Announcement

Collapse
No announcement yet.

Offscreen buffer

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Offscreen buffer

    I'm running an experiment with 600 small (30-40kb each) images which has been running fine. However, in the last couple of days it has started freezing and crashing the experiment. Occasionally i get the message "Unable to allocate memory for offscreen buffer" within a Superlab Error window.

    It is running on an Intel Macbook with 1Gb memory and lots of available disk space. I have run the same basic experiment on a G4 iBook with on problems.

    Has anyone got any suggestions???

  • #2
    First of all, 1GB of RAM isn't enough for any Intel Mac, but regardless, that wouldn't cause crashes or error messages to appear.

    Second, the size of the image on disk has little bearing on how much RAM it will require when decompressed. A 76kB PNG I found on my desktop would use about 1.8MB of RAM when decompressed (width*height*32 bits/pixel). On the other hand, even loading a dozen of those at the same time isn't going to stress a machine with 1GB of RAM.

    In the Experiment Options dialog, select "Memory Management," and then set it to load stimulus files "For only one trial at a time." In addition, make sure you aren't loading all 600 of those image files in the same trial.


    I hope this helps!

    Comment


    • #3
      Offscreen buffer revisited

      I know this has been a while, but i was led to believe this issue had been resolved.

      I have an experiment that was created in SL 1.75 for Mac (Yes, it's that old!!) by a PhD student on a iBook running OS 9.2. It has 800 stimuli images in ONE Trial. This runs and displays without a problem (I've just retested it).
      The Experiment was updated to run in SL 4 with only 600 stimuli on a Macbook running OS 10.5. These images are only about 30-40Kb each. This takes an age to load (as you'd expect, although the original didn't) but stops after displaying about 100 images. I've tried splitting the stimuli into 25 events per trial, but the delay in loading the next trial is quite noticeable. Changing to load all trials in Memory Management causes other issues, in that sometimes it doesn't start or reaches about 100 stimuli before stopping. Even reducing trials to 10 stimuli each causes it to stop after 100.

      Any thoughts???

      Comment


      • #4
        There are several reasons as to why the pictures are failing to load, e.g.:
        • Mac OS X itself requires considerably more memory than Mac OS 9 did.

        • On the Mac, SuperLab 1.75 supported only the PICT file format. PICT required far less disk space than other formats, but nevertheless could occupy a lot of memory when loaded, depending on the file contents.

          In other words, the fact that the images are "only about 30-40Kb" may not mean much, depending on the file.

        • SuperLab 4 uses more memory per picture. It's a tradeoff that we had to make on Mac OS in order to present stimuli with decent speed.

        And last but not least, if your trial consists conceptually of a single picture, then you really should be having just one picture per trial. You can probably recreate the experiment fairly quickly if you take advantage of the stimulus lists feature.

        Comment


        • #5
          That did it. Created a stimulus list and it works fine. Many Thanks. I don't know why i didn't think of it.

          Comment

          Working...
          X