Sign in with
Sign up | Sign in
Your question

DTR vs Desktop Performance?

Last response: in Laptops & Notebooks
Share
Anonymous
a b D Laptop
September 30, 2004 3:40:56 AM

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

I am considering a desktop replacement such as a Sager 4790 as an
alternative to updating a desktop system. I am concerned about
performance. I use my system to edit digital photos, and that uses a
lot of CPU, memory, and hard disk. I want to upgrade my existing
desktop because performance is limiting (it is a 1 GHz P3 with 512MB
RAM).

Does anyone have benchmark or other measurements that will indicate how
DTR performance will compare to a desktop with the same CPU, memory,
and similar speed (7200 RPM) hard drive?

Thanks.

Guy
Anonymous
a b D Laptop
September 30, 2004 9:54:37 AM

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

"Guy Scharf" <guy@spamcop.net> wrote in message
news:Xns9573DC967483guyspamcopnet@216.196.97.142...
> I am considering a desktop replacement such as a Sager 4790 as an
> alternative to updating a desktop system. I am concerned about
> performance. I use my system to edit digital photos, and that uses a
> lot of CPU, memory, and hard disk. I want to upgrade my existing
> desktop because performance is limiting (it is a 1 GHz P3 with 512MB
> RAM).
>
> Does anyone have benchmark or other measurements that will indicate how
> DTR performance will compare to a desktop with the same CPU, memory,
> and similar speed (7200 RPM) hard drive?

You wont get desktop performance especially for photoediting and HD DTR.
September 30, 2004 10:01:37 AM

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

Guy Scharf <guy@spamcop.net> wrote:
: I am considering a desktop replacement such as a Sager 4790 as an
: alternative to updating a desktop system. I am concerned about
: performance. I use my system to edit digital photos, and that uses a
: lot of CPU, memory, and hard disk. I want to upgrade my existing
: desktop because performance is limiting (it is a 1 GHz P3 with 512MB
: RAM).

: Does anyone have benchmark or other measurements that will indicate how
: DTR performance will compare to a desktop with the same CPU, memory,
: and similar speed (7200 RPM) hard drive?

After getting my Pentium 4 1.8GHZ desktop a few years ago, I got a
2.0GHZ Celeron laptop, both with 512MB of RAM. I was surprised that
the laptop was actually slightly faster on the few tasks I measured.
So I don't think you are looking at automatically slower performance
on a laptop. Just compare the CPU, memory, and hard drive speed as you
seem to be doing.

Andrew
--
----> Portland, Oregon, USA <----
*******************************************************************
----> http://www.bizave.com <---- Photo Albums and Portland Info
----> To Email me remove "MYSHOES" from email address
*******************************************************************
Related resources
Anonymous
a b D Laptop
September 30, 2004 12:22:22 PM

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

"Andrew" <usenetMYSHOES@bizaveMYSHOES.com> wrote in message
news:zcasdufyfyrkr2724654609@bizaveMYSHOES.com...
> Guy Scharf <guy@spamcop.net> wrote:
> : I am considering a desktop replacement such as a Sager 4790 as an
> : alternative to updating a desktop system. I am concerned about
> : performance. I use my system to edit digital photos, and that uses a
> : lot of CPU, memory, and hard disk. I want to upgrade my existing
> : desktop because performance is limiting (it is a 1 GHz P3 with 512MB
> : RAM).
>
> : Does anyone have benchmark or other measurements that will indicate how
> : DTR performance will compare to a desktop with the same CPU, memory,
> : and similar speed (7200 RPM) hard drive?
>
> After getting my Pentium 4 1.8GHZ desktop a few years ago, I got a
> 2.0GHZ Celeron laptop, both with 512MB of RAM. I was surprised that
> the laptop was actually slightly faster on the few tasks I measured.
> So I don't think you are looking at automatically slower performance
> on a laptop. Just compare the CPU, memory, and hard drive speed as you
> seem to be doing.

