* Need opinions on high-volume SATA models *

[Eric Gisin]

Unless your interested in a file server or web server, no.

http://www.storagereview.com/php/benchmark/compare_rtg_2001.php?typeID=10&testbedID=3&osID=4&raidco
nfigID=1&numDrives=1&devID_0=271&devID_1=270&devCnt=2

You made this claim in early December, and it is still wrong.
Here is my response again:

Not sure about the hyperthreading part but go to storagereview.com and check
out their Performance Database. In the High-End Drivemark 2000, the one w/o
NCQ performs more I/O's per second than the one with it. Not sure if that's
really significant in real world usage.

There are two drives with NCQ reviewed, the MaxLine III and the Barracuda
7200.7. The MaxLine performs better with NCQ in all tests. The Barracuda
performs better without NCQ, but only for desktop benchmarks.

===

Storage Review really screwed up with their NCQ conclusions.
You have to use controlled test: one SATA2 controller,
with the NCQ feature switched on and off.

 

[Odie Ferrous]

Try harder, from WinHEC 2004:
www.seagate.com/cda/newsinfo/newsroom/releases/article/0,1121,2102,00.html

I've read this article. It appears the only testing they did was the
copy of a 4GB file to the drive.

The WinHEC public demo says Barracuda 7200.7 with NCQ beats Raptor.

A google of DiamondMax-10 or Barracuda-7200.8 with "NCQ benchmark" gives lots
of results:

http://www.tweakers.net/reviews/553/1&lp=nl_en

My Dutch is not really good enough to understand that 100% - but from
the little I can make out, they are painting a rosy scenario for Hitachi
drives which, if it is true, makes me sceptical about the validity of
any testing they are doing. But that's just me - the world's biggest
sceptic.

http://babelfish.altavista.com/babelfish/trurl_pagecontent?lp=nl_en&trurl=http://www.tweakers
.net%2freviews%2f553%2f7

This doesn't say much for the Barracuda at all. But maybe it's in the
translation?

http://techreport.com/reviews/2005q1/maxtor-diamondmax10/index.x?pg=1

Interesting; they make out the Maxtor is quicker than the Seagate, yet
Seagate make no mention of this. Moreover, they conclude that the
Raptor is quicker than the Maxtor. If that is the case, and Seagate are
claiming their drive is the quickest, something is a little flawed.

http://www.gamepc.com/labs/view_content.asp?id=ncq300&page=8&cookie_test=1

Even more interesting; this throws a damp squib on Seagate's claims.

None meet the WinHEC claim, but they are within 20% of The Raptor 73GB IO/s.

Still worth checking out. I will go out and get some hardware and do my
own testing. Trouble is, in the UK, it is all still very expensive -
nearly twice the price of the same kit in the US.

Thanks for the links.


Odie

 

[Chuck U. Farley]

As I said:

Please point out something other than _opinion_ about why the testing _may_
be inaccurate. Storage review is well known for their unbiased drive
testing.

Presumably they enabled queuing in the driver as no drive decides by itself
whether it will queue or not. The xxxxxx queued commands may impose a bigger
overhead on the protocol and allow less data per time unit to be

transferred.

Presuming is just like ASSuming.

And since queuing involves the use of different commands and a mechanism
to keep track of the order of commands (tags) and the data that belongs
to them, that too may bring about the difference you see in that review.

Again, more opinion about what _may_ be wrong with their testing.

 

[Chuck U. Farley]

Unless your interested in a file server or web server, no.

http://www.storagereview.com/php/benchmark/compare_rtg_2001.php?typeID=10&testbedID=3&osID=4&raidco
nfigID=1&numDrives=1&devID_0=271&devID_1=270&devCnt=2

You made this claim in early December, and it is still wrong.

Guess you missed the comment above so let me repeat it slowly so maybe you
can understand:

"Unless you're interested in a file
server or web server, no."

Here is my response again:



There are two drives with NCQ reviewed, the MaxLine III and the Barracuda
7200.7. The MaxLine performs better with NCQ in all tests. The Barracuda
performs better without NCQ, but only for desktop benchmarks.

Where in the OP's post was there _anything_ mentioned about _server_
performance? Look at the link and you'll see the Seagate w/o NCQ wins every
single _desktop_ benchmark compared to the same drive, with the same model
number. Just like the OP from that other thread, this OP has _not_ mentioned
anything whatsoever about _server_ configuation. Can you not follow a
thread?

===

Storage Review really screwed up with their NCQ conclusions.
You have to use controlled test: one SATA2 controller,
with the NCQ feature switched on and off.

If you want to debate their testing methodology, take it up with
storagereview.com. I'll leave it up to the OP whether he thinks the people
at storagereview.com are more beleivable... or some bit-twister on Usenet
who's always finding something wrong with something when it doesn't support
his argument.

 

[Odie Ferrous]

As I said:


Please point out something other than _opinion_ about why the testing _may_
be inaccurate. Storage review is well known for their unbiased drive
testing.

They don't necessarily get that opinion in the UK.

Besides, their site accepts advertising, which clearly paves the way for
biased testing.


Odie

 

[J. Clarke]

They don't necessarily get that opinion in the UK.

Besides, their site accepts advertising, which clearly paves the way for
biased testing.

Actually, I did not "suggest" that the testing was "innacurate". I stated
that they, Storagreview, has in earlier tests addressing the same issue
seen differences in performance that _they_, not _I_, attributed to
different firmware.

If Mr. Farley has a problem with that I suggest he take it up with
Storagereview and not with me.

The particular pages that were linked appear to be part of an article that
has not yet been released in its entirety--I could not find the text that
goes with them, so it is not possible to comment on the methodology.
Perhaps when the article is complete one can determine more.

As for bias due to advertising, I fail to see how that would lead to a
Seagate drive without NCQ looking more capable than one _with_ NCQ. Since
Seagate is trying to sell their drives on the basis of the performance
benefits of NCQ I would expect that sort of bias to go the other way.

Regardless of any of this, the difference in performance was minuscule.

 

[Folkert Rienstra]

http://www.storagereview.com/php/be...&numDrives=1&devID_0=271&devID_1=270&devCnt=2

Guess you missed the comment above so let me repeat it slowly so maybe you
can understand:

"Unless you're interested in a file
server or web server, no."


Where in the OP's post was there _anything_ mentioned about _server_
performance?

Look at the link and you'll see the Seagate w/o NCQ

What *"the"* Seagate w/o NCQ ?

wins every single _desktop_ benchmark compared to the same drive, with the
same model number.

And where exactly does it say that these are 2 different drives and not one
and the same drive? Since when do manufacturers make 2 different drives
under the same model number (at the same moment in time)?
And even if early 7200.7 were delivered without NCQ capable firmware,
what makes you think that they can still be bought/ordered?

And even if they were of different firmware you can't compare them as they
are different drives (physically) and the difference may come from that alone.
Drives aren't exact replicas of each other, each one differs slightly and so do
the benchmark results.

Just like the OP from that other thread, this OP has _not_ mentioned
anything whatsoever about _server_ configuation.

But specifically mentioned Command Queueing (NCQ) and SCSI which is associated
with Server type use. He also said something about storing multimedia content.

Can you not follow a thread?

Pot, Kettle, Dyslexic?

If you want to debate their testing methodology, take it up with
storagereview.com. I'll leave it up to the OP whether he thinks the people
at storagereview.com are more beleivable...

or some bit-twister on Usenet

Which obviously includes you.

 

[Folkert Rienstra]

As I said:


Please point out something other than _opinion_ about why the testing _may_
be inaccurate.

Storage review is well known for their unbiased drive testing.

Except that they don't solely test the drive, they also test the OS influence
with it. That has nothing to do with being opinionated.

Presuming is just like ASSuming.

Like your ASSumption that there are actually two different versions of the ST3160827AS ?

Again, more opinion about what _may_ be wrong with their testing.

Well, that is your opinion, not mine.

 

[Folkert Rienstra]

http://www.storagereview.com/php/be...&numDrives=1&devID_0=271&devID_1=270&devCnt=2


According to this, the non-NCQ drive actually performed better.

Actually, it is the same drive.

After wondering if it was worth hunting down, now I'm wondering if it's
worth going out of my way to avoid.
Nope.


Or did I read it wrong?

Yes.

 

[Folkert Rienstra]

Hi,

I'm putting together the following computer :

MB : Asus A8N-SLI Deluxe
CPU : AMD 3200+ 2GB
RAM : Corsair TWINX 2GB (2x1) Cas2
VID : Asus EN6600GT 128MB (TVO/SLI)
AUDIO : Echo Gina3G
HD 1 : WD Raptor 74GB SATA (10,000rpm)
HD 2 : Seagate Barracuda 200GB SATA (7,200rpm)
DVD: Plextor 716SA SATA
PSU: Antec 480W NeoPower
CASE : Antec P160

The Raptor will serve as the OS (C drive. The only question mark
is the 2nd drive, which will be used to store docs (music, video, etc.)

The network administrator where I work tells me Seagate Barracudas,
which he recommends as they are eerily silent, support NCQ
(N____ Command Queueing, whatever that is) only at 250GB and up.

200GB models and lower, he claims, do not.

That's false. The 200GB 7200.7 does not, the 200GB 7200.8 does.
Some lower capacity 7200.7 models also *do* support NCQ.

I don't know what NCQ is, but for the small price difference, I
originally opted for the 250GB... but the store has none in stock and
doesn't plan to have any for a bit.

They recommended the 200GB, which they have plenty of.

Plenty of what, 7200.7 or 7200.8 ?

First question... is NCQ (Seagate SATA 250GB+) worth hunting down and
finding?

Not for the size. All 7200.7 models ending at 7 support NCQ.

If the drive doesn't support it, you can't experiment with it either.
If it does, you can, *if* the OS supports it.
You can still decide to disable it if you don't want to use it.

Next... I'm told by a competing store that they no longer carry Seagate SATA
drives *AT ALL* because they have a 50% return-rate on them. Apparently they
keep breaking. Naturally, the store I'm ordering my system from - which DOES
carry them - says they're great.

So my second question is... should I be looking at Seagate at all? Or is
there a significant leader in the 250GB range among Maxtor, Quantum
or Western Digital?

SCSI is out of my price range.

But the Raptor was not?

 

[Eric Gisin]

http://www.storagereview.com/php/benchmark/compare_rtg_2001.php?typeID=10&testbedID=3&osID=4&raidco

Guess you missed the comment above so let me repeat it slowly so maybe you
can understand:

"Unless you're interested in a file
server or web server, no."

I guess you are still an idiot and now a troll.
Look at the storage review performance database,
and compare the four drives marked NCQ.
For the mentally challenged: http://www.storagereview.com/comparison.html

The MaxLine III performs better with NCQ in all tests.
The Barracuda 7200.7 performs better with NCQ on the desktop.

Since you are really slow ...

The MaxLine III performs better with NCQ in all tests.
The Barracuda 7200.7 performs better with NCQ on the desktop.

The MaxLine III performs better with NCQ in all tests.
The Barracuda 7200.7 performs better with NCQ on the desktop.

I have never seen benchmarks that show NCQ doesn't help.

 

[Peter]

For the mentally challenged: http://www.storagereview.com/comparison.html

The Barracuda 7200.7 performs better with NCQ on the desktop.

Weird interpretation of benchmark results.

 

[Chuck U. Farley]

I guess you are still an idiot and now a troll.

Look at the storage review performance database,
and compare the four drives marked NCQ.

We're not comparing _four_ drives, we are comparing the _same_ drive, the
Seagate the OP was inquiring about. One with NCQ enabled and the _same_
drive with NCQ disabled. The OP's question was about _Seagate_ drives with
NCQ, _not_ Maxtor.

For the mentally challenged: http://www.storagereview.com/comparison.html

The MaxLine III performs better with NCQ in all tests.

Which has nothing to do with the Seagate drive in question,.

The Barracuda 7200.7 performs better with NCQ on the desktop.

BZZZZZTTTT. Incorrect. The Seagate with NCQ _enabled_was _slower_ on desktop
benchmarks than the _same_ drive with NCQ _disabled_. Now on the _server_
benchmarks, it was a different story but the OP is _not_ running a server so
those benchmarks are irrelevant..

