Setting PSCS2 for dual processor Athlon CPU

Discussion in 'Photoshop' started by mogur2, Jul 20, 2005.

  1. mogur2

    mogur2 Guest

    My new system has an Athlon dual processor CPU. Is there anything special that I should
    do with the preferences in PSCS2?

    Thanks
     
    mogur2, Jul 20, 2005
    #1
    1. Advertisements

  2. mogur2

    KatWoman Guest

    The chip is the speed of the processor no need to set up anything in PS.
    Dual processing and hyper thread just means your comp can do more than one
    program at a time wihtout suffering slowdowns. Someone on here posted an
    excellent test not to many days ago.
    what you have to set in preferences is your SCRATCH disks. the more space PS
    can use the better.
    So it is better to have more than one hard drive to run PS.

    KW

    from his post on:
    comp.graphics.apps.Photoshop JUL 17


    This is the definitive test for Photoshop using Hyper. Well, it is
    for me. Of course, I recommend that you do your own testing on your own
    system for the way that you work in Photoshop.

    Because I couldn't find any hard numbers to support a lot of claims, I
    just ran my own tests to see if Photoshop ran faster with HT on or off.
    These tests were run on my home-built computer. The motherboard is Intel
    D865GBF. The processor it Intel P4 3.2 GHz Prescott - and HT, of course.
    I have 2 GB of memory in 512 KB chunks of DDR400 using Dual Channel
    Mode. This is Corsair CL 2.5 memory bought in pair, but not the
    expensive Dual Channel tested stuff. I have a 120 GB Seagate (7200 rpm)
    as my system drive and second Scratch Disc for Photoshop. I use a
    Western Digital Raptor 36.7 GB 10,000 rpm drive for Window XP Pro's
    Pagefile AND for Photoshop's first Scratch Disc. Both HDs are connected
    by SATA 150.

    I still use Photoshop CS. I haven't had a compelling reason to upgrade
    yet. I'm also waiting for that first bug fix release; then I'll upgrade
    regardless. I am using the > 1 GB plugin from Adobe.

    My primary test was an action that included almost all the steps that I
    use in my pro wedding photography business. This is somewhat my workflow
    and a bit more. i.e. I never do all of this on any one picture. I do use
    all of this on some pictures. I didn't include ACR in this because that
    is usually interactive for me. My sharpening tool is Focus Magic and it
    only works interactively. Here is a high level summary of my action;
    default setting unless otherwise indicated:

    Convert to RGB
    Flatten Image
    Run CurveMeister 2.0 - loading a setting
    Create a Hue/Sat adjustment layer - set Saturation at 20% - Merge Down
    Run Noise Ninja - using automatic noise profile
    Run PTLens - this calls pano12.dll
    Shadow/Highlights
    Convert to 8 bit
    Run digital background filters - Sprayed Strokes, Spatter, Lighting
    Effects (not quite like I do backgrounds)
    Save As - JPEG 2000 lossless - in different folder
    Convert to B&W - using an action that does the 2 adjustment layers
    method (See Dr. Brown)
    Convert to Duotone - Tritone in my loaded setting
    Save As - PSD
    Close

    Well, that's what's important to me. You are free to do what's important
    to you. Of course, the filters are the big time eaters in this or any
    action. Focus Magic can take longer than any of these - if a lot of
    sharpening is needed. I doubt it would changed much though.

    I also ran Contact Sheet II. I put 4 on a page, but that didn't take
    very long to run. I then ran Dr. Brown's Image Processor 2.2 resizing
    and saving them all to JPEG, PSD, and TIFF with LZW.

    To use a more standard test I also ran SiSoftware Sandra Lite 2005.SR2a.
    I ran the standard Combined Performance Index Wizard. For CPU that seems
    to be a combination of Dhrystone and Whetstone. It did test CPU
    Arithmetic, CPU Multi-Media, Memory, and Storage. I have it set to not
    test Network, because I don't have a server to pound.

