SATA Laptops?

[Dan]

Archived from groups: comp.sys.laptops (More info?)

When are SATA laptops going to hit the mainstream?

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

Dan wrote:

> When are SATA laptops going to hit the mainstream?

What difference does it make? There is no benefit of any kind to be
obtained by using SATA in a laptop.

--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

The benefit of SATA 2.5" disks is mainly in 2 areas:
o New laptop SATA chipsets might offer RAID for them
o SATA 2.5" HD may support NCQ/TCQ - which re-orders (slow) seeking

The migration is eventually to twin 1.8" SATA drives for laptops,
allowing users to run them in RAID-1 (good) or RAID-0 (adventure).

NCQ/TCQ offers a 10-15% benefit in re-ordering seeks - but requires
support in the HD + Chipset + OperatingSystem. It's what SCSI does.
Electromechanical seeking is a major bottleneck in terms of speed.
--
Dorothy Bradbury

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

Dorothy Bradbury wrote:

> The benefit of SATA 2.5" disks is mainly in 2 areas:
> o New laptop SATA chipsets might offer RAID for them

IDE chipsets offer RAID and there are in fact laptops that have RAID onboard
now. So SATA gains you nothing there.

> o SATA 2.5" HD may support NCQ/TCQ - which re-orders (slow) seeking

I have yet to see NCQ/TCQ demonstrate any real-world benefits on single user
machines.

> The migration is eventually to twin 1.8" SATA drives for laptops,

No, the migration is, perhaps eventually to single 1.8" drives in laptops
allowing for smaller, lighter machines. If there was any demand for RAID
on laptops then the ones that have it would be more popular.

> allowing users to run them in RAID-1 (good) or RAID-0 (adventure).
>
> NCQ/TCQ offers a 10-15% benefit in re-ordering seeks - but requires
> support in the HD + Chipset + OperatingSystem. It's what SCSI does.
> Electromechanical seeking is a major bottleneck in terms of speed.

So please provide a reference to tests that show that there is any
difference in function that is perceptible to the user when NCQ is enabled
on a single-user machine with a single drive. Not benchmarks that show
that in some arcane ubernerdly sense there is a "performance improvement"
but something that the guy sitting at the machine running code other than
benchmarks actually sees.

This is an example of nerds seeking nerdware for nerdly reasons.

--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

J. Clarke wrote:
> Dorothy Bradbury wrote:
>
>> The benefit of SATA 2.5" disks is mainly in 2 areas:
>> o New laptop SATA chipsets might offer RAID for them
>
> IDE chipsets offer RAID and there are in fact laptops that have RAID
> onboard now. So SATA gains you nothing there.
>
>> o SATA 2.5" HD may support NCQ/TCQ - which re-orders (slow) seeking
>
> I have yet to see NCQ/TCQ demonstrate any real-world benefits on
> single user machines.
>
>> The migration is eventually to twin 1.8" SATA drives for laptops,
>
> No, the migration is, perhaps eventually to single 1.8" drives in
> laptops allowing for smaller, lighter machines. If there was any
> demand for RAID on laptops then the ones that have it would be more
> popular.
>
>> allowing users to run them in RAID-1 (good) or RAID-0 (adventure).
>>
>> NCQ/TCQ offers a 10-15% benefit in re-ordering seeks - but requires
>> support in the HD + Chipset + OperatingSystem. It's what SCSI does.
>> Electromechanical seeking is a major bottleneck in terms of speed.
>
> So please provide a reference to tests that show that there is any
> difference in function that is perceptible to the user when NCQ is
> enabled on a single-user machine with a single drive. Not benchmarks
> that show that in some arcane ubernerdly sense there is a
> "performance improvement" but something that the guy sitting at the
> machine running code other than benchmarks actually sees.
>
> This is an example of nerds seeking nerdware for nerdly reasons.

A 10K rpm IDE drive would be a good choice. If there were any. SATA
looks like a solution to cabling in smaller cases, maybe. Sort of like
PCI-Express: new tek looking for a driving force.

Q

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

Quaoar wrote:

> J. Clarke wrote:
>> Dorothy Bradbury wrote:
>>
>>> The benefit of SATA 2.5" disks is mainly in 2 areas:
>>> o New laptop SATA chipsets might offer RAID for them
>>
>> IDE chipsets offer RAID and there are in fact laptops that have RAID
>> onboard now. So SATA gains you nothing there.
>>
>>> o SATA 2.5" HD may support NCQ/TCQ - which re-orders (slow) seeking
>>
>> I have yet to see NCQ/TCQ demonstrate any real-world benefits on
>> single user machines.
>>
>>> The migration is eventually to twin 1.8" SATA drives for laptops,
>>
>> No, the migration is, perhaps eventually to single 1.8" drives in
>> laptops allowing for smaller, lighter machines. If there was any
>> demand for RAID on laptops then the ones that have it would be more
>> popular.
>>
>>> allowing users to run them in RAID-1 (good) or RAID-0 (adventure).
>>>
>>> NCQ/TCQ offers a 10-15% benefit in re-ordering seeks - but requires
>>> support in the HD + Chipset + OperatingSystem. It's what SCSI does.
>>> Electromechanical seeking is a major bottleneck in terms of speed.
>>
>> So please provide a reference to tests that show that there is any
>> difference in function that is perceptible to the user when NCQ is
>> enabled on a single-user machine with a single drive. Not benchmarks
>> that show that in some arcane ubernerdly sense there is a
>> "performance improvement" but something that the guy sitting at the
>> machine running code other than benchmarks actually sees.
>>
>> This is an example of nerds seeking nerdware for nerdly reasons.
>
> A 10K rpm IDE drive would be a good choice. If there were any. SATA
> looks like a solution to cabling in smaller cases, maybe. Sort of like
> PCI-Express: new tek looking for a driving force.

