Improve time with a large file



IBM's flagship sort product DFSORT for sorting, merging, copying, data manipulation and reporting. Includes ICETOOL and ICEGENER

Improve time with a large file

Postby tsdjim » Tue Dec 27, 2011 12:49 pm

I have a COBOL program using INPUT-PROCEDURE OUTPUT-PROCEDURE to do an internal sort (z/OS DFSORT 1.10) for around 7 million records and I am trying to get a performance improvement. The program is not compiled with FASTSRT. The job JCL has REGION=0M

Operations claim that with our previous sort product, it took less than half the time and used just one SORTWK file, whereas with DFSORT now it requires 22.

I have attached the SORT log below. Is there anything there that may give a clue on performance improvement and reduce SORTWK files.

- 21.30 MINUTES EXECUTION TIME
22 //SORTWK01 DD DSN=&&TEMP01,UNIT=3390,SPACE=(CYL,(3200,50),RLSE),
// DISP=(,DELETE)
//SORTWK02....... etc
43 //SORTWK22 DD DSN=DCBK.T4,UNIT=3390,SPACE=(CYL,(3100,50),RLSE),
// DISP=(,DELETE)
44 //SORTDIAG DD DUMMY
45 //OTSFMEX DD DSN=DBP2.SFF005.DA.RPTTEST,DISP=(,CATLG,DELETE),
// UNIT=3390,SPACE=(CYL,(600,150),RLSE),DATACLAS=COMPS,
// DCB=(RECFM=VB,LRECL=7004,BLKSIZE=0)
46 //SYSTSIN DD *
47 //SYSIN DD *
48 //IGZSRTCD DD *
OPTION FILSZ=E7000000,AVGRLEN=7006,MAINSIZE=MAX,DYNSPC=512

1ICE805I 0 JOBNAME: SFJF10DX , STEPNAME: SFAF105P
ICE802I 0 BLOCKSET TECHNIQUE IN CONTROL
ICE201I G RECORD TYPE IS V - DATA STARTS IN POSITION 5
ICE992I 0 RA 0 WR 0 TR 0
ICE751I 0 C5-K51707 C6-K51707 C7-K54603 C8-K51707 E4-K51707 C9-BASE E5-K55337 B0-K48417 B0-K48417 E7-K51707
ICE143I 0 BLOCKSET SORT TECHNIQUE SELECTED
ICE250I 0 VISIT http://www.ibm.com/storage/dfsort FOR DFSORT PAPERS, EXAMPLES AND MORE
ICE000I 0 - CONTROL STATEMENTS FOR 5694-A01, Z/OS DFSORT V1R10 - 07:29 ON TUE DEC 27, 2011 -
0 SORT FIELDS=(0083,0002,CH,A,0011,0042,CH,A)
RECORD TYPE=V,LENGTH=(007006,,,000005,)
OPTION FILSZ=E7000000,AVGRLEN=7006,MAINSIZE=MAX,DYNSPC=512
ICE201I G RECORD TYPE IS V - DATA STARTS IN POSITION 5
ICE992I 0 RA 0 WR 0 TR 0
ICE751I 0 C5-K51707 C6-K51707 C7-K54603 C8-K51707 E4-K51707 C9-BASE E5-K55337 B0-K48417 B0-K48417 E7-K51707
ICE193I 0 ICEAM2 INVOCATION ENVIRONMENT IN EFFECT - ICEAM2 ENVIRONMENT SELECTED
ICE089I 6 SFJF10DX.SFAF105P. , INPUT LRECL = 7006, TYPE = V
ICE093I 0 MAIN STORAGE = (MAX,67108864,67108864)
ICE156I 0 MAIN STORAGE ABOVE 16MB = (67011824,67011824)
ICE127I 0 OPTIONS: OVFLO=RC0 ,PAD=RC0 ,TRUNC=RC0 ,SPANINC=RC16,VLSCMP=N,SZERO=Y,RESET=Y,VSAMEMT=Y,DYNSPC=512
ICE128I 0 OPTIONS: SIZE=67108864,MAXLIM=1048576,MINLIM=450560,EQUALS=Y,LIST=Y,ERET=RC16 ,MSGDDN=SYSOUT
ICE129I 0 OPTIONS: VIO=N,RESDNT=NONE,SMF=NO ,WRKSEC=Y,OUTSEC=Y,VERIFY=N,CHALT=N,DYNALOC=N ,ABCODE=MSG
ICE130I 0 OPTIONS: RESALL=4096,RESINV=0,SVC=109 ,CHECK=Y,WRKREL=Y,OUTREL=Y,CKPT=N,COBEXIT=COB2
ICE131I 0 OPTIONS: TMAXLIM=6291456,ARESALL=0,ARESINV=0,OVERRGN=16384,CINV=Y,CFW=Y,DSA=64
ICE132I 0 OPTIONS: VLSHRT=N,ZDPRINT=Y,IEXIT=N,TEXIT=N,LISTX=N,EFS=NONE ,EXITCK=S,PARMDDN=DFSPARM ,FSZEST=N
ICE133I 0 OPTIONS: HIPRMAX=OPTIMAL,DSPSIZE=MAX ,ODMAXBF=0,SOLRF=Y,VLLONG=N,VSAMIO=N,MOSIZE=MAX
ICE235I 0 OPTIONS: NULLOUT=RC0
ICE750I 0 DC 0 TC 0 CS DSVXX KSZ 48 VSZ 48
ICE752I 0 FSZ=7000000 RE IGN=0 C AVG=7006 7006 WSP=63697130 E DYN=0 0
ICE915I 0 MOFSZ=62204,MOSZ=1418,MOSYS=1578(5),MOSTG=1578,MEML=17592186040320(3)
ICE880I 0 QP=89 QA=89 HI=99 LI=11 MI=99 TZ=363010 N1=363008 N2=363008 SZ=90 HN=1
ICE916I 0 MOFR=0002,MOVR=D0
ICE996I 0 ESM=762624,ESO=762624,ESR=0,ESP=4096,ESS=16384,CES=762624,HSZ=524288
ICE997I 0 HWSP=29932861,HMAX=403970,HES=0,ASV=403970,EQ=I1,HN=0
ICE898I 0 OMAX=213783,NMAX=403970,TMAX=403970,CMAX=0,HU=0,BUN=0,MD=N6,DU=0,DR=0,HN=0
ICE889I 0 CT=MAX , SB=3, L=0, D=0000, CCW=1MAM
ICE901I 0 W 01NP17 02NP17 03NP17 04NP17 05NP17 06NP17 07NP17 08NP17
ICE901I 0 W 09NP17 10NP17 11NP17 12NP17 13NP17 14NP17 15NP17 16NP17
ICE901I 0 W 17NP17 18NP17 19NP17 20NP17 21NP17 22NP17
ICE751I 1 B1-K47755 B2-K47755 FA-K47755 B3-K47755 EA-BASE F1-K38900 E8-K51707
ICE091I 0 OUTPUT LRECL = 7006, TYPE = V
ICE055I 0 INSERT 6891774, DELETE 6891774
ICE054I 0 RECORDS - IN: 0, OUT: 0
ICE134I 0 NUMBER OF BYTES SORTED: 48283768644
ICE253I 0 RECORDS SORTED - PROCESSED: 6891774, EXPECTED: 7000000
ICE098I 0 AVERAGE RECORD LENGTH - PROCESSED: 7006, EXPECTED: 7006
ICE165I 0 TOTAL WORK DATA SET TRACKS ALLOCATED: 896250 , TRACKS USED: 853515
ICE199I 0 MEMORY OBJECT STORAGE USED = 1418M BYTES
ICE180I 0 HIPERSPACE STORAGE USED = 0K BYTES
ICE188I 0 DATA SPACE STORAGE USED = 0K BYTES
ICE891I 1 67057192 WMAIN, 11992 CMAIN, MAX CALLOC, N SCN, B BA, 0 AZ, 0 BZ, NN QC, 0 CZ, 0 DZ, 1 PLE
ICE892I 1 7006 RIN 0 BLI 7010 BLO 7006 RUN 56664 BUN 2097 CPU 00 CVC
ICE893I 1 0 XIN 510 WIN 197119 GIN NDEY PFP3 B00 CM000 CIX UPTH LMV VS RUX
ICE894I 0 12 STR 12 MOR 1052 IPB 0 OPB 56901 CYL O MN
ICE881I 1 EQ=I1 DX=0 D2=0 D3=0 D4=89 AS=0 SA=0 SB=0 SC=0 HN=1
ICE896I 0 57009 SET 2563586 DEXTOT 852478 BLK 7000000 CSZ 115129 WE
ICE804I 1 SORTWK22 EXCP COUNT: 6096
ICE804I 1 SORTWK21 EXCP COUNT: 6087
ICE804I 1 SORTWK20 EXCP COUNT: 6089
1ICE804I 1 SORTWK19 EXCP COUNT: 6095
ICE804I 1 SORTWK18 EXCP COUNT: 2329
ICE804I 1 SORTWK17 EXCP COUNT: 2328
ICE804I 1 SORTWK16 EXCP COUNT: 2326
ICE804I 1 SORTWK15 EXCP COUNT: 2327
ICE804I 1 SORTWK14 EXCP COUNT: 2332
ICE804I 1 SORTWK13 EXCP COUNT: 6085
ICE804I 1 SORTWK12 EXCP COUNT: 6087
ICE804I 1 SORTWK11 EXCP COUNT: 6083
ICE804I 1 SORTWK10 EXCP COUNT: 6084
ICE804I 1 SORTWK09 EXCP COUNT: 6087
ICE804I 1 SORTWK08 EXCP COUNT: 6091
ICE804I 1 SORTWK07 EXCP COUNT: 6081
ICE804I 1 SORTWK06 EXCP COUNT: 6087
ICE804I 1 SORTWK05 EXCP COUNT: 6091
ICE804I 1 SORTWK04 EXCP COUNT: 6084
ICE804I 1 SORTWK03 EXCP COUNT: 6088
ICE804I 1 SORTWK02 EXCP COUNT: 6082
ICE804I 1 SORTWK01 EXCP COUNT: 6090
ICE052I 0 END OF DFSORT
tsdjim
 
