While doing some testing in CA Easytrieve release 6.4 versus release 11.6 involving VSAM files, we notice that CPU time was greater in CA Easytrieve release 11.6 than in CA Easytrieve release 6.4.. Why?
Regarding performance, we did testing and found that the compilation of a CA Easytrieve program is slower with release 11.x (due to Assembler vs. C++), but the execution performance was about the same. The release 11.x runtime did use more memory. Programs with release 11.x functionality will not run under NEWFUNC set to N.
Testing here of up to 10 million records shows a consistent difference of .11 CPU seconds. We have to attribute the time difference at a customer site to the varied factors such as number of tape input records, paths the logic
takes depending on the different input VSAM records, DCB change of primary input tape file (LRECL and BLOCKSIZE), possible CPU load, and other unknown external changes.
In conclusion, we found that the compilation of a CA Easytrieve program is slower with release 11.x, (due to Assembler vs. C++), but the execution performance was about the same. The release 11.x runtime did use more memory. So, for Compile-and-Go jobs, they will notice the jobs running slower for release 11.x. As more time is spent in execution, there will be less of a difference between the two releases.
And, of course, the previously compiled and linked CA Easytrieve application programs will continue to run as release 6.4, so their performances will be unchanged running as 11.6-c in compatibility mode.
Release: EDBMSU00200-11.6-Easytrieve-Report Generator-Option for DB2-MSU