If there were any 10K laptop drives of _any_ kind and if they could be made
in the necessary form factor without excessive power consumption or heat
generation. In a laptop drive cabling is not really an issue. Seagate
does have some 2.5" 10K RPM SCSI drives but they're expressly _not_ made
for laptop use.

>
> Q

--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

On 4/1/2005 10:05 PM, J. Clarke wrote:
> Dan wrote:
>
>
>>When are SATA laptops going to hit the mainstream?
>
>
> What difference does it make? There is no benefit of any kind to be
> obtained by using SATA in a laptop.
>

I thought I have read that SATA drives have faster transfer rate
compared to IDE. No?

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

dxyzc@nospam.com wrote:

> On 4/1/2005 10:05 PM, J. Clarke wrote:
>> Dan wrote:
>>
>>
>>>When are SATA laptops going to hit the mainstream?
>>
>>
>> What difference does it make? There is no benefit of any kind to be
>> obtained by using SATA in a laptop.
>>
>
> I thought I have read that SATA drives have faster transfer rate
> compared to IDE. No?

No. No drive currently in existence can sustain a transfer rate greater
than 100 Mb/sec. Anything above ATA/100 with a single drive is pure hype.
The limit is now, as it has been for a very long time, the rate at which
the mechanical parts can move bits past the head.

--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

what about those 15k scsi drives surely they come close, don't they?
"J. Clarke" <jclarke.usenet@snet.net.invalid> wrote in message
news:d2o4ds2tao@news4.newsguy.com...
> dxyzc@nospam.com wrote:
>
>> On 4/1/2005 10:05 PM, J. Clarke wrote:
>>> Dan wrote:
>>>
>>>
>>>>When are SATA laptops going to hit the mainstream?
>>>
>>>
>>> What difference does it make? There is no benefit of any kind to be
>>> obtained by using SATA in a laptop.
>>>
>>
>> I thought I have read that SATA drives have faster transfer rate
>> compared to IDE. No?
>
> No. No drive currently in existence can sustain a transfer rate greater
> than 100 Mb/sec. Anything above ATA/100 with a single drive is pure hype.
> The limit is now, as it has been for a very long time, the rate at which
> the mechanical parts can move bits past the head.
>
> --
> --John
> to email, dial "usenet" and validate
> (was jclarke at eye bee em dot net)

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

Veritech wrote:

> what about those 15k scsi drives surely they come close, don't they?

The latest Cheetah 15k4 drives are rated at "up to 96" which means that they
can do that for sequential transfers in the outermost zone under ideal
conditions.

> "J. Clarke" <jclarke.usenet@snet.net.invalid> wrote in message
> news:d2o4ds2tao@news4.newsguy.com...
>> dxyzc@nospam.com wrote:
>>
>>> On 4/1/2005 10:05 PM, J. Clarke wrote:
>>>> Dan wrote:
>>>>
>>>>
>>>>>When are SATA laptops going to hit the mainstream?
>>>>
>>>>
>>>> What difference does it make? There is no benefit of any kind to be
>>>> obtained by using SATA in a laptop.
>>>>
>>>
>>> I thought I have read that SATA drives have faster transfer rate
>>> compared to IDE. No?
>>
>> No. No drive currently in existence can sustain a transfer rate greater
>> than 100 Mb/sec. Anything above ATA/100 with a single drive is pure
>> hype. The limit is now, as it has been for a very long time, the rate at
>> which the mechanical parts can move bits past the head.
>>
>> --
>> --John
>> to email, dial "usenet" and validate
>> (was jclarke at eye bee em dot net)

--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

> So please provide a reference to tests that show that there is any
> difference in function that is perceptible to the user when NCQ is
> enabled on a single-user machine with a single drive.

That performed by an economic buyer who requires such:
o Developers
o Product Demonstrators
o Emergency Server use

Laptops are not necessarily single-user / single-thread.
o SOHO users
---- no gain - reduced electromechanical latency N/A or insignificant
o Developers, Product Demos, Server users
---- some gain - reduced electromechanical latency of some use

Laptops can deliver as Demo-Server or Emergency-Server:
o Dual-CPU + Dual-HD laptop + 64-bit PCI + Gigabit + Solaris/Windows
o dual Savvio 10k SCSI is 1) expensive, 2) thick HD, 3) hot running, 4) high IOPS
o dual NCQ SATA is 1) 7x cheaper, 2) thinner HD, 3) cool, 4) low IOPS

Such laptops are at 2000-3000$, not exactly out of power user range.

