Re: Swap files

"Stacey" <fotocord@yahoo.com> wrote in message
news:bd5tqb$p5bn9$2@ID-52908.news.dfncis.de...
> Neogenesis wrote:
>
> > I have Win ME and Win 2000 server on the same system. i have 3 hard
drives
> > and wondering if i was using windows 2000, where did i place the swap
> > files?
> >
> > Does it matter if i place it on a Fat partition or a NTFS partition?
> >
> > Window ME FAT HD 1 UDMA 100 Primary
> > Personal Files FAT HD 2 UDMA 66 Slave
> > Windows 2000 NTFS HD 3 UDMA 66 Primary
> >
> >
> > I know it's better to place the swamp files on another which is not
> > running the OS, so i guess HD 3.
> > If NTFS is better then i could re partition HD 2 half FAT and half NTFS.
> >
> >
>
>
> You want the swap on the fastest drive. UDMA 100 is faster than UDMA66 and
> fat32 is faster than NTFS. It's not better to put it on another drive than
> the OS if that drive is slower. I'll let you figure it out from there.
>
> --
>
> Stacey

how about spreading out the swamp file onto 2 HD? would 2 UDMA 66, FAT be
better or worse?

"Neogenesis" <hangwing@hotmail.com> wrote in message
news:1hCJa.9776$xF.590480@newsfep2-win.server.ntli.net...
>
> "Stacey" <fotocord@yahoo.com> wrote in message
> news:bd5tqb$p5bn9$2@ID-52908.news.dfncis.de...
> > Neogenesis wrote:
> >
> > > I have Win ME and Win 2000 server on the same system. i have 3 hard
> drives
> > > and wondering if i was using windows 2000, where did i place the swap
> > > files?
> > >
> > > Does it matter if i place it on a Fat partition or a NTFS partition?
> > >
> > > Window ME FAT HD 1 UDMA 100 Primary
> > > Personal Files FAT HD 2 UDMA 66 Slave
> > > Windows 2000 NTFS HD 3 UDMA 66 Primary
> > >
> > >
> > > I know it's better to place the swamp files on another which is not
> > > running the OS, so i guess HD 3.
> > > If NTFS is better then i could re partition HD 2 half FAT and half
NTFS.
> > >
> > >
> >
> >
> > You want the swap on the fastest drive. UDMA 100 is faster than UDMA66
and
> > fat32 is faster than NTFS. It's not better to put it on another drive
than
> > the OS if that drive is slower. I'll let you figure it out from there.
> >
> > --
> >
> > Stacey
>
> how about spreading out the swamp file onto 2 HD? would 2 UDMA 66, FAT be
> better or worse?
>
>

Swamp file?

I hate to mention typos, but that one deserves a place in the dictionary.
--
I believe in having an open mind,
but not so open that my brains fall out.

On Mon, 23 Jun 2003 13:28:30 +0100, "Neogenesis"
<hangwing@hotmail.com> wrote:

>
>"Neogenesis" <hangwing@hotmail.com> wrote in message
>news:1hCJa.9776$xF.590480@newsfep2-win.server.ntli.net...
>>
>> "Stacey" <fotocord@yahoo.com> wrote in message
>> news:bd5tqb$p5bn9$2@ID-52908.news.dfncis.de...
>> > Neogenesis wrote:
>> >
>> > > I have Win ME and Win 2000 server on the same system. i have 3 hard
>> drives
>> > > and wondering if i was using windows 2000, where did i place the swap
>> > > files?
>> > >
>> > > Does it matter if i place it on a Fat partition or a NTFS partition?
>> > >
>> > > Window ME FAT HD 1 UDMA 100 Primary
>> > > Personal Files FAT HD 2 UDMA 66 Slave
>> > > Windows 2000 NTFS HD 3 UDMA 66 Primary
>> > >
>> > >
>> > > I know it's better to place the swamp files on another which is not
>> > > running the OS, so i guess HD 3.
>> > > If NTFS is better then i could re partition HD 2 half FAT and half
>NTFS.
>> > >
>> > >
>> >
>> >
>> > You want the swap on the fastest drive. UDMA 100 is faster than UDMA66
>and
>> > fat32 is faster than NTFS. It's not better to put it on another drive
>than
>> > the OS if that drive is slower. I'll let you figure it out from there.
>> >
>> > --
>> >
>> > Stacey
>>
>> how about spreading out the swamp file onto 2 HD? would 2 UDMA 66, FAT be
>> better or worse?
>>
>
>And how about mixing 2 different UDMA speed? does this increase, decrease or
>no change?
>

No change so long as the HDDs have the same performance otherwise, but
these days with memory so cheap, the swap file configuration shouldn't
be of much concern, just add enough memory that it's never used.


