Re: Syncsort > DFsort migration
Syncsort's site says they support IBM's Integrated Accelerator for Z Sort coprocessor for z15 also. Not sure what the limits are when that is used. -Original Message- From: IBM Mainframe Discussion List On Behalf Of Sri h Kolusu Sent: Monday, August 28, 2023 12:39 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Syncsort > DFsort migration >> It was the thread in January 12th 2023, with subject "DFSORT maximum input >> records". The message that kicked it off was: Thanks Michael. It is good to know the other products limits. DFSORT has raised the limit with Sort accelerator to 65,536 times than the other product's limit. Thanks, Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
>> It was the thread in January 12th 2023, with subject "DFSORT maximum input >> records". The message that kicked it off was: Thanks Michael. It is good to know the other products limits. DFSORT has raised the limit with Sort accelerator to 65,536 times than the other product's limit. Thanks, Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
It was the thread in January 12th 2023, with subject "DFSORT maximum input records". The message that kicked it off was: From: Charles Hardee Subject: DFSORT maximum input records MIME-Version: 1.0 I am trying to sort roughly 4295217295, records using DFSORT. DFSORT issues a message that I have exceeded the sort input record maximum. It appears that this maximum is 4294967295. The limit imposed is because, apparently, DFSORT is handling "equals" by adding a fullword record number to the sort key. Syncsort MFX also has a 4,294,967,295 record limit when EQUALS is used, *but* it can be raised to 140,737,488,355,327 using the EXTCOUNT parm. -Original Message- From: IBM Mainframe Discussion List On Behalf Of Sri h Kolusu Sent: Monday, August 28, 2023 11:50 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Syncsort > DFsort migration >> I think it was mentioned in this list previously that EQUALS in DFSORT has a >> lower limit on maximum number of records than SYNCSORT. Michael, Can you please provide link to that topic that mentions that DFSORT has a lower limit than the other product? We do have limits and I *believe* it is the same as the other product, but if you are running on Z15 and higher you can use sort accelerator(with OPTION ZSORT) with DFSORT which can sort up to 9,223,372,036,854,775,807 records ( Nine quintillion two hundred twenty-three quadrillion three hundred seventy-two trillion thirty-six billion eight hundred fifty-four million seven hundred seventy-five thousand eight hundred seven) as mentioned here https://www.mail-archive.com/ibm-main@listserv.ua.edu/msg124898.html Btw it would a while for any customer to sort that many records. Thanks, Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
On Mon, Aug 28, 2023, at 11:59 AM, Tom Brennan wrote: > LOL "a while" > Don't worry, IBM would sell you Capacity-on-Demand while the job ran. Kirk Wolf Dovetailed Technologies http://coztoolkit.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
LOL "a while" On 8/28/2023 9:49 AM, Sri h Kolusu wrote: I think it was mentioned in this list previously that EQUALS in DFSORT has a lower limit on maximum number of records than SYNCSORT. Michael, Can you please provide link to that topic that mentions that DFSORT has a lower limit than the other product? We do have limits and I *believe* it is the same as the other product, but if you are running on Z15 and higher you can use sort accelerator(with OPTION ZSORT) with DFSORT which can sort up to 9,223,372,036,854,775,807 records ( Nine quintillion two hundred twenty-three quadrillion three hundred seventy-two trillion thirty-six billion eight hundred fifty-four million seven hundred seventy-five thousand eight hundred seven) as mentioned here https://www.mail-archive.com/ibm-main@listserv.ua.edu/msg124898.html Btw it would a while for any customer to sort that many records. Thanks, Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: [EXTERNAL] Re: Syncsort > DFsort migration
I don't care how fast your computer is, that would take a while to sort! LOL Rex -Original Message- From: IBM Mainframe Discussion List On Behalf Of Sri h Kolusu Sent: Monday, August 28, 2023 11:50 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: [EXTERNAL] Re: Syncsort > DFsort migration >> I think it was mentioned in this list previously that EQUALS in DFSORT has a >> lower limit on maximum number of records than SYNCSORT. Michael, Can you please provide link to that topic that mentions that DFSORT has a lower limit than the other product? We do have limits and I *believe* it is the same as the other product, but if you are running on Z15 and higher you can use sort accelerator(with OPTION ZSORT) with DFSORT which can sort up to 9,223,372,036,854,775,807 records ( Nine quintillion two hundred