Select Options.Data File Load Settings to bring up the dialog box to the left.  From this dialog, you can modify how BARfly loads data files.

• Ignore Pause Points:  Prevent incremental deserialization from occurring.  This forces the entirety of each file to be read at once from the disk; there will never be any need to continually press F5 to continue deserialization.
• Force No Fragmentation:  Load each file as a single contiguous portion into memory, and split into nodes later.  If you do not specify this option, BARfly will locate the largest free memory block for each node that must be deserialized, with the possibility that the file contents will become fragmented.
• Only load data files with matching extensions:  Prevent BARfly from loading data files with extensions that do not conform to an implementation file's FILEEXTENSIONS control variable.  If checked, this setting will automatically fail any attempts to open a data file with a non-matching extension, even if the file is valid.

There are trade-offs when considering whether or not to choose Force No Fragmentation.  Files generally load much faster and are maintained in memory blocks more efficiently when this setting is checked.  However, there might not always be enough contiguous memory available for such a task, which can prevent successful loading of some large files after memory has already become fragmented.

  See also: [Data file load settings] [Data file save settings] [Settings for loading and saving BAR implementation files]
[Node editing settings] [Tree control line info settings] [Subnode view settings]
[Dump view settings] [Text view settings] [Raw view settings]
[Memory and BAR_Services settings] [Message display settings] [Toggling tool windows]