Computer Graphics with Photoshop - Tests of PS-CS2 Scratch Disks, C:\, USB 2.0 & FW

Author

Tests of PS-CS2 Scratch Disks, C:\, USB 2.0 & FW

Hunt

2005-11-29, 10:14 pm

Here are the results of my PS-CS2 test of Scratch Disks, C:\, USB2.0 and FW-
400. First the system background: Toshiba P25-S670, Pentium 4 3.2GHz, 2GB DDR
-SDRAM, 80GB Ultra-100 7200RPM HDD (29.8GB free), Nvidia GeForce FX Go5700 w/
128MB VRAM 32bit color 1440x900, XP SP-1, Page File=1536MB, Hyper-Threading=
ON. Only programs open were: NAV, StopZilla, Version Cue, SpySweeper, all in
the SysTray and background.

Photoshop CS-2 (all updates), History States=20, Cache Level=6, MaxRAM=55% (
952MB, from 1731MB free).

The file: recent panorama 671.3MB, 16bit/Channel color, 15 Layers, 5
Adjustment Layers, 2 Paths, 2 Alpha Channels, 7084px x 2853 @ 300ppi, Scratch
Size=1.10GB.

The test: Choose full-sized Layer (sky), and run Levels - Load from Save,
Curves - Load from Save, Gaussian Blur (9.0), Neat Image - Auto Profile -
Apply, Smart Sharpen, Flatten Image. Purge All was run before each test, as I
noticed a 20" improvement when I first ran with my C:\ as my Scratch Disk.

With my "Startup" Disk as only Scratch Disk, time was 3:01.64' to build
Histograms and complete, and 2:01.65' with Histograms built. Maxtor 250GB w/
201GB free USB 2.0, time was 3:42.00' to build Histograms and complete, and 2
:17.82' with Histograms built. Same drive on FW-400, time was 3:37.78' to
build Histograms and complete, and 2:16.63' with Histograms built. For final
speed, Startup was fastest by 15", with FW-400 very slightly ahead of USB 2.0
by 01.19". All-in-all, practically a wash, but internal HDD was faster. Note
that I had 201GB free on that drive and that it is an Ultra-100. I assume that
a SATA, or SATA II HDD (or a 360 SCSI) would top this and a RAID0 would
probably add some more.

OK, Bill Hilton's figures were more noticeable, in favor of the C:\ Scratch
Disk. What are the differences? For one, my machine is running 2GB RAM, so
there is less Scratch Disk usage. Bill, did you note the Scratch Size of your
files? Does Hyper-Threading help, or hurt? Bill, refresh my memory, did you
run HT during your tests? One last thought - Bill, when you did your tests,
was the state of the Histogram(s) the same for all tests? As these tests were
so close, I chose to not gather other files, and do the same thing. Plus, I
had to get back to the edit on this panorama for my client's deadline, in
hopes of paying for a new MoBo plus 1.5TB of RAID0 HDDs for my workstation! It
appears that there is little slowdown with my system using either the USB 2.0
or FW-400 Scratch Disk v the HDD. If one is faced with the prospect of only a
single HDD, that is filling up v a removable, it seems that there isn't too
great a loss in speed, all things being otherwise equal to my machine. Last,
Bill do you have any suggestions on what I might have overlooked in my tests?
I had expected C:\ to be quantitatively faster, with FW-400 (maybe if I had
had FW-800 something might have changed?), and USB 2.0 being way down the
pile, but I didn't see that much difference.

Hunt

Harvey

2005-11-30, 3:15 am

Thank you Hunt for you extensive tests.