twenty-three quadrillion three hundred seventy-two trillion thirty-six billion eight hundred fifty-four million seven hundred seventy-five thousand eight hundred seven) as mentioned here https://urldefense.com/v3/__https://www.mail-archive.com/ibm-main@listserv.ua.edu/msg124898.html__;!!KjMRP1Ixj6eLE0Fj!o8g9lT1CLs5wPI1yiaiwwUXcdegzjzbU5QRAbhEHh2lxOJ87xzEEANRbPI_9SAT1jbDtNSMy1Qn6a3ziDwI$ Btw it would a while for any customer to sort that many records. Thanks, Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- The information contained in this message is confidential, protected from disclosure and may be legally privileged. If the reader of this message is not the intended recipient or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any disclosure, distribution, copying, or any action taken or action omitted in reliance on it, is strictly prohibited and may be unlawful. If you have received this communication in error, please notify us immediately by replying to this message and destroy the material in its entirety, whether in electronic or hard copy format. Thank you. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
>> I think it was mentioned in this list previously that EQUALS in DFSORT has a >> lower limit on maximum number of records than SYNCSORT. Michael, Can you please provide link to that topic that mentions that DFSORT has a lower limit than the other product? We do have limits and I *believe* it is the same as the other product, but if you are running on Z15 and higher you can use sort accelerator(with OPTION ZSORT) with DFSORT which can sort up to 9,223,372,036,854,775,807 records ( Nine quintillion two hundred twenty-three quadrillion three hundred seventy-two trillion thirty-six billion eight hundred fifty-four million seven hundred seventy-five thousand eight hundred seven) as mentioned here https://www.mail-archive.com/ibm-main@listserv.ua.edu/msg124898.html Btw it would a while for any customer to sort that many records. Thanks, Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: [EXTERNAL] Re: Syncsort > DFsort migration
I think it was mentioned in this list previously that EQUALS in DFSORT has a lower limit on maximum number of records than SYNCSORT. -Original Message- From: IBM Mainframe Discussion List On Behalf Of Pommier, Rex Sent: Monday, August 28, 2023 10:54 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: [EXTERNAL] Re: Syncsort > DFsort migration Gil, I had to look up "stable sort" as I hadn't heard the term. Yes that is what EQUALS does in both DFSort and Syncsort. FYI, I searched the Syncsort programmer's manual and it doesn't mention "stable sort" either. :-) Rex -Original Message- From: IBM Mainframe Discussion List On Behalf Of Paul Gilmartin Sent: Monday, August 28, 2023 10:36 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: [EXTERNAL] Re: Syncsort > DFsort migration On Mon, 28 Aug 2023 15:11:05 +, Sri h Kolusu wrote: > >EQUALS is the parm which comes into picture when your input has duplicates on >the key, and you want to retain the order of the duplicates. OPTION EQUALS >specifies whether the original sequence of records that collate identically >for a sort or a merge should be preserved from input to output. > Is this the behavior generally known in IT as a "stable" sort? I find no mention of stable sort in the DFSORT documentation. -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- The information contained in this message is confidential, protected from disclosure and may be legally privileged. If the reader of this message is not the intended recipient or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any disclosure, distribution, copying, or any action taken or action omitted in reliance on it, is strictly prohibited and may be unlawful. If you have received this communication in error, please notify us immediately by replying to this message and destroy the material in its entirety, whether in electronic or hard copy format. Thank you. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
>> Is this the behavior generally known in IT as a "stable" sort? I find no >> mention of stable sort in the DFSORT documentation. Gil, Yes, EQUALS parm which keeps the dup keys in the order and what some call it as "stable" sort. DFSORT documentation does not have the same "stable" definition, but essentially the same functionality. >> FYI, I searched the Syncsort programmer's manual and it doesn't mention >> "stable sort" either. :-) Rex, Thanks for the confirmation about the competitor product. Much appreciated. Thanks, Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: [EXTERNAL] Re: Syncsort > DFsort migration
