Regarding Background job vs Performance.

[Former Member]

Hi all,

I ran 2 similar programs in background through SE30 for comparing the performance.In SM37 the Duration(sec.) for the 2nd program is 3 times less than the first program after execution.But runtime analysis shows the performance of the first program is better than the second program by 1.5 times. Why is this difference!

My question here is, can we consider the time duration shown in SM37 for performance comparison? Why do we get different results in SM37 and SE30?

What factors affect the background process?

Please clarrify.

Regards,

Venkatesh Kumar S.

[Former Member]

Hi all,

I ran 2 similar programs in background through SE30 for comparing the performance.In SM37 the Duration(sec.) for the 2nd program is 3 times less than the first program after execution.But runtime analysis shows the performance of the first program is better than the second program by 1.5 times. Why is this difference!

My question here is, can we consider the time duration shown in SM37 for performance comparison? Why do we get different results in SM37 and SE30?

What factors affect the background process?

Please clarrify.

Regards,

Venkatesh Kumar S.

[christian_wohlfahrt]

Hi,

based on which number you come up with 1.5 times better? Total runtime or just percentages? I guess you had buffering for the 2nd job -> DB time got lower and you had a different %.

What are your numbers?

Regards,

Christian

[Former Member]

Hi,

In SM37, the duration(in secs) of 1st program is 3086 and 2nd program is 1112.

In SE30 the execution time ( Abap + database + R/3 (in microsecs) ) for the 1st program is 321889 and 2nd program is 438381. This is run for huge amount of data.If ran for less data 1st program takes less time in both SM37 and SE30.Why so?Please clarify.

Thanks for your help.

Regards,

Venkatesh.

[Former Member]

Hi,

In SM37, the duration(in secs) of 1st program is 3086 and 2nd program is 1112.

In SE30 the execution time ( Abap + database + R/3 (in microsecs) ) for the 1st program is 321889 and 2nd program is 438381. This is run for huge amount of data.If ran for less data 1st program takes less time in both SM37 and SE30.Why so?Please clarify.

Thanks for your help.

Regards,

Venkatesh.

[christian_wohlfahrt]

Hi Venkatesh!

You have overflows. The total in SE30 is just 0.3 sec and 0.4 sec - but in SM37 3086 and 1112 seconds. After a conversion into microseconds, you would have 3 086 000 000 - which is obviously to much.

For such long times, you can only look on the SM37 times - if you like to get more details, make info messages - they will be in the job log with a time stamp and you can inform about 'data read', 'data converted', 'data output' or something like this.

Regards,

Christian

[Former Member]

se30 is for the performance analysis.. but sm37 is the execution time for background processing which may be effected by many other parameters.

so for performance analysis you should always go for SE30 instead of sm37

[Former Member]

Hi,

Go through all performanace issues.

i think the second program having large records.