BTX Technology/Native Command Queuing

[AJ]

AJ wrote:
I'll bet NCQ gets
over-marketed to stand-alone users though too (because the HD manf's have nothing
new to offer this year?). I was chomping-at-the-bit (hoping) for 3GB/s. I'd actually buy
a new HD and use my exiting 80GB SATA for backups if 3 GB/s was available (twice

[...]

Anywayz... I digress. I sure would like 3GB/s drives though, especially 2.5 inch ones!

It'll be a while before we see consumer-level drives with 3
Giga*byte*/sec interfaces. The 3G*bit*/sec interface (SATA-II) will
be here soon, but it's only one tenth of that (300MByte/s).

My bad. That's what I meant: 3Gb/s, twice the throughut of today's
SATA drives.

Find me a SATA drive with even 1.5Gb/s and I'll buy it. Hint: datarate

I don't think so, cuz where I heard about it was at the drive
manufacturers sites where they describe SATA technology and they are
talking about individual drive specs.

Read the above again. It matters *not* how fast the interface is, if the
media can't keep up. ...and it cannot.

Perhaps that's the roadblock then at this time and why it's not available.
Read the above again: the poster said he thought the 3Gb/s spec was
a motherboard bus spec rather than a per-drive spec, to which I thought
not. The assumption made was that the 3Gb/s throughput would be
realized (duh, of course) and not be just a theoretical number. (Of course
if one couldn't wait for 3Gb/s SATA, one could do RAID and double their
throughput now).

Keep dreaming. Win install is a little more than disk I/O too.

But indeed very disk intensive. My guess is that a faster drive
("drive subsystem" for all the intensly literal here) would do
more to improve the times of that process than CPU horsepower.
Also, boot times and wake from hibernation would benefit.
(Aside: It takes <20 mins to install (unattended install) with my P4
2.4 GHz/800 FSB and SATA-I drive. Boot time about 15 s.)

AJ

 

[Roy Coorne]

Tony Hill wrote:

....

Obviously a Mac isn't an option for everyone, I don't own one either
though would like to get one if I could afford it.

Happy New Mac Year
http://www.thinksecret.com/news/0412expo2.html

Roy

 

[Tony Hill]

Tony Hill wrote:

...

Happy New Mac Year
http://www.thinksecret.com/news/0412expo2.html

Hmm.. that does sound MUCH better than the eMac and iMac with built-in
displays (my monitor is currently 5 years old, every other component
in my system has been upgraded two to 4 times in that timeframe... so
why would I want to tie the two together?!?). Unfortunately I don't
really want/need another desktop system (I've got 4 of 'em in the
house already... and I live on my own! :> ).

What I really want is one of the iBook systems, but with the cheapest
one being $1000, it's definitely out of my price range. Hell, at the
moment that $500 system is well out of my price range! :>

 

[keith]

AJ wrote:
I'll bet NCQ gets
over-marketed to stand-alone users though too (because the HD manf's have nothing
new to offer this year?). I was chomping-at-the-bit (hoping) for 3GB/s. I'd actually buy
a new HD and use my exiting 80GB SATA for backups if 3 GB/s was available (twice

[...]

Anywayz... I digress. I sure would like 3GB/s drives though, especially 2.5 inch ones!

It'll be a while before we see consumer-level drives with 3
Giga*byte*/sec interfaces. The 3G*bit*/sec interface (SATA-II) will
be here soon, but it's only one tenth of that (300MByte/s).

My bad. That's what I meant: 3Gb/s, twice the throughut of today's
SATA drives.

Find me a SATA drive with even 1.5Gb/s and I'll buy it. Hint: datarate

Points to
consider: 1) drives cannot deliver data fast enough internally to flood
even a SATA-I interface, and 2) Besides, I doubt that we will see many
drives offering 3Gb/s. If I understand it correctly, the 3Gb/s is
intended for between the mobo and "expander chips", so that you can
attach multiple physical drives to a single SATA-II connector at the
mobo level.

