Re: 4GB RAM
- From: "Just D." <no@xxxxxxxxxxx>
- Date: Mon, 24 Aug 2009 23:06:51 -0700
Hi Richard,
Mine is currently 224MB with a light load of 27 processes. (I'm writing
this
offline.) I have 1GB RAM, with a 3Ghz Pentium 4 CPU. (WinXP-pro-SP3.) Out
of
the box, the pagefile was set to Minimum 1524MB, and Maximum 3048MB. The
highest that it has ever gone is about 343MB. The minimum was apparently
It always depends on what you do with this machine. I'm a pro developer
having VS, SQL Servers, Web Servers, FTP Servers, MS Office, and many other
apps running at the same time. So when I was using 1GB of RAM it was toooo
slowly. 2GB is more or less acceptable, but 4 is much better for sure.
I was playing with that many years ago. If I disable swap having enough
memory at the same time some apps simply refuse to start. For example
Adobe Photoshop Elements can't work without virtual memory. That's studid
because even if one has a lot of memory he's stuck here.
First, disabling the paging file does not eliminate virtual memory, since
there will still be a hidden (minimum) 2MB file on the system drive.
Are you sure? Do you want me to send you a screen shot with both root dirs
C: and D: with no pagefile? I was surprised when I read that from you and
decided to test, maybe I skipped something important from the recent
updates, but no, I'm right, there is no any swap file now.
disk-based part. I assume that graphics programs need a lot more virtual
memory than is obvious because of the need to store complete information
to "Undo" editing changes and such like things. (I undo a lot! :)
Well, the graphics program usually creates its own swap file. Example -
Adobe Photoshop Elements or Photoshop, doesn't matter. You have to set the
drives to be used as the first swap drive, the second one... And it creates
its own temp files right on the root. It has nothing to the regular system
swap file, but if you disable it you can easily get a problem with Adobe
products. Maybe it's fixed now, I tried that last time many years ago. Well,
just tested again and got the message - "Could not start Adobe Photoshop
Elements because the volume Windows is using for the Virtual Memory doesn't
have enough space, which could lead..." The reason is that the main swap
file was disabled, but both drives have 20GB and about 350GB of free space
accordingly. :) Enough?
The "speed" may be an illusion. Yes, with 64KB clusters there would be
fewer clusters for both Windows and the MFT (Master File Table) to keep
track
of, and therefore electronic transfer of information would be quicker.
(And
less possibility of fragmentation with large clusters.) However, the drive
still has to read or write to the individual 512-byte sectors, with a
fixed RPM
rotation rate and fixed seek time for the mechanical read/write arm.
Well, I'm just talking that the whole system will copy large file much
faster. I usually edit huge video files, this is some kind of hobby. One DV
tape is about 14-15GB file. Have you ever tried to copy these files from
disk to disk? Try that with the different cluster size, you will be
impressed. After you copied a few tenths or greater GB size file it takes a
very long time to close the file on disk after the process is finished if
the cluster size is 4KB - default value. It works like a sharm if the
cluster size on the same disk is 64KB and it taqkes just a few seconds to
close the file. The same on 4KB clusters - from 20 seconds and longer. Trust
me, this effect is visible not only on the disk system. When you want to
copy something over the LAN you can find the speed too slow. Try to google
for "Jumbo Frame". You will probably find what the default block size is
when you transfer the file over the LAN and how much efforts it takes from
the network adapter and CPU to fragment this huge amount of data and then
reassemble it on the remote size. It's not only theory, it's very serious to
be true.
Depending where the page file partition is physically located, relative to
the location of the primary system and program files, you may end up with
slower performance while multi-tasking with a lot of page swaps swinging
the head between two widely separated locations in partitions on the same
physical drive. A pagefile on a separate physical drive that was not in
competition with action on the system drive could be faster, (if it is not
on an older slower drive. :)
This is from a very old FAQ, I think over 20 years ago, and this is true. It
was especially true on the very old MFM drives if you have any idea what
that is. It's just a question of how you tune your computer up, there are
many tricks to do that. Btw, according to one of your phrases you've never
seem a fragmented swap file. Take a closer look and you will find it. I know
that because it was always some kind of game to get this file as
unfragmented as possible. I think if you run:
defrag.exe c: -a -f
on several different machines it will show the swap file fragmentation.
Another wrinkle is the way Windows sets up the location of the pagefile on
the volume in the first place, placing it widely separated from the
initial system files. The outermost tracks have the highest number of
sectors per
Do you want me to send you a screen shot with the location of my swap file
when I use it? :) And realistically the files that the system uses more
often should be located as close as possible, say, in the middle of the disk
or in the first third, depending on the disk occupation. :) In this case the
seek delay will be the shortest. Some speedisk apps in the history were able
to get the statistics of the file usage using the last access attribute of
the file moving it to the fastest accessible area. I'm sure you're not aware
of that.
track and therefore the fastest access times, and Windows XP is better
than previous versions because it will move frequently used files closer
to the
beginning of the drive to improve access speed, (and it uses prefetch to
Yea, and then make the heads fly between the beginning of the disk and
working areas with the user files. :)
I think it's incorrect, partially or at all. It was a very ancient idea to
start using disk from the very beginning filling the first fragments that
have been met. I guess it ended with MS DOS 3.31 or so over 20 years ago.
Even MS DOS 6.21 was allowing a terrible fragmentation and it was writing
files using absolutely different schema. A little later different disk cache
systems appeared and some of them even started writing the files trying to
move the heads as less as possible and basically keeping the heads in the
middle of the disk to shorten the access speed.
speed things up also,) but it does not position the pagefile all that
close to the beginning of the drive near the fastest outermost track.