    I did all this on nine (9) pictures. I picked a pretty good variety of
    pictures that I regularly do. These were a variety of file formats,
    color depths, color spaces, sizes, etc. The action ran for between 4 and
    6 minutes for the action tests. I ran the tests multiple times, but it
    didn't matter; there was no significant deviation.

    I ran the tests on my computer with no other applications running with
    both HT on and off. I left the standard boot apps running. So, my avast!
    AV software and Sygate firewall were running, but not doing anything
    that I could see.

    Since HT's big benefit seems to be the ability to run multithreaded apps
    and/or multiple apps at once. I also ran all the tests with a background
    app running. Frankly, I don't have any app that runs consistently. I do
    have Windows Media Player and iTunes. They run pretty consistently, but
    not hard enough for testing. So, I used SiSoftware Sandra's Burn-In
    Wizard to pound the two CPU burn-in tests. I ran these as Normal
    Priority with the CPU level set at 33%. That was just my own gut feeling
    of a realistic background level. Most of the time I don't have that much
    running in background, but I do get spikes higher than that too.

    Since the action test is most significant to me (and probably you in
    this newsgroup), I'll start with those results....

    With no background application running, my computer runs 7% faster with
    HyperThread turn on. With the background app simulation running, my
    computer ran 15% faster with HT turned on. This is just the Photoshop
    action test.

    The running of the Contact Sheet II showed no significant difference
    between HT on or off. Part of the reason for this may be because 9
    pictures wasn't long enough test. Ideally I should run this on many more
    files, but it still may not be that big of a difference. The other
    reason may be described below in the Sandra Memory and Storage results.

    The Dr. Brown's Image Processor test the times were longer, but still
    pretty short. Without a background app, there was no difference between
    HT on or off. With a background app running, HT gave a 10% increase. I
    don't think this is very significant though.

    Sandra's Memory and Storage test results show that there is no
    difference between HT on or off. Since HyperThreading is a CPU
    utilization enhancement, it doesn't affect the the slower technologies
    of memory and hard disks. So, the tests show no improvement.

    Since Contact Sheet II and Dr. Brown's Image Processor are mostly doing
    disk access (the slow part anyway), they shouldn't be much affected by a
    CPU enhancement. They do some image resizing, but that appears to be
    pretty efficient and doesn't change things enough to matter.

    I did not run Sandra's benchmarking with a background app. For one
    thing, it wouldn't let me run it's benchmark and burn-in wizards at the
    same time. Then again, benchmark apps do take up all the tested
    resources. So a background app doesn't make much sense.

    Sandra's CPU Arithmetic benchmark is 21% faster with HT on. The CPU
    Multi-Media benchmark runs 34% faster with HT on.

    In no test did I find that anything ran slower with HyperThreading turn
    on. Everything ran just as fast or faster with HT on. It seems to run
    just as promised; multithreaded apps and multi apps running at the same
    time can use efficiencies in the CPU allowed by HT.

    It doesn't seem to help IO intensive processes, but it shouldn't either.
    Then again, it doesn't hurt them either. If you have CPU intensive apps
    - and who doesn't - HT can only help. Photoshop certainly seems to be
    improved by having HT on. My gut feeling matches this. My experience
    using Photoshop told me that it was most helpful when I had other
    applications running at the same time. Since I usually do have other
    apps running, it has been useful. It is good to finally have some hard
    numbers to support that gut feeling.

    Of course, this isn't the final, universal answer either. It is the
    first hard numbers that I have seen - from someone other than Intel or
    AMD. Don't forget that you are welcome to change anything in my testing
    setup and run your own tests. I encourage you to do just that. Then post
    your own results in this august newsgroup.

    A comforted and happy user,
    Clyde
     
    KatWoman, Jul 20, 2005
    #2
    1. Advertisements

  3. mogur2

    Hecate Guest

    There is a *big* difference though, between dual processing and
    hyperthreading - the former is physical, the latter is virtual.

    And the Athlon Dual is faster than the Pentium D with HT.

    --

    Hecate - The Real One

    Fashion: Buying things you don't need, with money
    you don't have, to impress people you don't like...
     
    Hecate, Jul 20, 2005
    #3
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.