Help to interpret HD Tune results please

On Tue, 22 Jan 2013 07:46:59 +1100, Franc Zabkar
wrote:

On Tue, 22 Jan 2013 07:34:36 +1100, Franc Zabkar
put finger to keyboard and composed:

I looked, but I missed it. Sorry Jamie.

Sorry, that should have been "Jaimie".

Don't worry about it - I'm not fussy! Good info on the platters, I
didn't think that would be done under the one model number. I should
be less optimistic...

Cheers - Jaimie
--
haiku are easy
all you do is stop at the
seventeenth syllab

On Tue, 22 Jan 2013 09:47:01 +1100, Franc Zabkar
put finger to keyboard and composed:

What if Samsung took two 500GB platters and reduced their capacities
by carving 250GB from the outermost zones? They could do so by
reducing the bits-per-inch to a value that stays within the frequency
response limits of the older heads. That would account for the
plateau, and the burst rate, and the 500GB-per-platter data rate at
the innermost zones.

In retrospect, carving 25% of the total capacity from the outermost
zones seems excessive. Instead, what if the bulk of the capacity could
be cut by reducing the number of tracks per zone across the entire
platter?

For example, let's assume we have a 1TB drive with two 500GB platters
and 4 heads. We now wish to create a 750GB drive with two 375GB
platters and 4 heads. Let's assume that the average number of tracks
per zone is 1000 and 750, respectively. If we reduce the tracks per
zone from 1000 to 800, say, then the capacity would be cut from 1TB to
800GB. We could then carve the remaining 50GB from the outermost zones
by limiting their bit rate.

By doing it this way, we would be relaxing the requirements on the
track servo, as well as on the heads.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

On Tue, 22 Jan 2013 11:14:17 +1100, Franc Zabkar
wrote:

On Tue, 22 Jan 2013 09:47:01 +1100, Franc Zabkar
put finger to keyboard and composed:

What if Samsung took two 500GB platters and reduced their capacities
by carving 250GB from the outermost zones? They could do so by
reducing the bits-per-inch to a value that stays within the frequency
response limits of the older heads. That would account for the
plateau, and the burst rate, and the 500GB-per-platter data rate at
the innermost zones.

In retrospect, carving 25% of the total capacity from the outermost
zones seems excessive. Instead, what if the bulk of the capacity could
be cut by reducing the number of tracks per zone across the entire
platter?

For example, let's assume we have a 1TB drive with two 500GB platters
and 4 heads. We now wish to create a 750GB drive with two 375GB
platters and 4 heads. Let's assume that the average number of tracks
per zone is 1000 and 750, respectively. If we reduce the tracks per
zone from 1000 to 800, say, then the capacity would be cut from 1TB to
800GB. We could then carve the remaining 50GB from the outermost zones
by limiting their bit rate.

By doing it this way, we would be relaxing the requirements on the
track servo, as well as on the heads.

I don't think you would - the track would still be as narrow, and so
would require the same accuracy in servo and heads for pickup and
tracking. No change to requirements of either.

Your original suggestion requires just one change to the firmware -
stroke length limit - and would also boost the performance of the
drive relative to other 750gig devices, good for review purposes. This
new suggestion would require several changes and not boost
performance.

Cheers - Jaimie
--
There he saw Merry's feet still sticking out - the rest had already been
drawn further inside. Tom put his mouth to the crack and began singing
into it in a low voice. They could not catch the words, but evidently
Merry was aroused. -- J R R Tolkien

On Tue, 22 Jan 2013 00:18:08 +0000, Jaimie Vandenbergh
put finger to keyboard and composed:

Your original suggestion requires just one change to the firmware -
stroke length limit - and would also boost the performance of the
drive relative to other 750gig devices, good for review purposes. This
new suggestion would require several changes and not boost
performance.

I re-examined the curve for the fast drive. ISTM that, if
extrapolated, it could intersect the vertical axis at about 120MB/s or
less. If so, then this would suggest that it may have two slightly
shortstroked 400GB platters. However, this then leaves us with the
enigma of a 140MB/s burst rate and a 110MB/s plateau.

I tried to compare the drive against a single platter 500GB Samsung
HDD, but I only found the following Advanced Format model:

http://imageshack.us/scaled/landing/522/hd502hj.jpg

AF drives typically pack 10% more bits per track than their non-AF
equivalents, so maybe a 7200RPM 500GB non-AF Samsung drive would have
a maximum transfer rate of around 130MB/s (144.5 / 1.10). Therefore
Terry's drive probably doesn't have 500GB platters.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

In article , Franc Zabkar
writes

I wonder if HD Tune's Burst Rate result of 140MB/s can be believed. If
so, then this would suggest that the SATA interface is not a
bottleneck.