Dave

Neogenesis wrote:

>>
>>
>> You want the swap on the fastest drive. UDMA 100 is faster than UDMA66
>> and fat32 is faster than NTFS. It's not better to put it on another drive
>> than
>> the OS if that drive is slower. I'll let you figure it out from there.
>>
>> --
>>
>> Stacey
>
> how about spreading out the swamp file onto 2 HD? would 2 UDMA 66, FAT be
> better or worse?

Nope bad idea. Best option is outer edge of UDMA100 fat32 drive.

--

Stacey

"-" <nospam@nospam.com> wrote in message
news:3ef80cc5$0$12088$afc38c87@news.easynet.co.uk...
> How could you specify that you write the swap to the edge of the disk
> exactly?

You can do that by partitioning the drive. I can't remember if drives write
to the inner or outer tracks first (a google search should be able to tell
you). So you take advantage of that. Say you Make a partition half the size
of the drive - that partition is going to be on the inner (or outer) tracks
of the drive. Make another partition with the remaining space and that will
be on the outer (or inner) tracks.

Really though, it's not worth worrying about. People make a big deal out of
swap files and trying to optimise them, but theres not much point. Get
enough RAM so that you aren't hitting the swap file much (one interesting
side-note here is that, in modern OSs at least, you will always use the swap
file no matter how much RAM you have) and theres no problem. If you are
hitting the swap file a lot the solution is to get more RAM, not to optimise
the swap file.

Another interesting myth that has popped up in this discussion is that
ATA100 is faster than ATA66. Well, yes, the interface is faster, but the
drives that are attatched to them aren't (well they may be, but the
interface speed is no indication). Most current IDE hard drives can't
saturate an ATA66 interface - so moving to ATA100 makes very little
difference, certainly not a noticable difference.The same goes for ATA133
and SATA (150). There's a lot of marketting hype, but not much actual
difference.

Someone also suggested you put the swap file on the fastest drive - and that
it doesn't matter if it is the OS drive or not. That's bad advice. All of
the OS files are on that disk. It is also likely that program files, and
maybe also data files are on that disk. So whenever you access any of these
files (and this is the time when you are going to be using the swap file -
when reading new files into memory) you are using that hard drive. Put the
swap file on that same drive and you are now trying to read/write two
different parts of that drive at the same time. As above, I don't think it
really matters anyway, but still the information provided was incorrect and
I think it is worth pointing that out, if only for interest's sake.

One last thing. I said above you want to make sure you have enough RAM, but
I didn't say how to do that. One way is pretty obvious - look at your hard
drive activity light. If every time you open a new program, or switch
between open programs your hard drive suddenly goes crazy with activity
(that is swapping things into and out of virtual memory - known as
thrashing) then you really do need more RAM, as that's is a major bottleneck
in your system. If the problem isn't that obvious, you can use Task Manager
(assuming you use Windows). Load up your machine and use it normally for a
while. Play a game, have multiple apps open at once, load some large files
into photoshop..whatever you normally do. After a while open Task Manager
(right-click on the start bar and choose the obvious option, or
control-shift-escape) and choose the Performance tab. Under Commit Charge
there is a value called Peak. This is the most you have had in memory
(physical and virtual) in the current Windows session. If that is greater
than the amount of RAM you have (see Total under the Physical Memory
subheading if you are unsure) then you needed to hit the swap file. If it is
quite a bit more, then buying more RAM may well improve the speed of your
computer.

Gareth

- wrote:

>> Nope bad idea. Best option is outer edge of UDMA100 fat32 drive.
>
> How could you specify that you write the swap to the edge of the disk
> exactly?
>


Norton speed disk does this if the swap is a fixed size. MS's defragger
doesn't.

Given that the partition you specify is the first one on that disk that is..
--

Stacey

Gareth Church wrote:

>
> Another interesting myth that has popped up in this discussion is that
> ATA100 is faster than ATA66. Well, yes, the interface is faster, but the
> drives that are attatched to them aren't (well they may be, but the
> interface speed is no indication).