Right, but there's no laptop HD that can keep up with a late model 7200 or
10K RPM desktop HD and some laptops can't run the CPU at full speed
continuously. The heat causes throttling.
Anonymous
a b D Laptop
September 30, 2004 2:48:56 PM

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

Guy Scharf wrote:
>
> I am considering a desktop replacement such as a Sager 4790 as an
> alternative to updating a desktop system. I am concerned about
> performance. I use my system to edit digital photos, and that uses a
> lot of CPU, memory, and hard disk. I want to upgrade my existing
> desktop because performance is limiting (it is a 1 GHz P3 with 512MB
> RAM).
>
> Does anyone have benchmark or other measurements that will indicate how
> DTR performance will compare to a desktop with the same CPU, memory,
> and similar speed (7200 RPM) hard drive?
>

I have a fast DTR laptop, Athlon 64 3000+, but it's got a 4200 RPM drive
in it. In RAW conversion speed, and in PS batch job speed, there is
simply no comparison to my desktop system (Athlon 64 3200+, RAID 0+1).
The desktop RAW conversion times are like half of what they are on the
laptop.

In RAW conversion, I've found two bottlenecks. The first, and most
common is, not unexpectedly, CPU. But that's straightforward, if not
cheap, to fix. Once that's dealt with, you hit the bottleneck of disk
throughput. This is slightly harder to fix. While RAID 0 turns out not
to offer much, if any, real advantage in "everyday use", it most
assuredly DOES offer benefit to those of us who regularly manipulate
many large files. At least, as long as the RAID is implemented
properly, which means always a channel for each drive.

I don't have numbers, but could come up with them if my words don't
convince you :)  But bottom line, there's no way any laptop is going to
come anywhere near matching the performance of a desktop built
specifically for RAW conversion and digital image manipulation.

I'd be happy to provide more advice on building a good desktop if you
like. I just put together my personal killer DigiPhoto computer, and
I'm now very happy.

Lisa
Anonymous
a b D Laptop
September 30, 2004 3:04:00 PM

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

"Ron Reaugh" <rondashreaugh@att.net> wrote:

> You wont get desktop performance especially for photoediting and
> HD DTR.

Can you quantify that? Ignoring the HD for a moment, how much
difference in performance of a CPU bound and memory intensive task am I
likely to see between a DTR and a desktop both equipped with a P4 and 1
GB RAM?

For the hard drive, the Hitachi Travelstar 7K60 at 7200 RPM has specs
not quite as good as a 7200 RPM desktop drive. But I am not certain
how much of a difference that will make in real use. Probably
noticeable but I think not too serious a problem.

The tests I have seen show that this drive greatly outperforms the
typical laptop drive, which seem mostly to be 4200 RPM drives.

I am concerned though that the rated service life of a laptop drive is
far less than that of the usual desktop drive.

Can anyone point me to any published benchmarks that address these
questions?

Guy
Anonymous
a b D Laptop
September 30, 2004 5:05:20 PM

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

Guy Scharf wrote:
> "Ron Reaugh" <rondashreaugh@att.net> wrote:
>
>> You wont get desktop performance especially for photoediting and
>> HD DTR.
>
> Can you quantify that? Ignoring the HD for a moment, how much
> difference in performance of a CPU bound and memory intensive task am
> I likely to see between a DTR and a desktop both equipped with a P4
> and 1 GB RAM?
>
> For the hard drive, the Hitachi Travelstar 7K60 at 7200 RPM has specs
> not quite as good as a 7200 RPM desktop drive. But I am not certain
> how much of a difference that will make in real use. Probably
> noticeable but I think not too serious a problem.
>
> The tests I have seen show that this drive greatly outperforms the
> typical laptop drive, which seem mostly to be 4200 RPM drives.
>
> I am concerned though that the rated service life of a laptop drive is
> far less than that of the usual desktop drive.
>
> Can anyone point me to any published benchmarks that address these
> questions?
>
> Guy

