IBM Mainframe Forum

Hi ... I don know whther this is the right area to post this?
Hoping somebody will help me out....

I need to know how can i calculate MIPS consumption of Individual ID in Mainframe?
Give me some link that might meet my requirement.....
Because at lot of times we run jobs that consumes a lot of mips and we are in a situation to pay a lot....

How to calculate MIPS?
Can Jobs be used to find it?
Or only DBA can do that?

Pls help me out...........

The MIPS rating is based on processor type and Operating System. Performance guru Cheryl Watson offers several charts at http://www.watsonwalker.com/chart.html
I'm not sure why you care about CPU MIPS rating because chargeback is usually by CPU time or Service Units used. Your job will run in less time on a faster CPU but the chargeback per second is usually more.

Likewise, fewer Service Units are used on a faster CPU.

Using less CPU might involve application tuning or changing process flow.

It ha been brought has a strict rule that we should not be consuming Mips ...

it will better if i know how to find this mips

By MIPS we can only assume that mean CPU resource ?

Hello,

It ha been brought has a strict rule that we should not be consuming Mips ...

Someone is telling you not to use excessive cpu cycles.

You don't consume mips. . . MIPS is a fairly meaningless measurement of cpu speed.

I suspect that there is code that is poorly written and uses more system resources than acceptable.

k.... atleast may i know how it is calculated .....

How can i know my job is consuming these many MIPS?

So that i can change my logic.

Hello,

There is no calculation for mips that i know of - this is a rating or measurement of cpu "power". The vendor tells the client how many mips a cpu is rated.

If you run batch jobs, the cpu usage is shown in the spool.

If your system uses a performance monitor (like Strobe), the system may help isolate bottlenecks.

Often, when code uses excessive resources, the design needs to be changed. You may want to review the problem module(s) with someone senior on your project or possibly your dba.

Hi,
Probably I understand your problem. In our shop too, they would use MIPS to say CPU usage "This job is using too much MIPS" ...

What you actually mean is CPU time. I am not sure if you will have the means to know the CPU usage by user, your opps team may get you that. But on a day to day basis you can check the jobs in spool to get that::

n JESMSGLG of a completed job



You can see the stepwise CPU usage in min (.00, .01, .03 in this case) though most quick steps will show 0.00 and at the end the total CPU usage 0.08min in this case, though the total execution time was 2.98 min.


Also If you do DA (Active users) in SPOOL while the job is in EXECUTION you monitor the CPU usage, identify if it is looping & cancel it if required. This is specially helpful if you setup allows your to code JCL TIME parameters.

Hope it helps,

BChat

Hello,

In our shop too, they would use MIPS to say CPU usage "This job is using too much MIPS" ...

Repeated, incorrect use of a term does not make it correct. . .

We would all be better served if "they" would be taught proper technical terms. There is already way too much mis-communication and mis-using terms only adds to this.