Isn't burst rate a transfer between the drive's cache and the
controller? If so, 140MB/s would seem believable for SATA1 (150MB/s)
but slow for SATA2 (300MB/s).

--
(\_/)
(='.'=)
(")_(")

In article , Jaimie
Vandenbergh writes

Don't worry about it - I'm not fussy! Good info on the platters, I
didn't think that would be done under the one model number.

You would hope not. That was one of the reasons I asked Terry to post
the output from Device Mangler, in the hope that it would show a subtle
variation in the drive model numbers to reflect the difference in spec.
As it is, he might have to whip the lid off and look at the labels.

--
(\_/)
(='.'=)
(")_(")

On 21/01/2013 02:57, Mike Tomlinson wrote:
En el artículo , Terry
Pinnell escribió:

Thanks Mike. I'm not brave enough to make that mode change, especially
after reading further about XP sensitivity to this, BSODs etc!

I might give it a go if feeling brave enough one day.

I've changed from IDE to AHCI or RAID several times in XP, but always
with image backups, just in case.
IF you are lucky enough to have TWO HD controllers on your mobo, it
is painless. Luckily most of the systems I've done were mid-high
end Asus boards, so had both an Intel ICHx chipset and a second HD
controller, such as a JMicron, Silicon Image, Marvell etc.

Just disconnect the boot drive from the chipset (ICHx) controller
and plug it in to the 2nd controller. Check that it boots
normally (if not, check the 2nd controller is enabled and set to
IDE mode in BIOS - may need to swap back to ICH header and install
drivers for the 2nd controller if any settings were changed.)
Once booting normally from the 2nd controller, restart, go back
into the BIOS and change the main (ICH) controller to AHCI or RAID,
as required.
Boot XP and install the ICH AHCI/RAID drivers (typically, by
running the Intel RST setup.)
Shutdown, swap cable back to one of the ICH SATA headers and
reboot into XP which now has the correct drivers installed.

On a basic mobo with only an Intel ICH controller, it was more
painful. Most of those also have a PATA connector, which is
crucial for this method, as is a spare PATA drive:
1) Make image of SATA boot drive.
2) Restore image to temporary PATA drive.
3) Disconnect SATA cable, check/configure BIOS so it boots OK
to the temporary PATA drive.
4) Restart, change SATA to AHCI or RAID as needed in BIOS.
5) Boot XP on PATA, install SATA AHCI/RAID driver (typically by
installing Intel RST.)
6) Image PATA drive (which now has the correct drivers installed.)
7) Remove PATA drive, connect SATA drive then restore image from
step 6 to the SATA drive.
8) Boot to XP on SATA drive, which now had the correct drivers.

Hope someone finds these methods useful - no mucking about with
the Registry needed.

Cheers,
--
Rob

En el artículo , Rob
escribió:

I've changed from IDE to AHCI or RAID several times in XP, but always
with image backups, just in case.
IF you are lucky enough to have TWO HD controllers on your mobo, it
is painless.

[chomp]

Thanks for that, Rob. I have a Z77 chipset board, and I think all the
SATA interfaces (2 x SATA3 and 4 x SATA2) hang off the ICH chipset, so I
don't have the option of using a second controller to do the shuffling
you describe.

Searched the hard drive for msahci.sys, which isn't present, so if I
were to make changing from IDE mode to AHCI work I would need to install
the Intel AHCI driver, shut down, change the BIOS from IDE to AHCI, then
reboot and hopefully that would do the trick. I don't see the need to
install the full Intel RST software, shirley all it needs is the Intel
variant of msahci.sys to be present?

In any case, I'll make sure I have an image first

--
(\_/)
(='.'=)
(")_(")

On 26/01/2013 17:24, Mike Tomlinson wrote:

Searched the hard drive for msahci.sys, which isn't present, so if I
were to make changing from IDE mode to AHCI work I would need to install
the Intel AHCI driver, shut down, change the BIOS from IDE to AHCI, then
reboot and hopefully that would do the trick.

It used to be the case that the Intel AHCI driver would refuse to
install unless you already had AHCI enabled. I don't know if they
changed that recently.


--
John Jordan

En el artículo , John Jordan
escribió:

It used to be the case that the Intel AHCI driver would refuse to
install unless you already had AHCI enabled. I don't know if they
changed that recently.

Ta. Perhaps I can extract it from the CAB/.zip/whatever and stick it in
%windir%\system32.

--
(\_/)
(='.'=)
(")_(")