap being on it (along
with everything else).

once again, you're wrong, and as usual, you'll argue 'til the cows come
home rather than admit you're wrong.

> >> The only time swap is written to on any of my systems is
> >> shortly after a reboot, as essentially unused memory is
> >> swapped out to provide more space in RAM.
> >
> >then your system is not a typical system.
> >
> >> >> It's there to prevent a major disaster
> >> >> on the odd occasion when there isn't enough memory.
> >> >
> >> >odd occasion?? there is never enough physical memory unless you don't
> >> >do much of anything with the computer.
> >>
> >> There are lots of systems that have enough RAM installed
> >> to simply not need any installed swap space.
> >
> >and there are lots that aren't. so what?
> 
> You said it "never" happens.  Clearly it happens very often!

wrong on that too. what i said was that it never happens unless you
don't do much of anything with the computer. 

on a typical system, it's *very* normal to swap and i'm not talking
about edge cases where someone is just checking email or whatever.

> >nearly every computer in use today could benefit from more memory but
> >there's diminishing returns by adding more.
> 
> The diminishing return is not the amount of RAM used for program
> execution though.  Which is to say, swap space commonly has
> nothing to do with how much RAM is provisioned, nor the other
> way around.

the diminishing return is how much physical memory to have in the
system. at some point, it's just not worth the additional cost. 

as for the amount of swap space, it's dependent on how much ram is in
the system and how many apps the user wants to run at the same time.
that's common sense.

> >> That RAM
> >> is typically used as buffering for disk reads, so the
> >> more there is the faster the system appears to be.
> >
> >memory can be used for a lot of things, not just disk caching.
> 
> So?

so it's more than just what you say it is.

> >> But
> >> one effect is that there is virtually always enough RAM
> >> for program execution (generally with many GB more
> >> available at any given moment).
> >
> >nonsense.
> >
> >*one* app might fit into physical memory, but *all* apps won't fit and
> >switching between them *will* hit swap.
> >
> >and 'virtually always enough' are weasel words.
> 
> I've got a box here that typically shows a load average well into
> double digits when I'm editing images.  It has never swapped out
> a running program.

how much memory and how many apps are you running? without specifics,
your statement is meaningless.

here in the real world, where people typically have 4-8 gig systems,
they *will* swap in normal use, and that's even without memory
intensive apps. for those who use memory intensive apps, 16 gig (or
more) is wise.

[toc] | [prev] | [next] | [standalone]

#2471

nospam <nospam@nospam.invalid> wrote:
>In article <87fvio7c61.fld@barrow.com>, Floyd L. Davidson
><floyd@apaflo.com> wrote:
>> I've got a box here that typically shows a load average well into
>> double digits when I'm editing images.  It has never swapped out
>> a running program.
>
>how much memory and how many apps are you running? without specifics,
>your statement is meaningless.

And hence my statement had meaning, while your reply doesn't... :-)

Look up "load average" to begin with...

-- 
Floyd L. Davidson                         http://www.apaflo.com/
Ukpeagvik (Barrow, Alaska)                      floyd@apaflo.com

[toc] | [prev] | [next] | [standalone]

#2472

In article <87y4wg5lmd.fld@barrow.com>, Floyd L. Davidson
<floyd@apaflo.com> wrote:

> >> I've got a box here that typically shows a load average well into
> >> double digits when I'm editing images.  It has never swapped out
> >> a running program.
> >
> >how much memory and how many apps are you running? without specifics,
> >your statement is meaningless.
> 
> And hence my statement had meaning, while your reply doesn't... :-)
> 
> Look up "load average" to begin with...

you should start by looking up virtual memory and how it works in a
modern computer. 

they definitely swap, and often.

[toc] | [prev] | [next] | [standalone]

#2463

On Sat, 28 Jun 2014 15:55:14 -0400, nospam wrote:

>> "Even" with swap?  Swapping is something that should be relatively
>> rare.
> 
> nonsense. swapping happens as a matter of course on any desktop/laptop
> system because physical memory is always limited.
> 
>> It's there to prevent a major disaster on the odd occasion when there
>> isn't enough memory.
> 
> odd occasion?? there is never enough physical memory unless you don't do
> much of anything with the computer.

I usually show little or no swap used.  I have 2G of memory and 5 
workspaces.  If I download a particularly overbusy web page I might show 
a little, or if I'm running a fractal just for fun.

