Re: partioning HD w/NTFS


"Richard Urban" <richardurbanREMOVETHIS@xxxxxxxxxxx> wrote in message
news:%23%23YbHs7TGHA.5884@xxxxxxxxxxxxxxxxxxxxxxx
Please explain to all how multiple physical drives on the same IDE channel, which
operate consecutively - time sharing if you will, give an improvement over multiple
drives on two "separate" IDE channels which operate "concurrently" - at the same time?
We eagerly await the answer!

***
The latter is obviously the most beneficial to performance (2 physical drives on 2
separate IDE controllers).



This whole thread is about two drives on either one IDE channel or two drives on two IDE
channels, and the way the pagefile is handled in each scenario - if the pagefile is on a
second drive (either same IDE channel as the operating system drive or the secondary IDE
channel).

***
However, my original point still stands for IDE #0 controller containing two PHYSICAL
drives.

If the pagefile.sys is on drive #2 on that IDE controller, the head seek issue will have
been minimized (provided drive #2 doesn't have all kinds of intervening accesses to othe
files/folders). This is STILL an improvement over having the pagefile on physical drive
#1 of which the C:/boot partition (assumed here) usually has lots of background I/Os going
on stealing the head from potential pagefile.sys accesses (whether reads or writes).

To reiterate: obviously having each drive on its own IDE controller offers the maximum
POTENTIAL throughput. But depending on how one uses the PC dictates what desired
configuration you would want.

JL


--
Regards,

Richard Urban
Microsoft MVP Windows Shell/User

Quote from George Ankner:
If you knew as much as you think you know,
You would realize that you don't know what you thought you knew!

"PopS" <nobody@xxxxxxxxxxxxxxxxxxx> wrote in message
news:eDwLSa6TGHA.3888@xxxxxxxxxxxxxxxxxxxxxxx
Too wide a brushstroke there, I think:
That's true only if the drives are all on the same physical drive. Multiple physical
drives can and do give an improvement when they're set up properly.

Pop



"Richard Urban" <richardurbanREMOVETHIS@xxxxxxxxxxx> wrote in message
news:OnFdxftTGHA.1236@xxxxxxxxxxxxxxxxxxxxxxx
Wrong. If something wants to read from the pagefile at the same time the other drive
is writing to disk, the pagefile read gets bumped momentarily. It will have an adverse
effect on your system. It is better to leave it on drive C that to place it on a
second drive on the same IDE controller. You gain NOTHING by placing it on the second
drive - same controller. You, in fact, lose.

--
Regards,

Richard Urban
Microsoft MVP Windows Shell/User

Quote from George Ankner:
If you knew as much as you think you know,
You would realize that you don't know what you thought you knew!

"Jim Lewandowski" <jlewand@xxxxxxxxxxxxxxxxxxx> wrote in message
news:eunUf.3097$4L1.2303@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx

"Richard Urban" <richardurbanREMOVETHIS@xxxxxxxxxxx> wrote in message
news:ew7ffyhTGHA.2656@xxxxxxxxxxxxxxxxxxxxxxx
Placing the pagefile on a slave drive (connected to the same IDE channel) buys you a
sum total of nothing.

***
Not so. By having the paging file reside on another drive, at least he can minimize
head (seek) movement. That's worth about 10 ms saved per seek. Again, if the paging
file is the ONLY file being accessed on the second hard drive or the one primarily
being accessed (few intervening files read off same drive of other files/folders),
that would be of benefit as the heads would always be parked right where Windows
paging subsystem would like them to be.

Remember, lots of background CONTINUAL disk accesses can/do occur with Windows/XP
(NAV, system tasks, NTFS logging).

Additionally, when physical paging occurs, it is usually a SYNCHRONOUS process to
"real work" (user applications) within the system . IOW, an applications use of
memory has caused Windows to physically page out other blocks of memory. Is there
ever paging file activity at the same time for OTHER disk accesses? Not likely. It
can happen if one is launching a new program (reading exe/dll stuff into memory
forcing physical disk paging into action).


In fact, it may slow the system down. The
two drives write consecutively. When one drive is in read/write mode the other drive
CAN NOT do the same. The time is split between the two drives.

***
See head/seek placement issue above.



Place the pagefile on a hard drive on the secondary IDE controller and you can have
concurrent reading/writing of BOTH drives. Here you will likely see a gain if you
are using programs that make extensive use of the pagefile.

***
The final question is: what % of the time is your paging file PHYSICALLY being
written to/read from?

JL


--
Regards,

Richard Urban
Microsoft MVP Windows Shell/User

Quote from George Ankner:
If you knew as much as you think you know,
You would realize that you don't know what you thought you knew!

"jsph1961" <jsph1961@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:4206477E-03D0-46D5-9BEE-5F5B874B69C7@xxxxxxxxxxxxxxxx
I have 2 HDs on my PC, on the master (80GB) I have XP pro and everything
else. On the slave(40GB)I put a paging file and nothing else. I want to
partition so that my O/S is alone on the master ;so what size will suffice
for the O/S? I then want to put all the hot fixes into a folder on a seperate
partition. Then all my games on another then text files on another and so
on... Also concerning paging file. By having the paging file on a different
HD other than where the O/S is I'm concerned about the memory dump. How
large should I make the paging file on the master to handle the memory dump?
And how large should the paging file be on the slave which is accessed less?
I have 1 GB of memory to play with. Will the /S be able to access the games
even if they are on a HD without an O/S on it?












.

Relevant Pages

  • Re: Strange PEER error with Danis 506 1.81 generic question
    ... In the case of everything so far up until the Intel 915GAVL/GEVL and still in research 945 series motherboards, there have been two IDE controllers on the boards. ... Thus the primary hard disk controller from which the finished box will normally boot would be the Master on IDE controller #1. ... if there were, for example, two IDE hard drives, one would be the normal boot device as the Master on the primary IDE controller. ... to generate a report of scsi and "fake scsi" devices on a mixed scsi hard drives/ide dvd burner and dvd-rom system but I seem to recall that all scsi devices got listed before any ide devices - including the scsi scanner I had. ...
    (comp.os.os2.bugs)
  • RE: Why would drive run at UDMA33? (Segate 80GB)
    ... I have an PC with mfg date of 5/2003 and the motherboard manually ... IDE controller because it will force the IDE controller to step down ... drives but also mixing UDMA100 and UDMA66 and UDMA33 disk on any IDE ...
    (freebsd-questions)
  • Re: Server Migration Issue
    ... there the BIOS would automagically offer it as a choice in the boot menu. ... I belive it would _possibly_ be safer to use an IDE RAID controller (not ... It actually has an IDE controller which is connected to the ... >> to the SATA drives either by processes similar to Norton Ghost or by ...
    (microsoft.public.windows.server.sbs)
  • Re: Strange PEER error with Danis 506 1.81 generic question
    ... models has been that the burner must be master and there must be no other activity on the same IDE controller while burning is going on. ... I can understand this where we might be talking about an IDE hard drive setup with a burner. ... Also, when I did have two cd or dvd drives in the same system, I never had any success with cd to cd or dvd to dvd copies. ...
    (comp.os.os2.bugs)
  • Re: Server Migration Issue
    ... It actually has an IDE controller which is connected to the DVD ... But the server BIOS won't allow ... >> The new machine can only accept SATA drives while the current server ...
    (microsoft.public.windows.server.sbs)