Gil, I had to look up "stable sort" as I hadn't heard the term. Yes that is what EQUALS does in both DFSort and Syncsort. FYI, I searched the Syncsort programmer's manual and it doesn't mention "stable sort" either. :-) Rex -Original Message- From: IBM Mainframe Discussion List On Behalf Of Paul Gilmartin Sent: Monday, August 28, 2023 10:36 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: [EXTERNAL] Re: Syncsort > DFsort migration On Mon, 28 Aug 2023 15:11:05 +, Sri h Kolusu wrote: > >EQUALS is the parm which comes into picture when your input has duplicates on >the key, and you want to retain the order of the duplicates. OPTION EQUALS >specifies whether the original sequence of records that collate identically >for a sort or a merge should be preserved from input to output. > Is this the behavior generally known in IT as a "stable" sort? I find no mention of stable sort in the DFSORT documentation. -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- The information contained in this message is confidential, protected from disclosure and may be legally privileged. If the reader of this message is not the intended recipient or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any disclosure, distribution, copying, or any action taken or action omitted in reliance on it, is strictly prohibited and may be unlawful. If you have received this communication in error, please notify us immediately by replying to this message and destroy the material in its entirety, whether in electronic or hard copy format. Thank you. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
On Mon, 28 Aug 2023 15:11:05 +, Sri h Kolusu wrote: > >EQUALS is the parm which comes into picture when your input has duplicates on >the key, and you want to retain the order of the duplicates. OPTION EQUALS >specifies whether the original sequence of records that collate identically >for a sort or a merge should be preserved from input to output. > Is this the behavior generally known in IT as a "stable" sort? I find no mention of stable sort in the DFSORT documentation. -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Fwd: Syncsort > DFsort migration
>> The only issue I ran into (this was years ago) was that if you use the PARM >> in Syncsort of “EQUAL” expect different output in DFSORT ED, EQUALS is the parm which comes into picture when your input has duplicates on the key, and you want to retain the order of the duplicates. OPTION EQUALS specifies whether the original sequence of records that collate identically for a sort or a merge should be preserved from input to output. As a simple example: R1 R2 R3 If we sort these three records on positions 1-4 with EQUALS with DFSORT, the output is guaranteed to be: R1 R2 R3 because EQUALS says to keep records with the same key (duplicate records) in their original order. However, if we sort these three records on positions 1-4 with NOEQUALS with DFSORT the output can have those records in any order, e.g. R2 R3 R1 or R3 R1 R2 because NOEQUALS says the order of records is not guaranteed. So, with NOEQUALS it is difficult to compare the outputs. If you really want to compare, then you need to have EQUALS and then both products will give you the same results. I guess this is what has caused the notion that the two products behave differently from your earlier migration efforts. Further if you have any questions, please let me know. Thanks, Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Fwd: Syncsort > DFsort migration
The only issue I ran into (this was years ago) was that if you use the PARM in Syncsort of “EQUAL” expect different output in DFSORT.. We did not run into this but a friend did. There are items in Syncsort that are different in DFSORT. Syncsort made the decision that some of the (I do not have a manual in front of me but certain commands do not work the same or do not work in DFSORT) so be careful. Ed > Begin forwarded message: > > From: Richard McIntosh > Subject: Syncsort > DFsort migration > Date: August 26, 2023 at 6:52:58 AM CDT > To: IBM-MAIN@LISTSERV.UA.EDU > Reply-To: IBM Mainframe Discussion List > > We are starting to do a very rushed syncsort to DFsort migration. > I am assuming most of the sort ports are identical, but I bet not all of them. > Does anyone have a list of parms that need to be looked at? > Or anything thing else that I should be aware of. > > Thanks in advance for your time in replying. > > Richard M > > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