I no longer have the link, but toms hardware www.tomshardware.com did a
review of the Centrino (pentium M) that included benchmarks against
desktops and laptops. Additionally, you might consult the pcmark
database, the sisoft sandra benchmarks databases, etc. for some level of
quantitative comparisons among dtr and desktop components. Lisa Horton,
below, says in her personal systems the dtr is 50% of the destop's
performance in photo editing. My experience with benchmarks is that
there are very substantial dtr performance hits with RAM read/write
(30-50%), HD (4200 rpm, 50% at least), CPU performance (2.8Ghz P4M,
Intel 845 MP chipset, 20-30%). Overall, the DTRs are capable, but
differences in apparent performance are more than noticeable with
consumer level desktops. Only with video with a 64MB VRAM, ATI Radeon
Mobility does the dtr performance begin to approach desktop performance,
90% or greater of the desktop card. Benchmarking is fraught with
quantitation problems since there is very little guarantee that the
comparison database is not munged somehow; nevertheless, I believe
Lisa's 50% real world hit is not far off the mark.


Q
Anonymous
a b D Laptop
September 30, 2004 5:55:16 PM

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

Lisa Horton <Lisa091704@lisahorton.net> wrote:

> I have a fast DTR laptop, Athlon 64 3000+, but it's got a 4200 RPM
> drive in it. In RAW conversion speed, and in PS batch job speed,
> there is simply no comparison to my desktop system (Athlon 64
> 3200+, RAID 0+1). The desktop RAW conversion times are like half
> of what they are on the laptop.
>
> In RAW conversion, I've found two bottlenecks. The first, and
> most common is, not unexpectedly, CPU. But that's
> straightforward, if not cheap, to fix. Once that's dealt with,
> you hit the bottleneck of disk throughput. This is slightly
> harder to fix. While RAID 0 turns out not to offer much, if any,
> real advantage in "everyday use", it most assuredly DOES offer
> benefit to those of us who regularly manipulate many large files.
> At least, as long as the RAID is implemented properly, which means
> always a channel for each drive.
>
> I don't have numbers, but could come up with them if my words
> don't convince you :)  But bottom line, there's no way any laptop
> is going to come anywhere near matching the performance of a
> desktop built specifically for RAW conversion and digital image
> manipulation.
>
> I'd be happy to provide more advice on building a good desktop if
> you like. I just put together my personal killer DigiPhoto
> computer, and I'm now very happy.

I am very interested, especially in any performance data you have on
RAID.

My current desktop plan, which is cost limited, is for a Asus K8V SE
Deluxe motherboard with Athlon 64 3400+, 1 GB of RAM, Matrox G550 video
(maybe 550), all in an Antec Sonata case. I'll add more RAM later when
the price of a 1 GB stick drops. Other items, such as DVD RW, will be
salvaged from my current desktop. With a WD 250 GB SATA drive, total
cost will come to about $1K.

I had planned on a WD 2500JD 250 GB SATA drive as the primary drive.
As secondary drives, an existing but new Seagate 200 GB and an existing
used WD 100 GB ATA drives. That's a lot of disk space, but I use
automated backups of the used portions of the hard drives from several
systems to other hard drives, and keep multiple copies of the backups.
So at least half that total space will be for security.

Guy
Anonymous
a b D Laptop
September 30, 2004 6:27:46 PM

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

"Quaoar" <quaoar@tenthplanet.net> wrote:

> I no longer have the link, but toms hardware www.tomshardware.com
> did a review of the Centrino (pentium M) that included benchmarks
> against desktops and laptops. Additionally, you might consult the
> pcmark database, the sisoft sandra benchmarks databases, etc. for
> some level of quantitative comparisons among dtr and desktop
> components. Lisa Horton, below, says in her personal systems the
> dtr is 50% of the destop's performance in photo editing. My
> experience with benchmarks is that there are very substantial dtr
> performance hits with RAM read/write (30-50%), HD (4200 rpm, 50%
> at least), CPU performance (2.8Ghz P4M, Intel 845 MP chipset,
> 20-30%). Overall, the DTRs are capable, but differences in
> apparent performance are more than noticeable with consumer level
> desktops. Only with video with a 64MB VRAM, ATI Radeon Mobility
> does the dtr performance begin to approach desktop performance,
> 90% or greater of the desktop card. Benchmarking is fraught with
> quantitation problems since there is very little guarantee that
> the comparison database is not munged somehow; nevertheless, I
> believe Lisa's 50% real world hit is not far off the mark.

