Tucson Apple Core Newsletters! 
          Contact Us
TAC Home Meetings Newsletters Message Board TMUG Newsgroups
TAC Newsletter Topics

  • Printing from a IIgs to an Apple LaserWriter
  • Using 3.5" drives on an Apple IIe, IIc, IIc+, IIgs
  • ImageWriter II Printing Problem
  • Using Classic AppleWorks on Your new iMac
  • Apple IIc Power Supply
  • Requirements for the Macintosh Apple IIe Card
  • Apple IIgs Battery Replacement
  • Diskettes With Bad Blocks
  • Connecting an Apple IIgs to a Mac
  • Image Writer II Network Printing
  • About 3.5" drives for the Apple IIc
  • What is Interleave?
  • ProDos volumes and mapping

    What is Interleave? By Ridge Prevost, Tucson Apple Core

    Question:

    What is Interleave?

    Answer:

    Hard disks and floppy disks use spinning circular platters of magnetic media on which to store computer data. These platters are organized in an orderly arrangement of concentric tracks. Each track is further divided into sectors (blocks). Data is written to and then read from these sectors.

    When disk (hard or floppy) formatting (initializing) software is used, a disk's circular tracks and sectors of those tracks are "laid down" on the surface of the media. There is a sequential numbering, naming or addressing scheme that is applied by the software to each sector of a track as it is being created.When each sector of a track is numbered contiguously & consecutively (1,2,3,4,5...) a disk is said to have a 1:1 interleave.

    When the numbering sequence skips every other consecutive sector (1, skip the next one, 2, skip the next one, 3, skip the next one, 4, skip the next one, 5...) a disk is said to have a 2:1 interleave.

    If the numbering sequence skips two consecutive sectors (1, skip the next two, 2, skip the next two, 3, skip the next two, 4, skip the next two, 5...) a disk is said to have a 3:1 interleave.

    All sectors in a track get a number assigned to them; with interleave other than 1:1, the next higher numbered sector will not be the next sector on the track, it will be one or more sectors farther on, depending on the interleave ratio.

    Computers read data from the sectors of the tracks in numerical order of the sectors' assigned numbers. The sector labelled 1 is read first, followed by the sector labelled 2, etc.

    A computer that is fast enough to process disk data input/output (I/O) that is formatted at 1:1, will not process as many bytes per second if the interleave is 2:1 or higher, since it will be waiting for the next numbered sector while the skipped sector(s) pass under the drive's read/write (R/W) head.

    Computers that are slow at handling disk I/O will not process 1:1 data very efficiently, since by the time they are finished processing sector 1, sector 2 has already gone by and a wait is needed until the number 2 sector comes around again for processing. These computers will need 2:1, 3:1, or 4:1 interleave to process disk I/O most efficiently. Exactly which interleave is best is determined by testing.

    An Apple IIe or IIc handles disk I/O most efficiently at 4:1 for 3.5" diskettes. The IIgs does best with a 2:1 interleave. By association, I would think that the situation for hard drives (HD) would be similar, using Apple's High-Speed SCSI interface card. I don't know what would be correct for drives using the RAMFast SCSI card, or the IDE controller cards used with Applied Engineering and Ingenuity internal HD's.

    The RAMFast card streams data as fast as the HD can handle it by means of an on-board hi-speed static RAM cache buffer; data is passed to and from the computer at RAM speeds (very fast). HD interleave is relatively unimportant in this situation, although 1:1 would be optimum. Ohio Cache Systems originally developed the process for their floppy diskette controller cards to speed up disk I/O; the technology was later applied to the RAMFast SCSI HD interface card.

    The issue of interleave comes up most often when a 3.5" 800k diskette is initialized on an Apple IIe/c with Apple's utilities or Copy II Plus (4:1) and then seems slow in a IIgs, or when it's initialized on a IIgs using the GSOS Finder (2:1) and seems slow in a IIe/c. The IIe/c is slower at disk I/O processing than the IIgs.