My notbook had 512 of 2100, and an 80gb hd that is spinning at 4500.
Looking at the tests, maybe an external hd might not be any slower than my
internal. Of course I would bring up my ram to the max of 1g.
"Hunt" <noone@hunt.com> wrote in message
news:dmiqkb01u69@news1.newsguy.com...
> Here are the results of my PS-CS2 test of Scratch Disks, C:\, USB2.0 and
FW-
> 400. First the system background: Toshiba P25-S670, Pentium 4 3.2GHz, 2GB
DDR
> -SDRAM, 80GB Ultra-100 7200RPM HDD (29.8GB free), Nvidia GeForce FX Go5700
w/
> 128MB VRAM 32bit color 1440x900, XP SP-1, Page File=1536MB,
Hyper-Threading=
> ON. Only programs open were: NAV, StopZilla, Version Cue, SpySweeper, all
in
> the SysTray and background.
>
> Photoshop CS-2 (all updates), History States=20, Cache Level=6, MaxRAM=55%
(
> 952MB, from 1731MB free).
>
> The file: recent panorama 671.3MB, 16bit/Channel color, 15 Layers, 5
> Adjustment Layers, 2 Paths, 2 Alpha Channels, 7084px x 2853 @ 300ppi,
Scratch
> Size=1.10GB.
>
> The test: Choose full-sized Layer (sky), and run Levels - Load from Save,
> Curves - Load from Save, Gaussian Blur (9.0), Neat Image - Auto Profile -
> Apply, Smart Sharpen, Flatten Image. Purge All was run before each test,
as I
> noticed a 20" improvement when I first ran with my C:\ as my Scratch Disk.
>
> With my "Startup" Disk as only Scratch Disk, time was 3:01.64' to build
> Histograms and complete, and 2:01.65' with Histograms built. Maxtor 250GB
w/
> 201GB free USB 2.0, time was 3:42.00' to build Histograms and complete,
and 2
> :17.82' with Histograms built. Same drive on FW-400, time was 3:37.78' to
> build Histograms and complete, and 2:16.63' with Histograms built. For
final
> speed, Startup was fastest by 15", with FW-400 very slightly ahead of USB
2.0
> by 01.19". All-in-all, practically a wash, but internal HDD was faster.
Note
> that I had 201GB free on that drive and that it is an Ultra-100. I assume
that
> a SATA, or SATA II HDD (or a 360 SCSI) would top this and a RAID0 would
> probably add some more.
>
> OK, Bill Hilton's figures were more noticeable, in favor of the C:\
Scratch
> Disk. What are the differences? For one, my machine is running 2GB RAM, so
> there is less Scratch Disk usage. Bill, did you note the Scratch Size of
your
> files? Does Hyper-Threading help, or hurt? Bill, refresh my memory, did
you
> run HT during your tests? One last thought - Bill, when you did your
tests,
> was the state of the Histogram(s) the same for all tests? As these tests
were
> so close, I chose to not gather other files, and do the same thing. Plus,
I
> had to get back to the edit on this panorama for my client's deadline, in
> hopes of paying for a new MoBo plus 1.5TB of RAID0 HDDs for my
workstation! It
> appears that there is little slowdown with my system using either the USB
2.0
> or FW-400 Scratch Disk v the HDD. If one is faced with the prospect of
only a
> single HDD, that is filling up v a removable, it seems that there isn't
too
> great a loss in speed, all things being otherwise equal to my machine.
Last,
> Bill do you have any suggestions on what I might have overlooked in my
tests?
> I had expected C:\ to be quantitatively faster, with FW-400 (maybe if I
had
> had FW-800 something might have changed?), and USB 2.0 being way down the
> pile, but I didn't see that much difference.
>
> Hunt
>

Hunt

2005-11-30, 6:14 pm

In article <skbjf.657214$1i.317478@pd7tw2no>, harveyleroux@shaw.ca says...
>
>Thank you Hunt for you extensive tests.
>
>My notbook had 512 of 2100, and an 80gb hd that is spinning at 4500.
>Looking at the tests, maybe an external hd might not be any slower than my
>internal. Of course I would bring up my ram to the max of 1g.
>"Hunt" <noone@hunt.com> wrote in message
>news:dmiqkb01u69@news1.newsguy.com...
[SNIP]

Yes, I'd go with max RAM first (even though much notebook RAM is not the
bargain that desktop RAM is). AND, the point of which interface offers the
greatest advantage will be moot to most NOT on a laptop, as a few fast HDDs is
always better.

Bill Hilton has made a few recommendations for phase II of the tests, and I
will post what they show, maybe next week.

Hunt