Thank you; that is very helpful. Lisa's and your reports have
convinced me not to bother with a high power DTR. Instead, I'll do a
massive upgrade to the desktop and get a lower powered laptop for the
occasional use that I need one for.

I understand well the problems with benchmarking. I've benchmarked
everything from huge mainframes to PC's in my career. Effective
benchmarks are very hard work to create.

Guy
Anonymous
a b D Laptop
September 30, 2004 7:57:06 PM

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

There is no way a single laptop drive, even one
of the fastest such as the IBM/Hitachi E7K60 or
the Toshiba MK-5024GAY can keep up with one or
more 3.5" desktop drives running at 7,200,
10,000 or 15,000 rpm.

Photoediting is not only CPU intensive, it also
requires very high memory and i/o bandwidth. As
to benchmarks, why don't you test yourself the
the workloads that are of interest to you on
the laptop of your choice before you decide
buying it?


dk


"Guy Scharf" <guy@spamcop.net> wrote in message
news:Xns95745C40C8EA9guyspamcopnet@216.196.97.142...
> "Ron Reaugh" <rondashreaugh@att.net> wrote:
>
> > You wont get desktop performance especially for photoediting and
> > HD DTR.
>
> Can you quantify that? Ignoring the HD for a moment, how much
> difference in performance of a CPU bound and memory intensive task am I
> likely to see between a DTR and a desktop both equipped with a P4 and 1
> GB RAM?
>
> For the hard drive, the Hitachi Travelstar 7K60 at 7200 RPM has specs
> not quite as good as a 7200 RPM desktop drive. But I am not certain
> how much of a difference that will make in real use. Probably
> noticeable but I think not too serious a problem.
>
> The tests I have seen show that this drive greatly outperforms the
> typical laptop drive, which seem mostly to be 4200 RPM drives.
>
> I am concerned though that the rated service life of a laptop drive is
> far less than that of the usual desktop drive.
>
> Can anyone point me to any published benchmarks that address these
> questions?
>
> Guy
Anonymous
a b D Laptop
September 30, 2004 7:57:07 PM

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

"Dan Koren" <dankoren@yahoo.com> wrote:
> Photoediting is not only CPU intensive, it also
> requires very high memory and i/o bandwidth. As
> to benchmarks, why don't you test yourself the
> the workloads that are of interest to you on
> the laptop of your choice before you decide
> buying it?

I often test before buying. But before testing, I collect as much
information, anecdotal and otherwise, as I can. Testing one of the
laptops I am considering requires driving three hours *each* way, plus
ferry fees. Another is as close as an hour away, in a different
direction of course. And some are available only by mail order. It's
much more productive to ask people who have already made the tests what
they have discovered.

Guy
Anonymous
a b D Laptop
September 30, 2004 10:58:34 PM

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

Lisa Horton <Lisa091704@lisahorton.net> wrote:
> I'll be presumptuous and assume that you'd like comments on your
> plan :) 
>
> One of the great features of the K8V SEDX is that it's got LOTS of
> disk channels. Given what we now know about RAID 0, I would
> suggest thinking in the direction of one fast medium size drive
> for boot, Windows, and Apps, and a SATA RAID 0 array for photo
> storage. One of your existing drives could be useful as a
> Photoshop swap drive, you'll certainly have the drive channels for
> it.

Thank you; this has all been very helpful. I'm going to rethink the
disk configuration. Did you assemble the RAID configuration yourself
or have a shop do it? (If you did it, I'm going to remember your name
when it comes time to do it myself. :-))

Which video card did you select, and are you happy with your choice?