Interesting application benchmark
o Laptop -- Single-User Laptop + Twin NCQ HDs
o Application -- Multi IE windows loading + Email Outlook Search + A/V Scan running
o Assessment -- Rank time benefit based on amount of day spent doing activity (task WBS)
o Most users -- at tail-end of the distribution vs other options - diminishing returns

Many more may get a marginal benefit from a higher-rpm laptop HD -- which costs.
Any benefit from NCQ on a laptop HD may be insignificant -- but doesn't cost.

Outside of blade servers its benefit may be easy on benchmarks, hard on reality.


>> The migration is eventually to twin 1.8" SATA drives for laptops,
>
> No, the migration is, perhaps eventually to single 1.8" drives in laptops
> allowing for smaller, lighter machines. If there was any demand for RAID
> on laptops then the ones that have it would be more popular.

You are arguing Product-Mix:
o Power orientated buyers -- twin 1.8" drives where formly one 2.5" drive
---- that need not be in RAID
---- I've seen enough China Flotherm models to know it IS planned
o Weight orientated buyers -- single 1.8" drives saving mainly space
---- 2.5" to 1.8" offers only a small weight saving
o Ruggedness -- lighter drives transfer less shock into boards & vice-versa
---- small factor, but one which avoids using more costly flexible cable links

So I would not poo-poo NCQ SATA on a laptop.
There is a theoretical benefit - and potentially a real one.

Some users spend lots of time doing I/O with multiple-threads, and may link
into a network - share Prj2003 files, Exchange - effectively pushing to multi-user.
A single bod sat at a laptop is not necessarily "Single-Sser SOHO Word App".
--
Dorothy Bradbury

 

[Guest]

Archived from groups: comp.sys.laptops (More info?)

Dorothy Bradbury wrote:

Below a see a lot of bullshit and no facts. Sorry, but if that's your pitch
you blew the sale and don't let the door hit you in the ass on the way out.

>> So please provide a reference to tests that show that there is any
>> difference in function that is perceptible to the user when NCQ is
>> enabled on a single-user machine with a single drive.
>
> That performed by an economic buyer who requires such:
> o Developers
> o Product Demonstrators
> o Emergency Server use
>
> Laptops are not necessarily single-user / single-thread.
> o SOHO users
> ---- no gain - reduced electromechanical latency N/A or insignificant
> o Developers, Product Demos, Server users
> ---- some gain - reduced electromechanical latency of some use
>
> Laptops can deliver as Demo-Server or Emergency-Server:
> o Dual-CPU + Dual-HD laptop + 64-bit PCI + Gigabit + Solaris/Windows
> o dual Savvio 10k SCSI is 1) expensive, 2) thick HD, 3) hot running, 4)
> high IOPS o dual NCQ SATA is 1) 7x cheaper, 2) thinner HD, 3) cool, 4) low
> IOPS
>
> Such laptops are at 2000-3000$, not exactly out of power user range.
>
> Interesting application benchmark
> o Laptop -- Single-User Laptop + Twin NCQ HDs
> o Application -- Multi IE windows loading + Email Outlook Search + A/V
> Scan running o Assessment -- Rank time benefit based on amount of day
> spent doing activity (task WBS) o Most users -- at tail-end of the
> distribution vs other options - diminishing returns
>
> Many more may get a marginal benefit from a higher-rpm laptop HD -- which
> costs. Any benefit from NCQ on a laptop HD may be insignificant -- but
> doesn't cost.
>
> Outside of blade servers its benefit may be easy on benchmarks, hard on
> reality.
>
>
>>> The migration is eventually to twin 1.8" SATA drives for laptops,
>>
>> No, the migration is, perhaps eventually to single 1.8" drives in laptops
>> allowing for smaller, lighter machines. If there was any demand for RAID
>> on laptops then the ones that have it would be more popular.
>
> You are arguing Product-Mix:
> o Power orientated buyers -- twin 1.8" drives where formly one 2.5" drive
> ---- that need not be in RAID
> ---- I've seen enough China Flotherm models to know it IS planned
> o Weight orientated buyers -- single 1.8" drives saving mainly space
> ---- 2.5" to 1.8" offers only a small weight saving
> o Ruggedness -- lighter drives transfer less shock into boards &
> vice-versa ---- small factor, but one which avoids using more costly
> flexible cable links
>
> So I would not poo-poo NCQ SATA on a laptop.
> There is a theoretical benefit - and potentially a real one.
>
> Some users spend lots of time doing I/O with multiple-threads, and may
> link into a network - share Prj2003 files, Exchange - effectively pushing
> to multi-user. A single bod sat at a laptop is not necessarily
> "Single-Sser SOHO Word App".

--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)