[toc] | [prev] | [next] | [standalone]

#2465

Larry Blanchard <lblanch@fastmail.fm> wrote:
>On Sat, 28 Jun 2014 15:55:14 -0400, nospam wrote:
>
>>> "Even" with swap?  Swapping is something that should be relatively
>>> rare.
>>
>> nonsense. swapping happens as a matter of course on any desktop/laptop
>> system because physical memory is always limited.
>>
>>> It's there to prevent a major disaster on the odd occasion when there
>>> isn't enough memory.
>>
>> odd occasion?? there is never enough physical memory unless you don't do
>> much of anything with the computer.
>
>I usually show little or no swap used.  I have 2G of memory and 5
>workspaces.  If I download a particularly overbusy web page I might show
>a little, or if I'm running a fractal just for fun.

Compare that with the server I use for image editing.
It runs a 64 bit Linux, with 8 CPUs, has 32GB of RAM,
and commonly the load average can be well into double
digits and I've seen it go to 3 digits.

It has never swapped out a running program.  It shows
just under 100k of swap space being used, all of which
is code used to boot the system and then forever more is
idle.

At the moment it shows 20 GB of RAM used for disk
buffers.  Which is very nice when all 8 CPUs are
chugging away full blast at disk i/o while processing
images that are more than 100 MB in size.

This is not an uncommon configuration for this type of
work.

nospam has it backwards!  The little personal machines
are the ones that don't do much, and they are likely to
have limited RAM and may actually see swapping.  It is
larger systems with seriously large work loads that are
configured not to need swap space, because it is just
too expensive in terms of wasted CPU time to let a
system swap.  And really, an 8 CPU system with only 32
GB of RAM is not exactly a large system.

-- 
Floyd L. Davidson                         http://www.apaflo.com/
Ukpeagvik (Barrow, Alaska)                      floyd@apaflo.com

[toc] | [prev] | [next] | [standalone]

#2481

On Sat, 28 Jun 2014 15:55:14 -0400, nospam <nospam@nospam.invalid> wrote:

> nonsense. swapping happens as a matter of course on any desktop/laptop
> system because physical memory is always limited.

With kde running on the host, and 5 vb sessions running (4 linux with kde,
and one Windows 7 session running ...

$ free -m
              total       used       free     shared    buffers     cached
Mem:         15996      15589        406          0        172       3399
-/+ buffers/cache:      12017       3979
Swap:        32767          0      32767

See http://rudd-o.com/linux-and-free-software/tales-from-responsivenessland-why-linux-feels-slow-and-how-to-fix-that

Regards, Dave Hodgins

-- 
Change nomail.afraid.org to ody.ca to reply by email.
(nomail.afraid.org has been set up specifically for
use in usenet. Feel free to use it yourself.)

[toc] | [prev] | [next] | [standalone]

#2484

On 6/29/2014 12:23 PM PT, David W. Hodgins typed:

>> nonsense. swapping happens as a matter of course on any desktop/laptop
>> system because physical memory is always limited.
>
> With kde running on the host, and 5 vb sessions running (4 linux with kde,
> and one Windows 7 session running ...
>
> $ free -m
 > ...

Here's mine so far:
$ free -m
              total       used       free     shared    buffers     cached
Mem:          2012       1382        629          0        184        550
-/+ buffers/cache:        648       1364
Swap:         3813        107       3706

Also, VirtualBox was exited since yesterday afternoon PDT. Its swap has 
grown overnight.
-- 
"I don't understand them anymore, these people that travel the 
commuter-trains to their dormitory towns. These people call themselves 
human, but, by a pressure they do not feel, are forced to do their work 
like ants. With what do they fill their time when they are free of work 
on their silly little Sundays?" --Antoine de saint-Exupéry, 'Wind, Sand, 
and Stars,' 1939
    /\___/\         Ant(Dude) @ http://antfarm.ma.cx (Personal Web Site)
   / /\ /\ \                Ant's Quality Foraged Links: http://aqfl.net
  | |o   o| |
     \ _ /        If crediting, then use Ant nickname and AQFL URL/link.
      ( )         If e-mailing, then axe ANT from its address if needed.
A song (i/wa)s playing on this computer: Poom - Les voiles (ripped from 
YouTube -- 6-15)

[toc] | [prev] | [next] | [standalone]

#2486

Mine:

snowball:528$ free -m
             total       used       free     shared    buffers     cached
Mem:         32216      31499        716         28       2368      22881
-/+ buffers/cache:       6250      25965
Swap:        65535          0      65535

[toc] | [prev] | [next] | [standalone]

Page 1 of 3  [1] 2 3  Next page →

Back to top | Article view | comp.os.linux.hardware

csiph-web