View previous topic :: View next topic |
Author |
Message |
issac1029 Intermediate
Joined: 10 Dec 2005 Posts: 159 Topics: 75
|
Posted: Fri Aug 17, 2007 10:28 am Post subject: how the time cost if I mark DUMMY for SORTOUT? |
|
|
how the time cost if I mark DUMMY for SORTOUT?
pgm=sort
sortin dd "very big file"
sortout dd DUMMY
does this will much quicker than I put a real file for output?
I tought dummy will not have I-O cost, am I right?? |
|
Back to top |
|
|
Frank Yaeger Sort Forum Moderator
Joined: 02 Dec 2002 Posts: 1618 Topics: 31 Location: San Jose
|
Posted: Fri Aug 17, 2007 11:09 am Post subject: |
|
|
It depends. The best way to find out is to try it yourself. Is there some reason you can't do that?
But if you need me to try it for you, tell me the RECFM and LRECL of the input file, the number of records in the input file, whether you're sorting or copying, the BLKSIZE for the output file for the case where SORTOUT is not DUMMY, etc.
Just for fun, I tried copying an input file with 1000000 records, RECFM=FB and LRECL=300. When I used BLKSIZE=27900 for the SORTOUT data set, the run with SORTOUT DUMMY took more CPU time. When I used BLKSIZE=300 for the SORTOUT data set, the SORTOUT DUMMY run took less CPU time. YMMV. _________________ Frank Yaeger - DFSORT Development Team (IBM)
Specialties: JOINKEYS, FINDREP, WHEN=GROUP, ICETOOL, Symbols, Migration
DFSORT is on the Web at:
www.ibm.com/storage/dfsort |
|
Back to top |
|
|
amargulies Beginner
Joined: 10 Jan 2007 Posts: 123 Topics: 0
|
Posted: Tue Aug 21, 2007 10:23 am Post subject: |
|
|
issac,
Please post your control cards so we can better address your question.
Also, which sort utility are you currently using? _________________ Alissa Margulies
SyncSort Mainframe Product Services
201-930-8260
zos_tech@syncsort.com |
|
Back to top |
|
|
s_shivaraj Beginner
Joined: 21 Sep 2004 Posts: 140 Topics: 14 Location: Chennai, India
|
Posted: Wed Aug 22, 2007 11:12 am Post subject: |
|
|
Frank Yaeger,
Quote: | The DUMMY parameter specifies that no device is to be allocated to this DD statement and that no disposition processing or I/O is to be done. All other JCL parameters on a DUMMY DD statement have to be syntactically correct. Use the DUMMY parameter if you know you will not need a file in a job step. |
If there is No I/O, how the Block Size is going to affect , i was under assumption even writing to buffer Block will be skipped if its Dummy, correct me if i am wrong..
I am trying to understand how the Block Size plays a role in case of Dummy .. _________________ Cheers
Sivaraj S
'Technical Skill is the Master of complexity, while Creativity is the Master of Simplicity' |
|
Back to top |
|
|
Frank Yaeger Sort Forum Moderator
Joined: 02 Dec 2002 Posts: 1618 Topics: 31 Location: San Jose
|
Posted: Wed Aug 22, 2007 12:02 pm Post subject: |
|
|
Quote: | When I used BLKSIZE=27900 for the SORTOUT data set, the run with SORTOUT DUMMY took more CPU time. When I used BLKSIZE=300 for the SORTOUT data set, the SORTOUT DUMMY run took less CPU time. |
I was comparing a copy to a SORTOUT DD DSN=... data set to a copy with a SORTOUT DD DUMMY data set. The different BLKSIZEs were for the SORTOUT DD DSN=... data set, not for the SORTOUT DD DUMMY data set. BLKSIZE can have a big impact on performance for non-DUMMY data sets. My point was that you have to be careful what you're comparing SORTOUT DD DUMMY to, and that you should really do the comparison yourself with your own job and data rather than ask somebody else to do it for you. _________________ Frank Yaeger - DFSORT Development Team (IBM)
Specialties: JOINKEYS, FINDREP, WHEN=GROUP, ICETOOL, Symbols, Migration
DFSORT is on the Web at:
www.ibm.com/storage/dfsort |
|
Back to top |
|
|
s_shivaraj Beginner
Joined: 21 Sep 2004 Posts: 140 Topics: 14 Location: Chennai, India
|
Posted: Wed Aug 22, 2007 3:36 pm Post subject: |
|
|
Sorry for that misunderstanding _________________ Cheers
Sivaraj S
'Technical Skill is the Master of complexity, while Creativity is the Master of Simplicity' |
|
Back to top |
|
|
|
|