I don't think so, cuz where I heard about it was at the drive
manufacturers sites where they describe SATA technology and they are
talking about individual drive specs.

Read the above again. It matters *not* how fast the interface is, if the
media can't keep up. ...and it cannot.

Perhaps that's the roadblock then at this time and why it's not available.
Read the above again: the poster said he thought the 3Gb/s spec was
a motherboard bus spec rather than a per-drive spec, to which I thought
not. The assumption made was that the 3Gb/s throughput would be
realized (duh, of course) and not be just a theoretical number. (Of course
if one couldn't wait for 3Gb/s SATA, one could do RAID and double their
throughput now).

You're still not getting it. The interface could be 300Gb/s and it
wouldn't matter. The data can't get to/from the platter that fast. The
spec you want to be looking for is the "sustained transfer rate" or STR.
The interface rate is meaningless.

But indeed very disk intensive. My guess is that a faster drive ("drive
subsystem" for all the intensly literal here) would do more to improve
the times of that process than CPU horsepower. Also, boot times and wake
from hibernation would benefit. (Aside: It takes <20 mins to install
(unattended install) with my P4 2.4 GHz/800 FSB and SATA-I drive. Boot
time about 15 s.)

Win2K install on my Opteron 146 is a tad faster (haven't measured it, but
it's noticable) than on my K6-III, even though the K6-III has a
newer/faster drive. It's not like installation is an important benchmark
though.

 

[cquirke (MVP Win9x)]

The interface could be 300Gb/s and it wouldn't matter. The data
can't get to/from the platter that fast.

I'm wondering about that, given the expectation that the HD's on-board
RAM should to some extent de-couple the data transfer from raw platter
data speed. Even if UDMA means the CPU isn't actually tied up with
the data transfer between RAM and HD unit, there may still be impact
on the system if this transfer hits a ceiling.

The issue may arise only when multiple devices are on the bus, and may
itself be limited by PCI if the UIDE is still chained to that.

--------------- ----- ---- --- -- - - -

Dreams are stack dumps of the soul

 

[daytripper]

I'm wondering about that, given the expectation that the HD's on-board
RAM should to some extent de-couple the data transfer from raw platter
data speed.

"Some extent" is easy to calculate when you add the context of, say, an 8MB
drive cache.

The answer is: not much beneficial effect - when you want to transfer a
random, 9+ MB file.

Spiral Transfer Rate is nigh immutable when setting system io bandwidth
expectations...

 

[AJ]

On Wed, 29 Dec 2004 09:14:03 +0000, AJ wrote:



AJ wrote:
I'll bet NCQ gets
over-marketed to stand-alone users though too (because the HD manf's have nothing
new to offer this year?). I was chomping-at-the-bit (hoping) for 3GB/s. I'd actually buy
a new HD and use my exiting 80GB SATA for backups if 3 GB/s was available (twice

[...]

Anywayz... I digress. I sure would like 3GB/s drives though, especially 2.5 inch ones!

It'll be a while before we see consumer-level drives with 3
Giga*byte*/sec interfaces. The 3G*bit*/sec interface (SATA-II) will
be here soon, but it's only one tenth of that (300MByte/s).

My bad. That's what I meant: 3Gb/s, twice the throughut of today's
SATA drives.

Find me a SATA drive with even 1.5Gb/s and I'll buy it. Hint: datarate
<> thoughput. If you can't get it to/from the platter, it doesn't matter.

Points to
consider: 1) drives cannot deliver data fast enough internally to flood
even a SATA-I interface, and 2) Besides, I doubt that we will see many
drives offering 3Gb/s. If I understand it correctly, the 3Gb/s is
intended for between the mobo and "expander chips", so that you can
attach multiple physical drives to a single SATA-II connector at the
mobo level.

