Laserfiche WebLink
<br />11. OVERLAYING <br />An overlay scheme can be used to reduce the core memory require- <br />ments for program execution. A program that is not overlayed normally <br />has all of its instructions (i.e., main program, subprogram, and <br />library routines) in core during the whole time that the program is <br />in execution. An overlayed program has only a part of its instructions <br />in core at any time during execution. When properly over1ayed, the <br />time required to execute the program will be only slightly greater than <br />for a non-over1ayed program. However, an improperly over1ayed program <br />may require several hundred percent more time to execute. <br />HEC-2 is distributed to users as a non-overlayed program that will <br />execute in about 32K (decimal) words (varies from machine to machine) <br />of core storage on computers that have more than 32 bits per computer <br />word. Diagram A, Figure 6 (on page 35) shows the proportional use of <br />core on a typical machine. <br />Note that the blank common area for data storage is reused in <br />three separate blocks. In the distribution version of HEC-2 '(Diagram <br />A), each of the three blocks of blank common are approximately equal <br />in size. The blank common area utilized for SUMPO has been adjusted <br />to be equivalent to the block required for FLDHAZ and the block <br />required for all of the other subprograms. In subprogram SUMPO <br />blank common is used as a buffer space for data storage of previously <br />computed results that have been written to TAPE95 (see Section 10 - <br />Memory Management). If this space is increased, the core requirements <br />of the program increase proportionally as shown in Diagram C. For <br /> <br />33 <br />