Algorithm Improves Performance when Converting Split Track to Full Track
Original Publication Date: 1994-Apr-01
Included in the Prior Art Database: 2005-Mar-26
Candelaria, SK: AUTHOR [+1]
This disclosure describes a method to improve the performance of converting a Cache Split Track Image to a Cache Full Track Image when a format write operation occurs to a Cache Split Track Image.
Algorithm Improves Performance when Converting
SplitTrack to Full
disclosure describes a method to improve the performance
of converting a Cache Split Track Image to a Cache Full Track Image
when a format write operation occurs to a Cache Split Track Image.
operations to tracks in Cache that were originally
saved as split track images require a re-ordering of the storage
address segment pointers for each record entry within the Track
Information Block (TIB) so the new image of the track matches that of
a standard full track.
storage addressing hardware provides a Segmented
Array (SGA) that allows for accessing up to a maximum of sixteen
storage segments. As current track sizes only require a maximum of
five storage segments, four extra segments in the Segment Array are
used to repeat the first four segment numbers so a split track can be
format written without having to reload the Segment Array.
A split track
image is built from a Read or Write operation
that begins at some record R sub n, beyond the Home Address (HA), and
progresses through Index and continues operation on the same track
through HA until reaching record R sub n-1 The resulting segment
allocation for the split track image is shown in Fig. 1.
having to reload the SGA, the Format write operation
can be done to the Split Track Image directly. Once the Format Write
is complete, the segment pointer for the Home Address Record will be
at an offset that is grea...