partition and speed

[Rob Hines]

I just built a new system and partitioned the disk into 4 partitions. I
wanted to multiboot the system and have a choice on where to put each
partition - W2000, XP and Linux. Before I start installing the OSes, I
want to determine which partition will give the best performance for each.

Can anyone tell me a) if data stored in the C: partition is closest to
the spindle, and b) if there's a difference in performance between the
partition closest to the spindle vs. partition fartherest away from it.

 

[JAD]

this is a joke right?

further away from the spindle? and you think your going to tell the
difference....like the faster speed you get when you only use a 3rd of
a cd rom disk cause it burns from the center out?

 

[Shep©]

I just built a new system and partitioned the disk into 4 partitions. I
wanted to multiboot the system and have a choice on where to put each
partition - W2000, XP and Linux. Before I start installing the OSes, I
want to determine which partition will give the best performance for each.

Won't matter.

Can anyone tell me a) if data stored in the C: partition is closest to
the spindle, and b) if there's a difference in performance between the
partition closest to the spindle vs. partition fartherest away from it.

Nope.You will slow down the system no matter where you put the O/S coz
you is asking silly questions



--
Free Windows/PC help,
http://www.geocities.com/sheppola/trouble.html
remove obvious to reply
email (e-mail address removed)
Free songs to download and,"BURN" :O)
http://www.soundclick.com/bands/8/nomessiahsmusic.htm

 

[David Maynard]

Won't matter.

It surely will.

Nope.You will slow down the system no matter where you put the O/S coz
you is asking silly questions

It isn't a silly question. Modern hard drives use zoned bit recording,
meaning the number of sectors per track varies depending on where the track
is located, inner to outer, for the obvious reason that an outer track has
a larger circumference than the inner tracks and can, therefor, hold more
sectors (at the same linear bit density).

For example, the IBM Deskstar 34GXP (model DPTA-373420) has a media
transfer rate of between approximately 171 Mb/s and 284 Mb/s depending
where on the disk you are reading: that drive has 12 different zones with
272 sectors in its innermost zone and 452 sectors on its outside tracks.

Which means it's 66% faster on the outer tracks than on the inner; not a
trivial or 'silly' difference.

 

[Spajky]

Before I start installing the OSes, I

It surely will.

I agree, but differences should not be much IMHO

Modern hard drives use zoned bit recording,
meaning the number of sectors per track varies depending on where the track
is located, inner to outer, for the obvious reason that an outer track has
a larger circumference than the inner tracks and can, therefor, hold more
sectors (at the same linear bit density).

For example, the IBM Deskstar 34GXP (model DPTA-373420) has a media
transfer rate of between approximately 171 Mb/s and 284 Mb/s depending
where on the disk you are reading: that drive has 12 different zones with
272 sectors in its innermost zone and 452 sectors on its outside tracks.

Which means it's 66% faster on the outer tracks than on the inner; not a
trivial or 'silly' difference.

thats quite a lot, but benching not real life situation (teorethical
max speed); using cumulative results with "indexed or scored" results
should give less than 10% real life performance (PCMark2002 for
example.

the average speed also has to do something IMHO with HD firmware; for
example my QuantumF+AS 20Gb gives best results in the middle of the
its capacity; I was also surprised to see that results since before I
thought should be C part fastest (with 2HDs I had before, it was so).
Some benching can reveal interesting behavior sometimes ...

 

[David Maynard]

I agree, but differences should not be much IMHO



thats quite a lot, but benching not real life situation (teorethical
max speed); using cumulative results with "indexed or scored" results
should give less than 10% real life performance (PCMark2002 for
example.

It also makes a difference what assumptions the benchmark program is using.
He's not talking about where one 'one large partition', with everything
jumbled together, is it 'faster'. He's talking about partitioning the drive
and using them for specific purposes.

the average speed also has to do something IMHO with HD firmware; for
example my QuantumF+AS 20Gb gives best results in the middle of the
its capacity;

How did you restrict a general purpose benchmark program to 'the middle'?

I was also surprised to see that results since before I
thought should be C part fastest (with 2HDs I had before, it was so).
Some benching can reveal interesting behavior sometimes ...

When I said "66%" faster I was talking about the linear data rate only.
Latency and seek times are the same so it depends, as usual, on what one is
doing with it. Jumping around a gaggle of small files isn't going to show
much, if any, difference because latency and seek time effects will swamp
the track data rate differences.

