Hi :)
wrt Virtual Memory/pagefile.sys/Swap on Windows the trick seems to be
to set it as a fixed value.
Find
"System Properties" - Advanced tab - Performance (top 3rd) Settings -
Performance Settings - Advanced tab here too - Virtual Memory (bottom
section) Change
There will be about 3 pop-ups open around now.
Use the radio buttons there to change to a "Custom size". This really
needs to be greater than Ram but not more than 2xRam (else it gets
confused and may even reduce performance while tripping over it's own
shoelaces). It has to be greater than Ram because when hibernating
(perhaps sleeping too?) the contents of Ram gets written to Virtual
Memory. But giving it too much just confuses space just confuses
things so just under 2xRam is good but over that might get annoying.
Make sure the same number is in both the top and bottom boxes. Often
there is a recommendation for how much to set it too and it's usually
not a bad idea to follow that advice. I've only seen it give a crazy
suggestion once or twice out of hundreds of machines.
Ok, now it gets a bit fiddly. You have to click on the "Set" button
before clicking on "Ok" otherwise it forgets and you have to re-type
the numbers again. Then you click "Ok" on each of the pop-ups in
turn. Again if you don't it's not harmful, just annoying because it
forgets.
Of course if you have already been using your machine for a while then
Virtual Memory is already quite fragmented so this will only 'stop' it
getting worse. It wont improve things. Also when i say 'stop' it will
continue to suffer normal system rot and there are other factors such
as registry fragmentation that will continue. So, it fixes just 1
problem out of many.
When trying to resurrect an ancient and much used machine i would
initially set Virtual Memory to 0. Then defrag quite a lot and then
plonk a fairly huge file onto the system. Then reset the Virtual
Memory to a respectable size and get rid of the huge file. In theory
i hoped that would force all the Virtual Memory file to be contiguous
and out of the way.
Gnu&Linux does NOT SUFFER from fragmentation until the drive is
something like 96% full, not sure of the exact figure but definitely
over 90% (it's always that extra just 1 episode/movie of Star Trek).
Files might well be fragmented much lower than that despite the
elegant way that files are carefully placed in Ext2,3,4 with plenty of
room all around them to allow them to grow. There is a limit to how
much that policy can really work of course. However even when files
are fragmented there seems to be a better system for tracking where
all the bits are so the read/write head can anticipate and plan ahead
a bit better.
So what i find odd is that despite that Gnu&Linux doesn't use a Swap
file by default! One of the main rules in Gnu&Linux is that for any
'rule' there is always at least 1 version or distro or something that
deliberately breaks that rule but in the case of Swap i haven't found
one yet. They all seem to follow it! They all seem to use a separate
Swap partition or don't use Swap at all.
In Windows, which can't cope with fragmented files and couldn't (until
fairly recently) defrag system files people insist on setting Virtual
Memory to fragment as quickly as possible. Sometimes they set it to
have a fixed lower amount and only vary the top-off but that still
means the file gets read and re-written elsewhere and fragmented.
Normally by default it's set to keep changing size according to how
much of it is needed. That sounds good in theory. When you need more
memory it just expands to fill up more hard-drive space when you need
less it releases some of it. You can get Gnu&Linux to use a swap-file
just the same instead (or as well as) having a separate fixed swap
partition. Unfortunately Windows file-systems such as the various
Fats (vFat, Fat32 etc) and Ntfs are carefully designed to make sure
files fragment quite quickly and end up with bits scattered all over
the place.
Say you have file A that is 20units long and the next file B is 10.
Then you delete A and write a file C that is 30 units. Now you have
20units of C followed by 10 units of B followed by the remaining 10 of
C. If you now delete B and copy A back then you get 20 of C, followed
by 10 of A followed by the 10 remaining of C and then the last 10 of
A. So when you try reading a file the read/write head lurches around
the drive trying to find the various shopped up parts of the file. If
that file is a frequently accessed system file such as Virtual Memory
then it can significantly reduce performance.