Guy
Anonymous
a b D Laptop
September 30, 2004 10:58:35 PM

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

Guy Scharf wrote:
>
> Lisa Horton <Lisa091704@lisahorton.net> wrote:
> > I'll be presumptuous and assume that you'd like comments on your
> > plan :) 
> >
> > One of the great features of the K8V SEDX is that it's got LOTS of
> > disk channels. Given what we now know about RAID 0, I would
> > suggest thinking in the direction of one fast medium size drive
> > for boot, Windows, and Apps, and a SATA RAID 0 array for photo
> > storage. One of your existing drives could be useful as a
> > Photoshop swap drive, you'll certainly have the drive channels for
> > it.
>
> Thank you; this has all been very helpful. I'm going to rethink the
> disk configuration. Did you assemble the RAID configuration yourself
> or have a shop do it? (If you did it, I'm going to remember your name
> when it comes time to do it myself. :-))

Oh, I built the whole machine myself. The RAID array is not difficult
to set up. It's primarily a matter of reading and following
instructions, honestly. That, and being careful with cabling. I mark
some of the cables.

Mechanicaly, the only concern is heat usually. A typical mid tower with
4x 3.5 bays means all four drives are going to be stacked together. A
case with a front fan eliminates that problem usually. But at this
level of performance, I think it prudent to consider cooling as you
think out your system.

Logically, you need to turn on the right RAID controller in the BIOS,
and for 2x2 (0+1) that means the Promise, not the Via controller. When
you load Windows, it will tell you to press F6 if you need to load RAID
drivers, which you will.

>
> Which video card did you select, and are you happy with your choice?

I have a Matrox Millenium G400. Great card for Photoshop. I'm kind of
taking a break from serious photography, so right now I have an MSI
Radeon 9800 Pro installed. For gaming, I overclock it to better than
9800XT speeds, and it's a pretty good performer. Even though Nvidia
dominates the frame rate race right now, I chose ATI for image quality.
Nvidia uses an ordered array for anti-aliasing, like a square array.
ATI rotates that array by 45 degrees, which results in, IMO, superior
image quality. Of course all of this, being in the 3D domain, is
irrelevant for digital imaging.

Getting kind of far afield here, if you want more or more specific
advice you may email me.

Getting back a little on topic, a serious DTR laptop is a viable choice
for digital imaging if you're willing to sacrifice some performance for
portability. My laptop has a card reader that's blazing fast, making it
good for dumping cards. While RAW conversion and image manipulation
times can't match a high end desktop, they're more than competitive with
a mainstream desktop. And, you can take it with you. With a somewhat
beefy inverter, I can run the laptop and my tabloid size photo printer
on the battery in my van, giving me shoot to print capability anywhere.

Lisa
Anonymous
a b D Laptop
October 1, 2004 2:23:05 AM

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

"Guy Scharf" <guy@spamcop.net> wrote in message
news:Xns95745C40C8EA9guyspamcopnet@216.196.97.142...
> "Ron Reaugh" <rondashreaugh@att.net> wrote:
>
> > You wont get desktop performance especially for photoediting and
> > HD DTR.
>
> Can you quantify that? Ignoring the HD for a moment, how much
> difference in performance of a CPU bound and memory intensive task am I
> likely to see between a DTR and a desktop both equipped with a P4 and 1
> GB RAM?

There are a bunch of "depends". My concern is that in a laptop
configuration how can a P4 3.2 get properly cooled. Look at the kind of
cooling such systems need in desktop configuration. Then ask yourself how
it's possible in a laptop. In some cases what one finds is that the CPU
gets throttled to reduce heat production under continuous CPU load.

> For the hard drive, the Hitachi Travelstar 7K60 at 7200 RPM has specs
> not quite as good as a 7200 RPM desktop drive. But I am not certain
> how much of a difference that will make in real use. Probably
> noticeable but I think not too serious a problem.

It'll be a nontrivial difference depending on how much disk I/O the app load
does.

> The tests I have seen show that this drive greatly outperforms the
> typical laptop drive, which seem mostly to be 4200 RPM drives.

No doubt but also greatly underperforms a top desktop 7200 RPM drive.

> I am concerned though that the rated service life of a laptop drive is
> far less than that of the usual desktop drive.
>
> Can anyone point me to any published benchmarks that address these
> questions?
!