Posts: 18
Joined: Mon Nov 03, 2008 1:38 am
Has thanked: 0 time
Been thanked: 0 time

Re: Improve time with a large file

Postby BillyBoyo » Tue Dec 27, 2011 2:05 pm

If you have the SORT in the Cobol program and you want it to run faster, why do you not use FASTSRT? I think you'll see a big difference in that, as, as far as I understand it, DFSORT will be doing the IO for the files, not Cobol (via its routines) and DFSORT is considerably faster at that sort of thing.

I don't know why anyone would be worried about how many work datasets a particular product happens to use.

You have LRECL 7006, pre-determined average record-length 7006 and calculated average record-length 7006. Are you sure your file is actually variable?
BillyBoyo
Global moderator
 
Posts: 3804
Joined: Tue Jan 25, 2011 12:02 am
Has thanked: 22 times
Been thanked: 265 times

Re: Improve time with a large file

Postby tsdjim » Tue Dec 27, 2011 9:32 pm

I thought FASTSRT was only applicable when USING/GIVING was used , unless I am wrong.

I took the 7006 figure from the PROCESSED in the message
ICE098I 0 AVERAGE RECORD LENGTH - PROCESSED: 7006, EXPECTED: 7006

I mentioned the work datasets because, as I said, a previous sort product used only 1 sortwork for the same sort.
tsdjim
 
Posts: 18
Joined: Mon Nov 03, 2008 1:38 am
Has thanked: 0 time
Been thanked: 0 time

Re: Improve time with a large file

Postby skolusu » Tue Dec 27, 2011 11:35 pm

tsdjim,

Please send the complete sysout including JES messages to our performance expert at betten@us.ibm.com . He is on vacation right now, but he will check it once he gets back.
Kolusu - DFSORT Development Team (IBM)
DFSORT is on the Web at:
www.ibm.com/storage/dfsort
skolusu
 
Posts: 586
Joined: Wed Apr 02, 2008 10:38 pm
Has thanked: 0 time
Been thanked: 39 times


Return to DFSORT/ICETOOL/ICEGENER

 


  • Related topics
    Replies
    Views
    Last post