In Gnu&Linux it is reckoned that you can significantly increase
performance by putting your system files, particularly your log files,
on a different hard-drive from your data. i mean a proper hard-drive
not just a different partition on the same physical device. The main
reason for putting your data (all in /home) on a separate partition is
not to do with routine performance. it's more about making the system
more robust. it allows you to install a completely new OS without any
risk to your data (but still back-up anyway of course). In theory you
can have several different OSes all using the same /home although that
gets a bit messy if they have the same DE. it works a bit better if
you have 1 KDE one, 1 Gnome(ish), and maybe 1 of any of the rarer ones
(does Unity count as 1 of the rarer ones? i'd say it does but i'm sure
others disagree). Otherwise you find all your different OSes use the
same wallpaper and look the same (big yawn that is) and you don't get
the benefit of the different design teams interesting work.
Something i haven't really tried much, or at least can't remember the
result, is putting all the Virtual Memory on a separate physical
hard-drive. There is an option to split Virtual Memory across several
different hard-drives/partitions some of which might be physically
different drives but i'm not sure whether doing that is good or bad.
Errr, i haven't mentioned Bsd or Apple because i just haven't played
around with them that much. They don't seem to slow down as much as
Windows so i guess they have a similar set-up to Gnu&Linux or have
some neat work-around that might not translate well to Gnu&Linux let
alone Windows.
Regards from
Tom :)
------------------------------------------------------------------------
*From:* Andrew Brown <andrewbr@icon.co.za>
*To:* Tom Davies <tomdavies04@yahoo.co.uk>
*Cc:* Virgil Arrington <cuyfalls@hotmail.com>;
users@global.libreoffice.org
*Sent:* Wednesday, 31 July 2013, 8:48
*Subject:* Re: [libreoffice-users] 4.0.3
Hi Tom
Interesting post. Agree, sometimes these software wars becomes
irksome,
as my late mother and father used to say and raised us with this
motto
"how do you know you don't like it if you have not tried it". This
was
from our young years with foodstuffs that traditionally many young
children don't / have never tried, up to the real things in life.
But I
am in a similiar vein in what MS charge for their O/S and Office
suites
when they are riddled with known and unknown bugs.
At least I have always tried to keep an open mind, and thankfully was
raised on other O/S's (not necessarily desktop/workstation
friendly) and
systems pre-dating MS. I cut my teeth on IBM VAX, Pick, LISP,
FORTRAN,
COBOL, AT&T and SCO Unix, CP/M, BASIC and Xerox GEM, before the
adventure into IBM and MS systems with the very first and crude
DOS, and
then Apple O/S starting some 36 years ago.
I can with experience say I have tried them all, and why my entire
business and home office is OSS and FOSS, even to desktop. I give my
staff the choice of MS or FOSS, thankfully they all eventually
migrate
to FOSS, which allows me to plow the monies recovered from ongoing
and
unnecessary licensing fees into better, faster and more up to date
hardware. Even to the level of my servers.
To end off, the major difference I have between MS software and FOSS,
and you covered briefly in your reply, is that when one discovers
a bug,
or has a problem, one can get a solution or have it fixed promptly
without waiting for a major release or service pack, unlike
proprietory
and closed code. This is the same for malware, it takes so long
for the
commercial software to produce a fix and prevention compared to it
almost being a non-entity in FOSS.
I would be intrigued and grateful, if you could email me
privately, your
tweaks you do for the virtual memory slowdown of it's
fragmentation (by
the way MS refers to it as the pagefile). And that's another
feather in
FOSS's cap, one never has fragmentation or needs to defragment it,
unlike MS. I might know or remember them, but it's not coming to
memory
as I type this.
Regards
On 30/07/2013 03:27 PM, Tom Davies wrote:
> Hi :)
> I think disdain is possibly closer than hatred. I think bioth
are quite far away from the reality though. I think it's simply
that people would rather develop tools that are more robust and
less susceptible to malware and slow-downs.
>
>
> I think once you start using OpenSource tools you begin to
realise that MS seem to have deliberately built-in vulnerabilities
and their slow-downs. FOSS doesn't seem to suffer anything like
as much, although a bit of "system rot" is inevitable in almost
any system.
>
> I'm just installing Win7 on a handfull of machines and am able
to make a couple of tweaks that prevent their "Virtual Memory"
from getting so heavily fragmented. In previous versions of their
OS i have found it significantly reduces the slow-downs if you can
do this early on. On Win7 it takes an extra couple of clicks but
it's still really easy. I always wonder why the default is to set
it to fragment as quickly as possible. It's only with Win7 that
their de-fragger tool can defrag system files such as the Virtual
Memory (err that is Swap to Gnu&Linux geeks lol).