>>., an application sort execution, such as SAS-invoked PROC SORT under DFSORT >>does not exploit zSORT/SORTL (SAS INSTITUTE responded "not planned >>enhancement, and customers are not asking for it either.") Scott, Program invoked Sorts can exploit zSORT/SORTL if DFSORT can do the reads and writes. So, passing SORTDD= via parameter list will help to ensure that you can use zsort if you don't run into any other restrictions. >- unsure about how a "new" DFSORT site Any specific reason for using the word "new" ? DFSORT site has been the same for a long time. https://www.ibm.com/support/pages/dfsort or ibm.com/storage/dfsort ( will be re-directed to first link) is generally updated with every release or with any new enhancements. >> converting from use of PGM=SYNCSORT, PGM=SYNCGENR, PGM=SYNCTOOL, will "under >> the covers via a program alias" will get the DFSORT-equivalent - surely >> there must be some documented knob-turning with DFSORT >> install/implementation at IBM.COM ?? DFSORT doesn't have "under the cover" aliases for the other product, but will provide the necessary JCL to create the alias if requested. Thanks, Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
On Sat, 26 Aug 2023 11:52:58 +, Richard McIntosh wrote: >We are starting to do a very rushed syncsort to DFsort migration. >I am assuming most of the sort ports are identical, but I bet not all of them. >Does anyone have a list of parms that need to be looked at? >Or anything thing else that I should be aware of. > >Thanks in advance for your time in replying. > >Richard M Some relevant points/considerations/concerns: - SYNCSORT (now SYNCSORT MFX re-branded with Precisely marketing over-reach) use of /$ORTPARM DD (e.g., ELAP, VSCORE/VSCORET, others) and DFSORT uses //SORTCNTL DD or //DFSPARM DD - clearly may involve external JCL-changes. - DFSORT generates enhanced SYSOUT diagnostics when //SORTDIAG DD DUMMY coded; otherwise with SYNCSORT uses $ORTPARM DD control parameters - "out of the box" DFSORT ICEMAC definitions pretty well tuned; otherwise unsure about current-methodology with SYNCSORT. - SYNCSORT (namely ZPSaverSuite) may have zIIP-offload exploitation; otherwise, DFSORT not-so-much - an IBM-declared statement, presumed that the z15/z16 zSORT/SORTL on-chip approach taken, but only "compliant" SORT-executions (e.g., an application sort execution, such as SAS-invoked PROC SORT under DFSORT does not exploit zSORT/SORTL (SAS INSTITUTE responded "not planned enhancement, and customers are not asking for it either.") - unsure about "native SORT" using SYNCSORT (more current release/version/maintenance) and z15/z16 zSORT/SORTL (Accelerator). - SYNCSORT may be implemented with SAS application PROC SYNCSORT; otherwise SAS PROC SORT using DFSORT calls for SAS CONFIG parameter SORTBLKMODE. - DB2 sort-operation considerations - unknown about if/how/where SYNCSORT might get invoked; otherwise, with DFSORT consider IBM PH03207 and later documentation available for z15/z16 zSORT/SORTL exploitation. - unsure about how a "new" DFSORT site, converting from use of PGM=SYNCSORT, PGM=SYNCGENR, PGM=SYNCTOOL, will "under the covers via a program alias" will get the DFSORT-equivalent - surely there must be some documented knob-turning with DFSORT install/implementation at IBM.COM ?? - z/OS address space control/mgmt of REGION / MEMLIMIT (job- and job-step level) influences how a HOST-SORT solution can exploit memory over disk for SORT-WORK operations. - With a DFSORT implementation, active SMF type 16 (especially pay attention to DFSORT complaining when it cannot exploit zSORT/SORTL for whatever reason); otherwise, SYNCSORT has some near-equivalent sort invocation/performance data source when implemented with the SMF=(ON,nnn|208) parameter. - Likely both SYNCSORT and DFSORT behave as-expected when SORT-WORK allocations are made dynamically; otherwise consider HOST-SORT behavior when JCL-coded SORT-WORK DD allocations are made/used (e.g., DSNTYPE=LARGE, no VOL-ADD support for either DFSORT or SYNCSORT. Again, coding //SORTDIAG DD DUMMY will enhance the diagnostic-messages output sent to //SYSOUT DD for DFSORT operations (unsure if SYNCSORT has a special DD-name opportunity or otherwise is only driven with $ORTPARM DD override with MSG= parameter coding technique.) - by all means affordedexploit / engage available IBM/ISV/vendor SUPPORT services to ensure your solution implementation is as optimized as possible -- fortunately both IBM DFSORT and PRECISELY SYNCSORT technical support teams are always very helpful, in my past experience. Scott Barry SBBTech LLC -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Syncsort > DFsort migration
>> We are starting to do a very rushed syncsort to DFsort migration. Richard, Thank you for your interest in migrating to DFSORT. I will send you an offline mail with a couple of attachments which describes in detail about the parms and the needed changes. Thanks, Sri Hari Kolusu DFSORT Development IBM Corporation -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Syncsort > DFsort migration
We are starting to do a very rushed syncsort to DFsort migration. I am assuming most of the sort ports are identical, but I bet not all of them. Does anyone have a list of parms that need to be looked at? Or anything thing else that I should be aware of. Thanks in advance for your time in replying. Richard M -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN