10000rpm or SATA-II?

[Curious George]

Because it's true.

It's also disk technology independent. Its connection to the topic is
tenuous at best.

What's the point of spreading half-truths?

Then why do you do it so often?

It's been proven true by many people, and if you had
bothered to read up on tech or test it yourself you'd have
also seen this.

You're talking about nothing more than using a cheap desktop (intended
for extremely light use & low performance reqs) instead of a true
"workstation" or "server" for a "workstation" or "server" workload.
Has little to do with the topic or technology being discussed. It's
trivial because the solution is so commonplace & it doesn't even come
up for a lot of users.

This is completely disk technology independent. It relates to
integration, scaling, PCI & certain controller limitations not disk
technology _generally_. Its connection to the topic is indeed tenuous
at best.

Let's put it another way...

Why SATA or SATA II?
The increase in transfer rate from the faster bus, right?

But you're totally wrong. SATA II is usually 1.5 Gb/s just like SATA
1. Or rather SATAII does not _require_ 3 Gb/sec. Many SATAII devices
currently on the market are indeed 1.5 Gb/sec.

Get a clue and stay on topic for once for Pete's sake.

http://www.sata-io.org/namingguidelines.asp

 

[Pelysma]

I understand that, but I can't decide which would be better to get. Which
would be better for me considering it is for my home computer or is there
not much difference overall?

The way I understand it, the faster access time would be better if I am
regularly accessing the drive but only for small amounts of data and that
a higher transfer rate would be better if I were accessing the drive for
long periods of time for large files? However, I have no idea what the
difference is for these for either a 10000rpm drive or SATA-II drive(ATA
300).

As a home user you will rarely even notice the faster access times: slow
ones are the blink of an eye in human terms. Probably the only exception is
when you are loading a complex application such as a game, which has a very
large number of small files to read into memory from the disk. If the game
must access the disk in mid-action for a number of files, this could lead to
some lagging, but again the difference between fast and really fast isn't
much. These things do matter in a many-user business environment, which is
why people pay big bucks to get it all.

You will, however, notice faster or slower transfer rates when you save a
large file, so if you have to choose, I'd choose faster transfer rates over
faster access times.

My humble opinion.

As to your original question, whether 10,000 RPM or better interface
technology will get you more performance, I have no experience with that.

 

[Curious George]

Why SATA or SATA II?
The increase in transfer rate from the faster bus, right?

Wrong.

Not only does SATA II not mean 3Gb/sec interface but you're using
"transfer rate" cluelessly. In these last posts you're interchanging
the concepts of "databurstrate" from drive cache, with bus
"burstrate", with drive "STR".

"Databurstrate" from drive cache & "STR" are both disk specific and
are basically independent of interface. "Burst rate" is clockrate of
the bus times number of bytes per clock cycle. It's a fixed number
that does more to affect "overhead" than "transfer rate." But that's
not a significant issue when there is such a huge disparity between
device & interface rates. You can't just throw all these concepts in
a pot, stir them together to get everyone confused, and then claim
that a drive on a 3Gb/sec interface will transfer data twice as fast
as on a 1.5Gb/sec pipe. Or that 3GB/sec interfaces will overwhelm
32/33 pci with half the # connections as 1.5Gb/sec. It doesn't work
that way. The disks basically call the shots.

 

[kony]

Wrong.

Not only does SATA II not mean 3Gb/sec interface but you're using
"transfer rate" cluelessly. In these last posts you're interchanging
the concepts of "databurstrate" from drive cache, with bus
"burstrate", with drive "STR".

Actually no, you are now first to suggest the
interchangeability of the two.

"Databurstrate" from drive cache & "STR" are both disk specific and
are basically independent of interface.

WRONG. The drive's STR potential is, in practice, also
dependant on the STR capability of the bus it's connected
to.

"Burst rate" is clockrate of
the bus times number of bytes per clock cycle.

Actually, it's what the drive can transfer, not the upper
ceiling on what it might transfer if ignoring any further
details of the transfer.

For example, consider any drive ever made. None will
achieve this theoretical "burst rate" through your simple
calculation. Ignoring all but two maximal limitors then
ignoring everything else will not be useful. The drives
you're using RIGHT NOW, do not achieve their theoretical
burst rate.

It's a fixed number
that does more to affect "overhead" than "transfer rate." But that's
not a significant issue when there is such a huge disparity between
device & interface rates.

I don't know how to break this to you, but this simplistic
description of burst rate is not news. It's laughable that
you use it to try to argue about a topic where your lack of
further understanding is the limitation.

You can't just throw all these concepts in
a pot, stir them together to get everyone confused, and then claim
that a drive on a 3Gb/sec interface will transfer data twice as fast
as on a 1.5Gb/sec pipe. Or that 3GB/sec interfaces will overwhelm
32/33 pci with half the # connections as 1.5Gb/sec. It doesn't work
that way. The disks basically call the shots.

The disk is ONE of many potential limits in transfer rate.
Often it is the slowest part, often it isn't. "yes and no".
Get some hardware and test it yourself, you're wasting our
time so long as you don't ever bother to test your theories.

 

[kony]

You're talking about nothing more than using a cheap desktop (intended
for extremely light use & low performance reqs) instead of a true
"workstation" or "server" for a "workstation" or "server" workload.

Please quote where the OP suggested anything other than a
PC. Regardless, the fact that it does apply on a PC does in
itself make my statement true, that it's NOT only the disk,
as I wrote "yes and no".

But you're totally wrong. SATA II is usually 1.5 Gb/s just like SATA
1. Or rather SATAII does not _require_ 3 Gb/sec. Many SATAII devices
currently on the market are indeed 1.5 Gb/sec.

Realize that the theoretical bus speed is _only_ theoretical
and additionally most certainly bound by the slowest bus
used in it's logical connection to the rest of the system.

Get a clue and stay on topic for once for Pete's sake.

http://www.sata-io.org/namingguidelines.asp

Funny, I might've written same thing.
I suggest you stop pretending to know and do a bit of
testing. Quite simply, it's a proven fact that the bus a
drive is connected to, and the bus THAT bus is connected to,
does have the potential to (and can be observed to without
particularly difficult or unique system configurations)
further degrade performance below the already-derated
theoretical performance of the drive itself.

Stop blabbering about it and start doing some testing.
Otherwise it looks foolish when there are articles a plenty
on the net that observe the very same things I have in
actual testing.

 

[Curious George]

Please quote where the OP suggested anything other than a
PC. Regardless, the fact that it does apply on a PC does in
itself make my statement true, that it's NOT only the disk,
as I wrote "yes and no".

Wow! Are you crazy? stupid? Or intentionally distorting the
discussion?

Realize that the theoretical bus speed is _only_ theoretical
and additionally most certainly bound by the slowest bus
used in it's logical connection to the rest of the system.

Really it's maximal - not realized to the end user maximal, but
maximal nonetheless. But this lexical distinction doesn't save you
from your goof.

Funny, I might've written same thing.

No you wouldn't because you thought SATAII was faster.

I suggest you stop pretending to know and do a bit of
testing. Quite simply, it's a proven fact that the bus a
drive is connected to, and the bus THAT bus is connected to,
does have the potential to (and can be observed to without
particularly difficult or unique system configurations)
further degrade performance below the already-derated
theoretical performance of the drive itself.

Stop blabbering about it and start doing some testing.
Otherwise it looks foolish when there are articles a plenty
on the net that observe the very same things I have in
actual testing.

Like your thorough testing & detailed raw data? Sorry but claimind to
test but misunderstand the HW & workloads won't get you anywhere.

 

[Curious George]

Actually no, you are now first to suggest the
interchangeability of the two.

Because you're not acknowledging your goofs. Yes I'm the first to
catch them

WRONG. The drive's STR potential is, in practice, also
dependant on the STR capability of the bus it's connected
to.

So the disk spins slower when PCI is congested? ROFLOL

Learn your terms.

Actually, it's what the drive can transfer, not the upper
ceiling on what it might transfer if ignoring any further
details of the transfer.

For example, consider any drive ever made. None will
achieve this theoretical "burst rate" through your simple
calculation.

There is no calculation - only definitions you continue to
neglect/misunderstand

Ignoring all but two maximal limitors then
ignoring everything else will not be useful. The drives
you're using RIGHT NOW, do not achieve their theoretical
burst rate.



I don't know how to break this to you, but this simplistic
description of burst rate is not news. It's laughable that
you use it to try to argue about a topic where your lack of
further understanding is the limitation.

Right. So you agree I'm right.
I don't know how to break it to you but that doesn't go very far to
invalidate my argument.

The disk is ONE of many potential limits in transfer rate.
Often it is the slowest part, often it isn't. "yes and no".
Get some hardware and test it yourself, you're wasting our
time so long as you don't ever bother to test your theories.

Funny talk about testing but you don't know what happens when you run
an ATA-133 drive in ATA-100 mode or U320 in U160 mode.

I hate to break it to you but if say the sound is hickuping when you
have the sound card, a disk & nic going - it's not bc you're
saturating the PCI bus. It's probably a driver or other issue.


Hey, keep shiiting in circles. It's fun to watch you step in it over
& over & over.

 

[kony]

Wow! Are you crazy? stupid? Or intentionally distorting the
discussion?

Most people LEARN as they go along.
You seem to go straight from "I know it all" to a complete
departure from reality.

Really it's maximal - not realized to the end user maximal, but
maximal nonetheless. But this lexical distinction doesn't save you
from your goof.

Whatever, my goof was actually replying to you at all
instead of just stating the facts in a separate post to the
OP.

No you wouldn't because you thought SATAII was faster.

Funny, I thought I'd posted that the bus it was connected to
limits both of them when it's PCI.

Tell me something, are you off your meds again?

Like your thorough testing & detailed raw data? Sorry but claimind to
test but misunderstand the HW & workloads won't get you anywhere.

Frankly I'm happy to leave you entangled in your own
delusions. In the meantime, others may realize that
connecting SATA I or II to a PCI bus, means there is a real
potential that other heavy PCI bus traffic will make that
bus the bottleneck, not the drives. It does't really matter
if you agree since the bus has no feelings on the matter,
cares not what you're thinking.

 

[kony]

So the disk spins slower when PCI is congested? ROFLOL

You are the first person to mention "spins".
I really do think you have some need for medication.

Learn your terms.


There is no calculation - only definitions you continue to
neglect/misunderstand

This is somewhat interesting, you seem totally disconnected
from the post I made and just inventing things as you go
along.

The interesting part of it is that you seem to actually
think you're doing something productive, either that or you
have no concern at all about whether your time is wasted or
not.

Right. So you agree I'm right.

I agree that you have a crude grasp of what "burst rate" is.

I don't know how to break it to you but that doesn't go very far to
invalidate my argument.

Your arguement seems to hinge around ignoring the slowest
link in the data transfer process.

Funny talk about testing but you don't know what happens when you run
an ATA-133 drive in ATA-100 mode or U320 in U160 mode.

I can't help but smile... you're really off your rocker.

I hate to break it to you but if say the sound is hickuping when you
have the sound card, a disk & nic going - it's not bc you're
saturating the PCI bus. It's probably a driver or other issue.

It "can" be another issue, like latency. It can also be bus
saturation. 133MB/s isn't very much these days,
particularly with a drive bursting in ATA133 mode and having
sustained transfer abilities beyond 70MB/s. It's no wonder
there is saturation when you add Gigabit networking (now
common on even mid-grade new motherboards) and sound.

Hey, keep shiiting in circles. It's fun to watch you step in it over
& over & over.

Then we're both amused. Glad to brighten up your day.

 

[Curious George]

ROFLOL!

Aren't you the one that likes to go on & on about "theories that don't
pan out in practice" & "minor issues" that have "limited impact on the
average user" ?

Flooding the thread with bs won't cover up your mistakes. Everyone
lost interest a few posts ago with your original errors staring them
in the face. Rehashing your errors, overstatement, & drug crazed
fantasy won't get you anywhere.

But if your ego really needs to post another 50 incoherent pages bc
you think having the last word=being right- knock yourself out.

 

[kony]

ROFLOL!

Aren't you the one that likes to go on & on about "theories that don't
pan out in practice" & "minor issues" that have "limited impact on the
average user" ?

Flooding the thread with bs won't cover up your mistakes. Everyone
lost interest a few posts ago with your original errors staring them
in the face. Rehashing your errors, overstatement, & drug crazed
fantasy won't get you anywhere.

But if your ego really needs to post another 50 incoherent pages bc
you think having the last word=being right- knock yourself out.

Curious George, you don't seem so amused anymore.
Don't you realize that after you started making things up to
suit your arguements, that lost credibility made it
pointless to try and backtrack?

If (as you wrote) you really wanted to talk about errors,
you would've just stuck to them but instead tried to divert
attention.

The original error was quite clear- you can't grasp that
it's quite possible to have a drive bottlenecked by a
controller sitting on the PCI bus. You've tried to
side-step and spew BS to avoid this central issue but you're
only fooling yourself- it is a fact and reproducible by
anyone who cares to try (or accidentally realizes it through
routine uses of a system). In a very basic isolated
benchmark stressing the drive, it may not be as evident or
apparent but most certainly is obvious with 2 or more drives
and/or during uses of a system where there is quite high
possibility other PCI devices will be contending for bus
time and/or bandwidth.

Since you one tried to be "Mr. SCSI Raid Expert", one would
think you would've already observed this, since a RAID0 or
better of modern drives will exceed the PCI bus potential
and result in a bottleneck, throughput below the potential
of the drives. I already suggested you actually TRY these
things you go on and on about but you seem quite
"hands-off", as if you don't even have the hardware to do
so.

 

[Fleabus]

Are there any 7200rpm drives on the market which can fully utilize SATA?

Hi:
At least two (the 80 and 250GB models) of the Hitachi DeskStar
SATA II 7200 RPM 8MB cache drives are shipping.
There is a 500GB SATA II 7200 RPM 16MB cache Hitachi DeskStar that
should appear soon.

http://www.hitachigst.com/portal/site/en/menuitem.8f07a3c3d3a7a12d92b86b31bac4f0a0/

And yes I know about the past IBM/DeathStar horror story. I don't
think Hitachi would risk its good name with a repeat performance.
I would think that these and other models from this company will be
the most exhaustively tested drives around because of that history.

BTW, Hitachi includes the software to "switch on" SATA II on the HDD
and you would need a chipset or controller card that supports SATA II.

Happy trails,

 

[Frobinrobin]

In answer to the original post, I would opt for 74GB Raptors (if your
budget allows it) stay away from the 36GB disks because they are slower
than the 74GB (apparently no TCQ on the 36GB).

I am basing my recomendations on personal usage of both the 2x80GB
sataii and the 2x36GB Raptors in a RAID 0 config <although Im
recomending the 74GB versions>
For starters, Windows starts quicker on the Raptors, disk use is
noticebly faster but they are slightly noisier and a little hotter (not
an issue for me)

If you are going for a single disk, again the Raptors kick butt - the
hitachi's sataii are just plain poor in a single config.. here are some
benches from hdtach: (from memory - if it matters I can get the actuals
for you - I'm just later for work :¬)

Sataii 80gb Hitachi (1 disk)
average read: 48mb/s
burst speed: 133mb/s

36GB Raptor (1 disk)
average read: 51mb/s
burst speed: honestly cant remember


-----
Sataii 80 Hitachi (2 disks in RAID 0)
average read: 97mb/s
burst speed: honestly cant remember

36GB Raptor (2 disks in RAID 0)
average read: 73mb/s
burst speed: 178mb/s

Now even though those results may look in favour of the sataii - but
with their 7200rpm they just perform poorly.
I was sorely unimpressed with the actual performance (as well as the
benches) of the sataii drive - I was expecting at least 60mb/s with
NCQ!

Stay with the Raptors.
Buy two of whatever disks you decide on - put them in RAID 0 for the
best performance ever.
However, if you're on a budget - opt for the sataii - I paid £40 a
disk for the hitachi's which is cheaper that most standard sata150
drives. the raptors cost me £60 each for the 36gb versions, those 74gb
raptors retail at £120.