I don't think so, cuz where I heard about it was at the drive
manufacturers sites where they describe SATA technology and they are
talking about individual drive specs.

Read the above again. It matters *not* how fast the interface is, if the
media can't keep up. ...and it cannot.

Perhaps that's the roadblock then at this time and why it's not available.
Read the above again: the poster said he thought the 3Gb/s spec was
a motherboard bus spec rather than a per-drive spec, to which I thought
not. The assumption made was that the 3Gb/s throughput would be
realized (duh, of course) and not be just a theoretical number. (Of course
if one couldn't wait for 3Gb/s SATA, one could do RAID and double their
throughput now).

You're still not getting it. The interface could be 300Gb/s and it
wouldn't matter. The data can't get to/from the platter that fast. The
spec you want to be looking for is the "sustained transfer rate" or STR.
The interface rate is meaningless.

Well context matters. You know what I meant. Point: more throughput.
Call it what you want. If you know what indeed is the roadblock to
getting 3Gb/s out the door, plz do tell, I'm curious.

Win2K install on my Opteron 146 is a tad faster (haven't measured it, but
it's noticable) than on my K6-III, even though the K6-III has a
newer/faster drive. It's not like installation is an important benchmark
though.

It is to me, for recovery purposes. Quite important actually.

AJ

 

[Kai Harrekilde-Petersen]

Well context matters. You know what I meant. Point: more throughput.
Call it what you want. If you know what indeed is the roadblock to
getting 3Gb/s out the door, plz do tell, I'm curious.

I don't think there's any "roadblock" to get drives with SATA-II
interfaces (3Gb/s) out of the door, except the "Why bother?" factor.
For the drive manufacturers there is little incentive to use SATA-2
interfaces, except, perhaps, a tiny improvement in latency.

There might also be a consideration from the drive manufacturers to
the fact that having to mux two 3Gb/s interfaces onto a single 3Gb/s
requires bigger buffers in the expander chips.


Kai

 

[keith]

On Wed, 29 Dec 2004 09:14:03 +0000, AJ wrote:



AJ wrote:
I'll bet NCQ gets
over-marketed to stand-alone users though too (because the HD manf's have nothing
new to offer this year?). I was chomping-at-the-bit (hoping) for 3GB/s. I'd actually buy
a new HD and use my exiting 80GB SATA for backups if 3 GB/s was available (twice

[...]

Anywayz... I digress. I sure would like 3GB/s drives though, especially 2.5 inch ones!

It'll be a while before we see consumer-level drives with 3
Giga*byte*/sec interfaces. The 3G*bit*/sec interface (SATA-II) will
be here soon, but it's only one tenth of that (300MByte/s).

My bad. That's what I meant: 3Gb/s, twice the throughut of today's
SATA drives.

Find me a SATA drive with even 1.5Gb/s and I'll buy it. Hint: datarate
<> thoughput. If you can't get it to/from the platter, it doesn't matter.

Points to
consider: 1) drives cannot deliver data fast enough internally to flood
even a SATA-I interface, and 2) Besides, I doubt that we will see many
drives offering 3Gb/s. If I understand it correctly, the 3Gb/s is
intended for between the mobo and "expander chips", so that you can
attach multiple physical drives to a single SATA-II connector at the
mobo level.

I don't think so, cuz where I heard about it was at the drive
manufacturers sites where they describe SATA technology and they are
talking about individual drive specs.

Read the above again. It matters *not* how fast the interface is, if the
media can't keep up. ...and it cannot.

Perhaps that's the roadblock then at this time and why it's not available.
Read the above again: the poster said he thought the 3Gb/s spec was
a motherboard bus spec rather than a per-drive spec, to which I thought
not. The assumption made was that the 3Gb/s throughput would be
realized (duh, of course) and not be just a theoretical number. (Of course
if one couldn't wait for 3Gb/s SATA, one could do RAID and double their
throughput now).

You're still not getting it. The interface could be 300Gb/s and it
wouldn't matter. The data can't get to/from the platter that fast. The
spec you want to be looking for is the "sustained transfer rate" or STR.
The interface rate is meaningless.

Well context matters. You know what I meant. Point: more throughput.
Call it what you want. If you know what indeed is the roadblock to
getting 3Gb/s out the door, plz do tell, I'm curious.

Then stop looking at the interface. It simply doesn't matter. The
interface has been faster than the platter for some time. It simply
doesn't matter.

It is to me, for recovery purposes. Quite important actually.

If you have to recover often enough for the install time to matter, you
had better be looking for the problem somewhere else! Why do you need to
recover that often? From scratch?! Yikes!

In the four-five years I've been using Win2K I've only had to reinstall
once (last week, in fact). It's a horribly time-consuming process (not
done yet), even if the installation of Win took zero time. There was a
time I did benchmarking (on NT4), and each test had to be "identical". I
did the install/setup once and then cloned drives from that "master". It
really didn't matter how long the install took.

 

[keith]

"Some extent" is easy to calculate when you add the context of, say, an 8MB
drive cache.

Which begs the question; why woud the data you want be in the drive's
cache at all (and not cached by the OS)?

 

[AJ]

On Thu, 30 Dec 2004 23:25:26 +0000, AJ wrote:


On Wed, 29 Dec 2004 09:14:03 +0000, AJ wrote:



AJ wrote:
I'll bet NCQ gets
over-marketed to stand-alone users though too (because the HD manf's have nothing
new to offer this year?). I was chomping-at-the-bit (hoping) for 3GB/s. I'd actually buy
a new HD and use my exiting 80GB SATA for backups if 3 GB/s was available (twice

[...]

Anywayz... I digress. I sure would like 3GB/s drives though, especially 2.5 inch ones!

It'll be a while before we see consumer-level drives with 3
Giga*byte*/sec interfaces. The 3G*bit*/sec interface (SATA-II) will
be here soon, but it's only one tenth of that (300MByte/s).

My bad. That's what I meant: 3Gb/s, twice the throughut of today's
SATA drives.

Find me a SATA drive with even 1.5Gb/s and I'll buy it. Hint: datarate
<> thoughput. If you can't get it to/from the platter, it doesn't matter.

Points to
consider: 1) drives cannot deliver data fast enough internally to flood
even a SATA-I interface, and 2) Besides, I doubt that we will see many
drives offering 3Gb/s. If I understand it correctly, the 3Gb/s is
intended for between the mobo and "expander chips", so that you can
attach multiple physical drives to a single SATA-II connector at the
mobo level.

I don't think so, cuz where I heard about it was at the drive
manufacturers sites where they describe SATA technology and they are
talking about individual drive specs.

Read the above again. It matters *not* how fast the interface is, if the
media can't keep up. ...and it cannot.

Perhaps that's the roadblock then at this time and why it's not available.
Read the above again: the poster said he thought the 3Gb/s spec was
a motherboard bus spec rather than a per-drive spec, to which I thought
not. The assumption made was that the 3Gb/s throughput would be
realized (duh, of course) and not be just a theoretical number. (Of course
if one couldn't wait for 3Gb/s SATA, one could do RAID and double their
throughput now).

You're still not getting it. The interface could be 300Gb/s and it
wouldn't matter. The data can't get to/from the platter that fast. The
spec you want to be looking for is the "sustained transfer rate" or STR.
The interface rate is meaningless.

Well context matters. You know what I meant. Point: more throughput.
Call it what you want. If you know what indeed is the roadblock to
getting 3Gb/s out the door, plz do tell, I'm curious.

Then stop looking at the interface. It simply doesn't matter. The
interface has been faster than the platter for some time. It simply
doesn't matter.

Well the implied (by the drive vendors) info is that there is a speed up
to be had in a future generation of SATA. No one is picking nits (I'm
not). Faster drives (more throughput) will be welcomed with open arms
(was the point). My guess is that SATA-I is close to the stated 1.5Gb/s
spec. My guess is too that vendors will be smarter than to advertise
3 Gb/s if it is only theoretical but not achievable because they can't really
get that much data to the interface in time. If a single drive has an
advertised 3 Gb/s interface on it, that implies that it can transfer data
at that rate (or thereabouts). Anything else begs for class action lawsuits.
So I'd say that the spec/rating indeed does matter and is significant because
it's an implied statement about a certain level of performance.

Anyone know if another platter density increase is coming soon and
will that help achieve 3 Gb/s? Will drives be moving away from magnetic
technology to cross that chasm?

If you have to recover often enough for the install time to matter, you
had better be looking for the problem somewhere else! Why do you need to
recover that often? From scratch?! Yikes!
In the four-five years I've been using Win2K I've only had to reinstall
once (last week, in fact). It's a horribly time-consuming process (not
done yet), even if the installation of Win took zero time. There was a
time I did benchmarking (on NT4), and each test had to be "identical". I
did the install/setup once and then cloned drives from that "master". It
really didn't matter how long the install took.

I'm forward thinking to when a consumer box could actually self-heal in
a short time. Currently it may take 20 mins (probably less) for the as-manufactured
configuration to be reinstalled. But then there's the onsite/user config and data also
(more work to be done in this area). The goal of course being a user-friendly or
hands off approach to fixing mucked-up systems. Doesn't apply to techies
like yourself.

AJ

 

[keith]

On Thu, 30 Dec 2004 23:25:26 +0000, AJ wrote:


On Wed, 29 Dec 2004 09:14:03 +0000, AJ wrote:



AJ wrote:
I'll bet NCQ gets
over-marketed to stand-alone users though too (because the HD manf's have nothing
new to offer this year?). I was chomping-at-the-bit (hoping) for 3GB/s. I'd actually buy
a new HD and use my exiting 80GB SATA for backups if 3 GB/s was available (twice

[...]

Anywayz... I digress. I sure would like 3GB/s drives though, especially 2.5 inch ones!

It'll be a while before we see consumer-level drives with 3
Giga*byte*/sec interfaces. The 3G*bit*/sec interface (SATA-II) will
be here soon, but it's only one tenth of that (300MByte/s).

My bad. That's what I meant: 3Gb/s, twice the throughut of today's
SATA drives.

Find me a SATA drive with even 1.5Gb/s and I'll buy it. Hint: datarate
<> thoughput. If you can't get it to/from the platter, it doesn't matter.

Points to
consider: 1) drives cannot deliver data fast enough internally to flood
even a SATA-I interface, and 2) Besides, I doubt that we will see many
drives offering 3Gb/s. If I understand it correctly, the 3Gb/s is
intended for between the mobo and "expander chips", so that you can
attach multiple physical drives to a single SATA-II connector at the
mobo level.

I don't think so, cuz where I heard about it was at the drive
manufacturers sites where they describe SATA technology and they are
talking about individual drive specs.

Read the above again. It matters *not* how fast the interface is, if the
media can't keep up. ...and it cannot.

Perhaps that's the roadblock then at this time and why it's not available.
Read the above again: the poster said he thought the 3Gb/s spec was
a motherboard bus spec rather than a per-drive spec, to which I thought
not. The assumption made was that the 3Gb/s throughput would be
realized (duh, of course) and not be just a theoretical number. (Of course
if one couldn't wait for 3Gb/s SATA, one could do RAID and double their
throughput now).

You're still not getting it. The interface could be 300Gb/s and it
wouldn't matter. The data can't get to/from the platter that fast. The
spec you want to be looking for is the "sustained transfer rate" or STR.
The interface rate is meaningless.

Well context matters. You know what I meant. Point: more throughput.
Call it what you want. If you know what indeed is the roadblock to
getting 3Gb/s out the door, plz do tell, I'm curious.

Then stop looking at the interface. It simply doesn't matter. The
interface has been faster than the platter for some time. It simply
doesn't matter.

Well the implied (by the drive vendors) info is that there is a speed up
to be had in a future generation of SATA. No one is picking nits (I'm
not). Faster drives (more throughput) will be welcomed with open arms
(was the point).

Stop looking at the press-releases then. Look at the specifications.

My guess is that SATA-I is close to the stated 1.5Gb/s
spec. My guess is too that vendors will be smarter than to advertise
3 Gb/s if it is only theoretical but not achievable because they can't really
get that much data to the interface in time.

You're would be guessing wrong. Perhaps you'd like to run the numbers. I
just pulled up the spec (you can easily drill for it on Maxtor's site) for
the Maxtor 7200RPM MaXLine Plus II 250GB/300GB drives. Now I just picked
these more or less at random, so there may be a faster drive out there in
the desktop space somewhere, but this is a relatively recent PATA/SATA
drive, with some decent specs. It's "Maximum STR is _up_to_ 59MB/s". That
"up to" part means that this is only possible on certain cylinders, and is
certainly not the average, or minimum. Last I checked 59MB/s is a ways
away from 1.5Gb/s.

If a single drive has an
advertised 3 Gb/s interface on it, that implies that it can transfer
data at that rate (or thereabouts). Anything else begs for class action
lawsuits.

Only from the uninformed. Become informed!

So I'd say that the spec/rating indeed does matter and is
significant because it's an implied statement about a certain level of
performance.

You have a *lot* to learn, grasshopper.

Anyone know if another platter density increase is coming soon and will
that help achieve 3 Gb/s? Will drives be moving away from magnetic
technology to cross that chasm?

See above, grasshopper.

I'm forward thinking to when a consumer box could actually self-heal in
a short time.

....and you're using Windows? ...and OE? Yikes, you really are green!

Currently it may take 20 mins (probably less) for the
as-manufactured configuration to be reinstalled. But then there's the
onsite/user config and data also (more work to be done in this area).
The goal of course being a user-friendly or hands off approach to fixing
mucked-up systems. Doesn't apply to techies like yourself.

Non-techies aren't re-installing anything, other than perhaps from
manufacturer's images. Besides, what difference does the OS make
considering all the time is gettign back all the applications and data?

 

[AJ]

On Sun, 02 Jan 2005 00:53:00 +0000, AJ wrote:


On Thu, 30 Dec 2004 23:25:26 +0000, AJ wrote:


On Wed, 29 Dec 2004 09:14:03 +0000, AJ wrote:



AJ wrote:
I'll bet NCQ gets
over-marketed to stand-alone users though too (because the HD manf's have nothing
new to offer this year?). I was chomping-at-the-bit (hoping) for 3GB/s. I'd actually buy
a new HD and use my exiting 80GB SATA for backups if 3 GB/s was available (twice

[...]

Anywayz... I digress. I sure would like 3GB/s drives though, especially 2.5 inch ones!

It'll be a while before we see consumer-level drives with 3
Giga*byte*/sec interfaces. The 3G*bit*/sec interface (SATA-II) will
be here soon, but it's only one tenth of that (300MByte/s).

My bad. That's what I meant: 3Gb/s, twice the throughut of today's
SATA drives.

Find me a SATA drive with even 1.5Gb/s and I'll buy it. Hint: datarate
<> thoughput. If you can't get it to/from the platter, it doesn't matter.

Points to
consider: 1) drives cannot deliver data fast enough internally to flood
even a SATA-I interface, and 2) Besides, I doubt that we will see many
drives offering 3Gb/s. If I understand it correctly, the 3Gb/s is
intended for between the mobo and "expander chips", so that you can
attach multiple physical drives to a single SATA-II connector at the
mobo level.

I don't think so, cuz where I heard about it was at the drive
manufacturers sites where they describe SATA technology and they are
talking about individual drive specs.

Read the above again. It matters *not* how fast the interface is, if the
media can't keep up. ...and it cannot.

Perhaps that's the roadblock then at this time and why it's not available.
Read the above again: the poster said he thought the 3Gb/s spec was
a motherboard bus spec rather than a per-drive spec, to which I thought
not. The assumption made was that the 3Gb/s throughput would be
realized (duh, of course) and not be just a theoretical number. (Of course
if one couldn't wait for 3Gb/s SATA, one could do RAID and double their
throughput now).

You're still not getting it. The interface could be 300Gb/s and it
wouldn't matter. The data can't get to/from the platter that fast. The
spec you want to be looking for is the "sustained transfer rate" or STR.
The interface rate is meaningless.

Well context matters. You know what I meant. Point: more throughput.
Call it what you want. If you know what indeed is the roadblock to
getting 3Gb/s out the door, plz do tell, I'm curious.

Then stop looking at the interface. It simply doesn't matter. The
interface has been faster than the platter for some time. It simply
doesn't matter.

Well the implied (by the drive vendors) info is that there is a speed up
to be had in a future generation of SATA. No one is picking nits (I'm
not). Faster drives (more throughput) will be welcomed with open arms
(was the point).

Stop looking at the press-releases then. Look at the specifications.

Again, I don't think the vendors would open themselves up to class action
as easily as you imply.

You're would be guessing wrong. Perhaps you'd like to run the numbers. I
just pulled up the spec (you can easily drill for it on Maxtor's site) for
the Maxtor 7200RPM MaXLine Plus II 250GB/300GB drives. Now I just picked
these more or less at random, so there may be a faster drive out there in
the desktop space somewhere, but this is a relatively recent PATA/SATA
drive, with some decent specs. It's "Maximum STR is _up_to_ 59MB/s". That
"up to" part means that this is only possible on certain cylinders, and is
certainly not the average, or minimum. Last I checked 59MB/s is a ways
away from 1.5Gb/s.

Actually, 59MB/s = 472Mb/s = 4.72 Gb/s.

My bad. But you should have known what I meant. (Though there's quite
a few ultra-techie types that have a hard time with contextual comprehension.
Are you one?).

By "advertised", I was thinking such as a prominent spec on the front of a
retail box with "3 Gb/s" (I don't care if it also said "interface" or "throughput"
or whatever. Performance is implied in that case and if it isn't there, it
wouldn't take much of a lawyer to argue "deception", "fraudulent" etc).


Aside: Does anyone know if another platter density increase is coming soon
and will that help achieve 3 Gb/s? Will drives be moving away from magnetic
technology to cross that chasm?

...and you're using Windows? ...and OE? Yikes, you really are green!

You have a point whippersnapper or are you in general an ad-hominem poster?

Non-techies aren't re-installing anything, other than perhaps from
manufacturer's images.

That was part of my scenario and why I said <20 mins (images apply faster
than an install takes).

Besides, what difference does the OS make
considering all the time is gettign back all the applications and data?

Reread my post. I mentioned that as part of the process and indicated
that's where more work needs yet to be done.

AJ

 

[Kai Harrekilde-Petersen]

Actually, 59MB/s = 472Mb/s = 4.72 Gb/s.

That's utterly plain wrong: 472Mb/s = 0.472Gb/s.

/PS: please bear in mind that the advertised 1.5Gb/s includes 25%
overhead for encoding (8B/10B). The maximum available bandwidth for
user-level data is 150MB/sec = 1200Mb/s.


Regards,

Kai

 

[AJ]

Actually, 59MB/s = 472Mb/s = 4.72 Gb/s.

LOL. My bad. "whippersnapper" is gonna jump all over this big time! LOL!