When used for a contiguous swap file, however, the data rate comes into play.

 

[JAD]

ROFLMAO............66% faster than WHAT,,,and that breaks down to
What? 5 milliseconds and you can tell with your own god given
talents that its faster..........silly question and a silly
answer.....

each.


Won't matter.

It surely will.

it.


Nope.You will slow down the system no matter where you put the O/S coz
you is asking silly questions

It isn't a silly question. Modern hard drives use zoned bit recording,
meaning the number of sectors per track varies depending on where the
track
is located, inner to outer, for the obvious reason that an outer track
has
a larger circumference than the inner tracks and can, therefor, hold
more
sectors (at the same linear bit density).

For example, the IBM Deskstar 34GXP (model DPTA-373420) has a media
transfer rate of between approximately 171 Mb/s and 284 Mb/s depending
where on the disk you are reading: that drive has 12 different zones
with
272 sectors in its innermost zone and 452 sectors on its outside
tracks.

Which means it's 66% faster on the outer tracks than on the inner; not
a
trivial or 'silly' difference.

 

[David Maynard]

ROFLMAO............66% faster than WHAT

What part of outer track vs inner track did you not understand?

,,,and that breaks down to
What?

284 Mb/s vs 171 Mb/s, for the example IBM Deskstar 34GXP.

5 milliseconds and you can tell with your own god given
talents that its faster.......... silly question and a silly
answer.....

Did you forget to take your medication today?

 

[JAD]

here in the real world makes little difference

ROFLMAO............66% faster than WHAT

What part of outer track vs inner track did you not understand?

,,,and that breaks down to
What?

284 Mb/s vs 171 Mb/s, for the example IBM Deskstar 34GXP.

5 milliseconds and you can tell with your own god given
talents that its faster.......... silly question and a silly
answer.....

Did you forget to take your medication today?

 

[David Maynard]

here in the real world makes little difference

It won't, to those who don't now how to utilize it.

 

[JAD]

look, at the outside we are looking at milliseconds of
difference,,,,,what would you be using, that would utilize the
microseconds that your talking about?(SETI?) Except the 'benchmark
log' boasting rights? Its cool to know there is a difference, but the
info makes little difference from the average users/consumers POV.

here in the real world makes little difference

It won't, to those who don't now how to utilize it.

 

[Spajky]

How did you restrict a general purpose benchmark program to 'the middle'?

with "demo" version it benches only C part
with "registered" (pro) version, you can chose in options which
partition to bench .. & to display as HD score ...

 

[David Maynard]

with "demo" version it benches only C part
with "registered" (pro) version, you can chose in options which
partition to bench .. & to display as HD score ...

I asked because you didn't say anything about setting up partitions
segmented across your drive to test specific areas of it.

 

[David Maynard]

look, at the outside we are looking at milliseconds of
difference,,,,

What you're looking at, on the example drive, it a data rate 66% faster
than on the inner most tracks.

I have no idea how you arrive at 'milliseconds'.

,what would you be using, that would utilize the
microseconds that your talking about?(SETI?)

Be rather silly to use a CPU bound task for gauging disk performance.

Except the 'benchmark
log' boasting rights? Its cool to know there is a difference, but the
info makes little difference from the average users/consumers POV.

There are a lot of things the 'average user' doesn't know about; one of
which probably is how to even partition the drive in the first place.

Although, if they've got a contiguous page file and are using Speed Disk
they don't need to as it automatically moves the page file to the front for
that very reason.

 

[JAD]

hehehehe you keep using this 66% faster, THAN WHAT? 66% FASTER
faster =time time in regards to HD's is milliseconds. Its not hard
to figure it out.

look, at the outside we are looking at milliseconds of
difference,,,,

What you're looking at, on the example drive, it a data rate 66%
faster
than on the inner most tracks.

I have no idea how you arrive at 'milliseconds'.

,what would you be using, that would utilize the
microseconds that your talking about?(SETI?)

Be rather silly to use a CPU bound task for gauging disk performance.

Except the 'benchmark
log' boasting rights? Its cool to know there is a difference, but the
info makes little difference from the average users/consumers POV.

There are a lot of things the 'average user' doesn't know about; one
of
which probably is how to even partition the drive in the first place.

Although, if they've got a contiguous page file and are using Speed
Disk
they don't need to as it automatically moves the page file to the
front for
that very reason.

 

[Spajky]

I asked because you didn't say anything about setting up partitions
segmented across your drive to test specific areas of it.

I have it partitioned in few partitions & the middle ones are the
fastest benched. I also run HDTach /read only, not registered/ & also
there was the highest "hill" in the bench around middle of its
capacity .... interesting this with my drive ... must be something in
firmware set so to minimize differences of performance depending on
data position on the platters ... I don´t know ...

 

[David Maynard]

hehehehe you keep using this 66% faster, THAN WHAT?

And the 'what' was explained: 284 Mb/s vs 171 Mb/s, for the example IBM
Deskstar 34GXP.

You ARE capable of calculating a percentage increase from 171 to 284,
aren't you?

66% FASTER
faster =time

Well, no. 'Faster' can be expressed as either a decrease in time to
accomplish something or an increase of things accomplished in a fixed time.
such as transferring 284 Mb/s vs 171Mb/s, but in neither case is it simply
'time'.

Number of things accomplished per fixed unit of time is the more common:
MPH, RPM, bps, and, the ever popular, furlongs per fortnight.

time in regards to HD's is milliseconds.

Time is time. 'Milli' is simply a decimal system prefix to seconds; e.g.
.001 seconds. Seconds, itself, is simply an arbitrary measure derived from
it's convenience in dividing up the time it takes the earth to rotate from
high noon to high noon.

To expound, 12 occurs in so many measurements from ancient times because it
is conveniently, for people unfamiliar with fractions, divisible by 2, 3,
and 4. 60 is the next convenient number, being divisible by 2, 3, 4, and 5.
Hence, two 12 hour periods in a day, 60 minutes per hour, and 60 seconds
per minute.

Its not hard
to figure it out.

Right: ((284/171)-1)*100=66.081871345029239766081871345029%

Which is why your difficulty in figuring it out is a mystery.

 

[JAD]

all results in nothing noticeable except to BM watchers......drop your
line somewhere else, this spot has gone silent.

hehehehe you keep using this 66% faster, THAN WHAT?

And the 'what' was explained: 284 Mb/s vs 171 Mb/s, for the example
IBM
Deskstar 34GXP.

You ARE capable of calculating a percentage increase from 171 to 284,
aren't you?

66% FASTER
faster =time

Well, no. 'Faster' can be expressed as either a decrease in time to
accomplish something or an increase of things accomplished in a fixed
time.
such as transferring 284 Mb/s vs 171Mb/s, but in neither case is it
simply
'time'.

Number of things accomplished per fixed unit of time is the more
common:
MPH, RPM, bps, and, the ever popular, furlongs per fortnight.

time in regards to HD's is milliseconds.

Time is time. 'Milli' is simply a decimal system prefix to seconds;
e.g.
..001 seconds. Seconds, itself, is simply an arbitrary measure derived
from
it's convenience in dividing up the time it takes the earth to rotate
from
high noon to high noon.

To expound, 12 occurs in so many measurements from ancient times
because it
is conveniently, for people unfamiliar with fractions, divisible by 2,
3,
and 4. 60 is the next convenient number, being divisible by 2, 3, 4,
and 5.
Hence, two 12 hour periods in a day, 60 minutes per hour, and 60
seconds
per minute.

Its not hard
to figure it out.

Right: ((284/171)-1)*100=66.081871345029239766081871345029%

Which is why your difficulty in figuring it out is a mystery.

 

[David Maynard]

all results in nothing noticeable except to BM watchers

No, you mean *you* don't notice it, although one wonders how you COULD as I
doubt you've ever bothered optimizing disk partitions. But I wouldn't be
too surprised as you also don't notice 7200 vs 5400 RPM either.

......drop your line somewhere else, this spot has gone silent.

This is the first I heard of facts and basic math being called a 'line'.

 

[Shep©]

It won't, to those who don't now how to utilize it.

<Mega big snip>
Dave.You have a lot of great tech info and knowledge but,"Jad" is
right.No matter where or what the O/P does with his partitions/drives
he'll screw up the system and slow it down so as to make all your help
worthless :/
Dats peeeeeple for you :/
I've benchmarked,tuned and done so much shit on this just to get me
online games an extra nano-second I know from whence I come.Den I go
and screw it up by installing some shit software :/



--
Free Windows/PC help,
http://www.geocities.com/sheppola/trouble.html
remove obvious to reply
email (e-mail address removed)
Free songs to download and,"BURN" :O)
http://www.soundclick.com/bands/8/nomessiahsmusic.htm