View previous topic :: View next topic
|
Author |
Message |
Amit Yadav
New User
Joined: 20 Sep 2007 Posts: 9 Location: PUNE
|
|
|
|
Hi All .
Could anybody explain ,
Which type of algorithim(heap sort , quik sort ..etc) is used by Dfsort .
Which type of algorithim will be good for mainframe syatem. |
|
Back to top |
|
|
Anuj Dhawan
Superior Member
Joined: 22 Apr 2006 Posts: 6250 Location: Mumbai, India
|
|
|
|
Hi,
Please elaborate your question. . . "type of algorithim" depends on "type of problem" you have in hand . . . |
|
Back to top |
|
|
superk
Global Moderator
Joined: 26 Apr 2004 Posts: 4652 Location: Raleigh, NC, USA
|
|
Back to top |
|
|
Amit Yadav
New User
Joined: 20 Sep 2007 Posts: 9 Location: PUNE
|
|
|
|
If you have to sort millions of records , and i have available SYNCSORT and DFSORT both , then how we can analyse that which will be best performance wise and how?
I have to use this in production environment. |
|
Back to top |
|
|
dbzTHEdinosauer
Global Moderator
Joined: 20 Oct 2006 Posts: 6966 Location: porcelain throne
|
|
|
|
**sigh**
different sites have different configurations.
how about testing?
and then comparing results.
both products have high volume control statements.
I would worry more about learning the control statements.
since you have both products,
maybe asking your production support personnel would be a better use of your time. |
|
Back to top |
|
|
mtaylor
Active User
Joined: 20 Feb 2009 Posts: 108 Location: Kansas City
|
|
|
|
As noted in the cited thread, these algorithms are usually covered by patents. As far as what algorithms are 'good' for mainframes, that's a bit of a nebulous question. The distinction that should concern you is 'in memory' sort versus 'file sort' since presumable on the mainframe you'll be sorting large disk/tape datasets.
There used to be an informal competition held at super computer conferences known as Terabyte Sort in which contestants developed various sorting algorithms (usually parallel) to sort terabyte sized files. There were many academic papers published from these contests detailing the algorithms.
Quote: |
If you have to sort millions of records , and i have available SYNCSORT and DFSORT both , then how we can analyse that which will be best performance wise and how?
I have to use this in production environment. |
Run tests with varying sized datasets upto millions of records then chart the results. Your chart should come close to a nlogn growth curve. |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
see ...
Donald Knuth, The art of Computer programming - Volume 3 ( sorting and searching )
for a theoretical approach |
|
Back to top |
|
|
|