If you had bothered to go to the link I provided, you would have seen the
_same_ Seagate drive with NCQ both enabled _and_ disabled compared. It's not
my fault if you can't comprehend the most basic of charting and comparison.
Let's see if I can help cure your ignorance. Go here:

http://www.storagereview.com/php/be...&numDrives=1&devID_0=271&devID_1=270&devCnt=2

and you'll see the _same_ drive with NCQ enabled and disabled. Notice the
one on the left that represents the drive _w/o_ NCQ enabled? Notice all the
orange under that column next to the desktop benchmarks? Orange means it
won the benchmark.Notice how the numbers for the i/o's per second are
_larger_ on the left than the ones on the right (you know, the one that
represents NCQ _enabled_).

And guess what more i/o's per second translate into? It means it's _faster_
than the one with NCQ enabled. You know, can process _more_ information in a
_shorter_ period. Now I don't really want to take a chance and confuse you
more, although I doubt that's possible at this point, but notice how the Low
Level Suite numbers are _exactly_ the same? Notice how the model numbers are
the same? Guess what that means? It's the _exact_ same drive. The _only_
variable is whether NCQ is enabled or disabled.

So, my answer to the OP was correct, hunting down a Seagate drive with NCQ
enabled is _not_ worth hunting down. NCQ was not meant for desktop
applications, at least not this implementation of it.

Since you are really slow ...

The one who is slow on the uptake in this exchange is blatantly clear.

I have never seen benchmarks that show NCQ doesn't help.

It's painfully obvious you don't comprehend benchmark testing, or basic
chart reading either.

Hopefully, I cured your ignorance but somehow I doubt it... given your
inability to comprehend the most basic graphical representation of data.

 

[Folkert Rienstra]

Wotanidiot.
Trying to hide his earlier ignorance by making a 4k post about a simple slip of the tongue.

We're not comparing _four_ drives, we are comparing the _same_ drive,

Glad you finally caught on. It took you a while.

the Seagate the OP was inquiring about.
Not.

One with NCQ enabled and the _same_ drive with NCQ disabled.

You can't disable it, you stupid troll. You can make use of it or you don't.
The decision is made in the driver, the OS and application, not the drive.
All the drive has to do is support it.

The OP's question was about _Seagate_ drives with NCQ, _not_ Maxtor.

Nonsense, it was about Seagate *and* about NCQ.

Which has nothing to do with the Seagate drive in question,.

Which OP didn't ask about, so what.
Other (Seagate) drives may well perform similar to the Maxtor when
used with and without queueing. That 7200.7 is a surpassed series.

BZZZZZTTTT. Incorrect.

You know what he meant.

The Seagate with NCQ _enabled_
Clueless.

was _slower_ on desktop benchmarks than the _same_ drive

with NCQ _disabled_.
Clueless.

Now on the _server_ benchmarks, it was a different story but the OP is
_not_ running a server so those benchmarks are irrelevant..

If you had bothered to go to the link I provided, you would have seen the

_same_ Seagate drive with NCQ both enabled _and_ disabled compared.

Bare nonsense.
You obviously don't have a clue about IDE/SATA command queueing.

It's not my fault if you can't comprehend the most basic of charting and comparison.

Let's see if I can help cure your ignorance.

Rotflol.

[blatant posturing snipped]

Notice how the model numbers are the same? Guess what that means?

Only a few days ago you obviously thought they were different drives.
This immense show of posturing doesn't make that go away, troll.

It's the _exact_ same drive. The _only_ variable is whether

NCQ is enabled or disabled.
Clueless.


So, my answer to the OP was correct, hunting down a Seagate drive with NCQ
enabled is _not_ worth hunting down.

And that shows you clueless.

NCQ was not meant for desktop applications, at least not this implementation of it.

You obviously haven't got a single clue of what you are talking about.
Tagged queueing works in favor whenever multiple commands are issued
at the same time.
On busy servers is where you reap the most benefit but that doesn't mean it
doesn't work on desktops, especially with OSes that load several files in parallel
when starting applications.

The one who is slow on the uptake in this exchange is blatantly clear.

Well, at least you got that right.

It's painfully obvious you don't comprehend benchmark testing, or
basic chart reading either.

Hopefully, I cured your ignorance but somehow I doubt it...

Well, in that case you two can shake hands.

given your inability to comprehend the most basic graphical
representation of data.

The one thing that you cleared up is that you are an obvious troll.