>
> Regards from
> Tom :)
>
>
>
>
>
>
>> ________________________________
>> From: Virgil Arrington <cuyfalls@hotmail.com
<mailto:cuyfalls@hotmail.com>>
>> To: Amit Choudhary <contact.amit.choudhary.india@gmail.com
<mailto:contact.amit.choudhary.india@gmail.com>>;
users@global.libreoffice.org <mailto:users@global.libreoffice.org>
>> Sent: Monday, 29 July 2013, 20:30
>> Subject: Re: [libreoffice-users] 4.0.3
>>
>>
>> I certainly hope the primary motive for FOSS such as LO is not
a disdain for
>> MS. I personally don't care how much money MS makes. I hope the LO
>> developers are motivated by a desire to produce a great product
that can be
>> used worldwide. Hatred usually doesn't provide a very effective
motive for
>> productive action.
>>
>> Virgil
>>
>> -----Original Message-----
>> From: Amit Choudhary
>> Sent: Monday, July 29, 2013 10:47 AM
>> To: users@global.libreoffice.org
<mailto:users@global.libreoffice.org>
>> Subject: Re: [libreoffice-users] 4.0.3
>>
>> On Mon, Jul 29, 2013 at 8:14 PM, Amit Choudhary
>> <contact.amit.choudhary.india@gmail.com
<mailto:contact.amit.choudhary.india@gmail.com>> wrote:
>>>
>>>
>>> On Mon, Jul 29, 2013 at 2:51 PM, Andrew Brown
<andrewbr@icon.co.za <mailto:andrewbr@icon.co.za>> wrote:
>>>> Hi Amit
>>>>
>>>> I understand where you are coming from, and the good news is,
in your
>>>> favour, that MS in both it's O/S and office suite are losing
market share
>>>> in a big way. Here's an article from Ubuntu founder and my
countryman
>>>> Mark Shuttelworth on his take on MS and Ubuntu. I like his
statement that
>>>> the no.1 bug in Linux has now been
> fixed/closed, in that MS no longer
>>>> dominates majority market share.
>>>
>>> But the numbers don't lie. I checked MS revenues and profits on
>>> finance.yahoo.com and it doesn't look like MS is losing market
share. MS
>>> losing share might be an illusion.
>>>
>> Period Ending Jun 30, 2012
>> Jun 30, 2011 Jun 30, 2010
>>
>> Net Income Applicable To Common Shares $16,978,000 $23,150,000
>> $18,760,000 (All numbers in thousands)
>>
>> Regards,
>> Amit
>>
>> --
>> To unsubscribe e-mail to:
users+unsubscribe@global.libreoffice.org
<mailto:unsubscribe@global.libreoffice.org>
>> Problems?
>>
http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
>> Posting guidelines + more:
http://wiki.documentfoundation.org/Netiquette
>> List archive: http://listarchives.libreoffice.org/global/users/
>> All messages sent to this list will be publicly archived and
cannot be
>> deleted
>>
>>
>> --
>> To unsubscribe e-mail to:
users+unsubscribe@global.libreoffice.org
<mailto:unsubscribe@global.libreoffice.org>
>> Problems?
http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
>> Posting guidelines + more:
http://wiki.documentfoundation.org/Netiquette
>> List archive: http://listarchives.libreoffice.org/global/users/
>> All messages sent to this list will be publicly archived and
cannot be deleted
>>
>>
>>
--
To unsubscribe e-mail to: users+unsubscribe@global.libreoffice.org
<mailto:unsubscribe@global.libreoffice.org>
Problems?
http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more:
http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/users/
All messages sent to this list will be publicly archived and
cannot be deleted
Privacy Policy |
Impressum (Legal Info) |
Copyright information: Unless otherwise specified, all text and images
on this website are licensed under the
Creative Commons Attribution-Share Alike 3.0 License.
This does not include the source code of LibreOffice, which is
licensed under the Mozilla Public License (
MPLv2).
"LibreOffice" and "The Document Foundation" are
registered trademarks of their corresponding registered owners or are
in actual use as trademarks in one or more countries. Their respective
logos and icons are also subject to international copyright laws. Use
thereof is explained in our
trademark policy.