= .472Gb/s is obviously the correct answer. Which seems low and wouldn't
explain why SATA drives are faster than ATA100 drives.

AJ

 

[Never anonymous Bud]

/PS: please bear in mind that the advertised 1.5Gb/s includes 25%
overhead for encoding (8B/10B).

You're still using a modem with stop and start bits??

 

[cquirke (MVP Win9x)]

Which begs the question; why woud the data you want be in the drive's
cache at all (and not cached by the OS)?

I suspect it's like ye olde 1:1 interleave MFM controllers.

With varying numbers of sectors per track, the raw data rate off
platter is going to vary, and that has to be coupled with the
constant-speed bus to the rest of the PC. If raw data from disk has
to wait for this bus, then one might have to wait for an extra spin of
the disk, which is ungood.

So the cache may be to allow raw disk to cache transfer at whatever
the off-platter data rate happens to be; once done, the heads can
carry on with whatever is supposed to happen next, while the cache RAM
feeds the PCI (or newer coupling) bus at whatever speed is OK.

By the time the heads have reached the next position, chances are the
cache will be empty and the whole process can begin again.

Note; there's a lot of conjecture there, such as; is there enough time
during a head click to empty a buffer? Which is faster, HD unit to PC
transfer rate or disk platter to cache RAM transfer rate? And so on.

-------------------- ----- ---- --- -- - - - -

Tip Of The Day:
To disable the 'Tip of the Day' feature...

 

[AJ]

On Mon, 03 Jan 2005 03:07:35 +0000, AJ wrote:

Non-techies aren't re-installing anything, other than perhaps from
manufacturer's images. Besides, what difference does the OS make
considering all the time is gettign back all the applications and data?

Eye on ball: the issue was the importance of drive speed in a recovery
process. The actual process (OS install, data restore, apps etc) isn't
important to the point: that drivespeed (sustained throughput) is more
important than CPU performance in regards to recovery (a rather lengthy
process). When dealing with images, even moreso since it's all disk and
little CPU. OK now?

AJ

 

[AJ]

You're would be guessing wrong. Perhaps you'd like to run the numbers. I
just pulled up the spec (you can easily drill for it on Maxtor's site) for
the Maxtor 7200RPM MaXLine Plus II 250GB/300GB drives. Now I just picked
these more or less at random, so there may be a faster drive out there in
the desktop space somewhere, but this is a relatively recent PATA/SATA
drive, with some decent specs. It's "Maximum STR is _up_to_ 59MB/s". That
"up to" part means that this is only possible on certain cylinders, and is
certainly not the average, or minimum. Last I checked 59MB/s is a ways
away from 1.5Gb/s.

Indeed that's the story. A FAQ the Seagate website:

"Disc drive data rates have not exceeded ATA100 limits yet, so why should I switch
to SATA?
The maximum internal data rate on an IDE disc drive today is around 72 Mbytes/sec.
The ATA/100 data transfer rate has not been reached, but one of the reasons IDE
performance is where it is today is the expandable data path PATA has allowed.
That data path in PATA has reached its limit. SATA allows disc drives to continue
to offer performance and reliability at cost parity to PATA. In addition, SATA interface
requires less voltage, meaning better power consumption and management in both
desktop and mobile applications. The thinner cable allows for flexible designs and
improved airflow in smaller form factors."

(Whippersnapper: your attitude made me question the credibility of what you say/said,
so I had to go search the web for the above source for information that I was able to
believe.)

It wouldn't be a far stretch of the imagination to think that when 3Gb/s shows up
on a drive box that the technology inside will have changed to where the 1.5Gb/s
rate has been eclipsed. Therefor I'll still keep an eye out for drives with that
specification, as it may (and I think probably will) be an idicator of increased
performance.

AJ

 

[AJ]

That's utterly plain wrong: 472Mb/s = 0.472Gb/s.

I know. I quickly afterward posted a correction.

AJ