Dreams... Open your eyes. There are apps letting you take a look at the
drive map. For example when I enabled the swap file in my system when the
system was installed, it appeared after all system and other files, apprx.
from 75% to 80% of the first partition, but it's just rough evaluation. The
physical drive had 2 partitions - 80 and the rest.
Another problem I noticed with another computer here, (WinXP-home-SP3,)
with an 83.2GB drive, with a hidden recovery partition, is that the 2.01GB
recovery partition is the first on the drive, so the beginning of the main
When one creates a partition, doesn't matter working or recovery one, he is
able to put it to any place. Btw, Acronis was taking the very last cylinder
of the disk for its own hidden recovery partition up to version 11, then in
version 2009 they started using the system (NTFS) partition, which is
nonsense. I wrote them what I'm thinking about this stupid solution, but
this is off topic here. I can just add that they switched from Linux based
Recovery Partition to Windows based.
Another interesting thing with that computer was after installing Avast!
antivirus, I pressed alt+ctrl+del to activate Task Manager with the
Performance tab with the CPU usage and PF usage graphs, and then activated
the full antivirus scan. The CPU usage immediately max'd out at 100% for
more than a minute with occasional downward spikes a few percent less. The
It was testing real processes in memory, that's why.
most amazing thing was the PF usage graph, which is ordinarily flat-lined
near the bottom, it became frantic and started jumping up and down like
wild. (Peak 532MB) I assume Avast! was examining every little process in
It was testing the files loading them one by one. :) If you enabled the deep
recursive search inside archives it could take this amount of memory.
memory, aggressively swapping pages to see if any viruses or other malware
were active. ("There's a new Sheriff in town!" :)
Really? Does it really need to swap processes to test them? :):):) There is
something really new. Actually if you know the entry point of the process...
The most important thing to remember is that, although the word "disk" in
the term Hard Disk Drive is singular, most desktop computer drives
actually
have from 1 to 4 platters inside with 2 heads per platter. That previously
mentioned 83.2GB drive has 1 platter with 2 heads, so there is a 2nd
outermost track area on the 2nd side of the platter where a small pagefile
partition could be set up on the fast track - simply position the
partition
I guess you never heard about interleave, about block of heads and how they
work, about different zones on the hard drives having different number of
sectors and in what year that appeared..., etc.
Keep in mind also that with NTFS format, the MFT is the first file put on
the partition, and there is a 12.5% space reserved for the MFT Zone on the
partition, so allow maybe an extra 20% more than the maximum pagefile size
for the partition size. The computer I'm currently using to post this
message has a 250GB drive with 2 platters and 4 heads, so it has 4
surfaces with outermost fast tracks. The drives in both computers are 3.5
inch form
And you're sure that all these surfaces and heads are working independently?
:)
the innermost track. That means disk read speeds are generally more than
twice as fast on the outer tracks than the inner ones.
Depends on the disk generation, controller, interface, the number of
defective sectors, remapping, the country origin, the hard drive quality and
couple thousand more factors. :) I hope you know what the remapping is?
I'm guessing that your laptop probably has a low profile 2.5 inch HDD, so
it would have a shorter distance for the read/write head to move between
the
innermost and outermost tracks. Using the 1.5x rule, with 3.24GB RAM, you
would need a minimum (initial) pagefile of 4.86GB, in at least a 6GB
partition. You should be able to set the maximum the same as the minimum
initial size. How big is your internal drive anyway?
Not 80 like your, it's half-a-TB. There are bigger ones, but they are 12mm
height and the current standard supports only 8.5-9 mm 2.5" drives.
BTW: You appear to be using Outlook Express with Avast! antivirus checking
your outgoing messages. Outlook Express sometimes malfunctions when a/v
scans messages. You might want to drop over to the OutlookExpress
Well, it scans mostly before they arrive, it works like an additional layer,
but when we scan the whole PC the mail base files will be scanned as well.
newsgroup and learn more about OE quirks, and check out these other links
here:
news://msnews.microsoft.com/microsoft.public.outlookexpress.general
http://www.oehelp.com/OETips.aspx
The smartest idea written at the link above is to completely disable
antivirus protection for email messages. :):):) Bravo!
You can definitely destroy the Inbox folder, but I never keep anything here,
that's first. Second - Avast works correctly comparing to Kaspersky's AVP.
It can destroy you the Inbox folder if you use an older version of this
software. I know these cases.
Just wondering - are you keeping all your messages in one folder? You're not
using any rules to automatically move them to different folders, right?
Why you don't need your anti-virus to scan your email
http://thundercloud.net/infoave/tutorials/email-scanning/index.htm
Yes-yes. But I'm not a boy to believe in everything I read on the Internet.
(Be happy! Be VERY happy! :)
--Richard
Well, I'm really happy, :)
Just D.
.
- References:
- 4GB RAM
- From: Just D.
- Re: 4GB RAM
- From: databaseben
- Re: 4GB RAM
- From: Just D.
- Re: 4GB RAM
- From: Richard
- 4GB RAM
- Prev by Date: Re: Device not recognized when mp3 player connected to USB port
- Next by Date: Re: Can't delete a folder and its contained file
- Previous by thread: Re: 4GB RAM
- Next by thread: SHRegGetValueW in .dll SHLWAPI.dll could not be located
- Index(es):
Relevant Pages
|