But ussually the ATA100 drives are newer/larger/faster than ATA66 drives but
agree neither saturates the ata66 format..


--

Stacey

"Stacey" <fotocord@yahoo.com> wrote in message
news:bdatmq$q468h$2@ID-52908.news.dfncis.de...
> > Another interesting myth that has popped up in this discussion is that
> > ATA100 is faster than ATA66. Well, yes, the interface is faster, but the
> > drives that are attatched to them aren't (well they may be, but the
> > interface speed is no indication).
>
> But ussually the ATA100 drives are newer/larger/faster than ATA66 drives
but
> agree neither saturates the ata66 format..

That's true. It must be at least a year or two since I've seen a new ATA66
drive for sale. I was replying more to someone else's comment that ATA100 is
"30% faster", which is fairly misleading.

Gareth

"Gareth Church" <gechurch@bigpond.com.au> wrote in message
news:WJ9Ka.963$p8.35745@newsfeeds.bigpond.com...
> "Stacey" <fotocord@yahoo.com> wrote in message
> news:bdatmq$q468h$2@ID-52908.news.dfncis.de...
> > > Another interesting myth that has popped up in this discussion is that
> > > ATA100 is faster than ATA66. Well, yes, the interface is faster, but
the
> > > drives that are attatched to them aren't (well they may be, but the
> > > interface speed is no indication).
> >
> > But ussually the ATA100 drives are newer/larger/faster than ATA66 drives
> but
> > agree neither saturates the ata66 format..
>
> That's true. It must be at least a year or two since I've seen a new ATA66
> drive for sale. I was replying more to someone else's comment that ATA100
is
> "30% faster", which is fairly misleading.

FreshDiagnose Ver.5.80 HDD Benchmarks.

Drive C: ATA133
Drive D: ATA66

Both on an ATA100 controller;

C: Ave. write speed: 16.2 Mb/s
Ave. read speed: 18.9 Mb/s

D: Ave. write speed: 9.1 Mb/s
Ave. read speed: 11.1 Mb/s

Don't be telling me there's no difference between ATA66 and higher.
--
~misfit~

==================

AMD Athlon XP1800+ T'bred 'B' core @ 1950Mhz. Standard HSF and vcore.


---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.491 / Virus Database: 290 - Release Date: 18/06/2003

"~misfit~" <misfit@'SPAMTRAP'orcon.net.nz> wrote in message
news:F0bKa.49342$JA5.866714@news.xtra.co.nz...
> > That's true. It must be at least a year or two since I've seen a new
ATA66
> > drive for sale. I was replying more to someone else's comment that
ATA100
> is
> > "30% faster", which is fairly misleading.
>
> FreshDiagnose Ver.5.80 HDD Benchmarks.
>
> Drive C: ATA133
> Drive D: ATA66
>
> Both on an ATA100 controller;
>
> C: Ave. write speed: 16.2 Mb/s
> Ave. read speed: 18.9 Mb/s
>
> D: Ave. write speed: 9.1 Mb/s
> Ave. read speed: 11.1 Mb/s
>
> Don't be telling me there's no difference between ATA66 and higher.

I have a few points to make here. First, I didn't say "there's no difference
between ATA66 and higher". I said most current drives can't saturate ATA66,
and going to a faster interface makes minimal difference. Second, the point
Stacey made was a fair one - no company has been selling new ATA66 drives
for quite a while now. That means that in general ATA66 drives are going to
be slower than ATA100 or better. But that is nothing to do with the
interface - just that the ATA66 drives are older so tend to be slower.
Third, showing the benchmarks for two drives, one ATA66 and another ATA100
proves nothing. I have no doubt I could find an ATA66 drive, and an ATA100
drive of about the same age (but slower due to other reasons) and use a
benchmark tool to show that the ATA100 drive is slower than the ATA66 drive.
That, like your test, would prove nothing other that one particular drive is
faster than another particular drive (on my particular machine). Fourth, if
you actually took the time to look at the results and think about them,
rather than just seeing that one number was bigger than another and jumping
to your conclusion, you will have noticed that even the ATA100 drive is well
below ATA33 speed (the number refers to how many megabytes the interface is
capable of transferring per second - eg. ATA33 can transfer 33 megabytes).
I'm afraid that has just proved my point - I said most drives can't saturate
ATA66 - you have show a drive that can't even saturate ATA33.

Gareth