Joined: 18 Jul 2007 Posts: 2150 Location: At my coffee table
Can you do buffers on the SORTIN? Do you need to do buffers on the SORTIN?
Does the INCLUDE take effect after the sort or before?
The RLSEs are not needed, are any of the SORTWKs needed? Is your specifing them limiting Syncsort's own abilities to dynamically allocating a better value/number?
Have you checked your sort manual?
Joined: 22 Apr 2006 Posts: 6258 Location: Mumbai, India
You might like to consider PARASORT as your SORTIN is on tape. Is yout sortin multi-volume? Then too ,you should be considering PARASORT. PARASORT allows you to read more than one volume at a time. PARASORT is fully documented in Chapter 10 of the SyncSort for z/OS Programmer's Guide.
Joined: 22 Apr 2006 Posts: 6258 Location: Mumbai, India
One more time - Tuning SyncSort is usually application specific. So, we'll be intrested to see "B" messages, if they are not already present in SYSOUT, you can add the following to the sort step(s) in question:
Joined: 22 Apr 2006 Posts: 6258 Location: Mumbai, India
seman18 wrote:
I remove the sortwk statement, however, it doesn't help.
Removing sortwork should not affect elasped time, if you code them sort will look for them (and they will be checked for syntax error) otherwise it'll pick, how many are neede, as per DYNALLOC parameter.
I hope other posts of mine are visible at other end (some body tell me )- you might like to consider them...
Anuj's suggestion of PARASORT is probably your best bet. Peter's ques about vol mount delays can be be addressed by specifying 2 tape units in your JCL. It also can negate the time lost in tape rewinds.
But that assumes the tape operator hasn't fallen asleep.
I got the SYNCSORT Programming guide, and use the synclist to get the following information
Code:
3420 3480/90 3490E 3590
|----------|----------|----------|----------|
4-WAY | NOT | NOT | PARE41 | NOT |
INPUT | POSSIBLE | POSSIBLE | PARE42 | POSSIBLE |
| | | PARE43 | |
| | | PARE44 | |
|----------|----------|----------|----------|
3-WAY | NOT | NOT | PARE31 | NOT |
INPUT | POSSIBLE | POSSIBLE | PARE32 | POSSIBLE |
| | | PARE33 | |
|----------|----------|----------|----------|
2-WAY | NOT | NOT | PARE21 | NOT |
INPUT | POSSIBLE | POSSIBLE | PARE22 | POSSIBLE |
|----------|----------|----------|----------|
and I got the dataset in tape's detail using listcat as following