Shouldn't the SU's consumed or SU/SEC RMF numbers have some relation to MSU's? Granted IBM touts 'Software MSU's' rather than the 'Hardware MSU's' but if you take RMF SU/SEC available for the processor vs what an AS consumed you should be able to work out a reasonable number.
Perhaps Imanol Aguirre's question is just a variation on the old 'how may MIPS does my application take?'. Ken Porowski AVP Systems Software CIT Group E: [EMAIL PROTECTED] -----Original Message---- Kelman, Tom > -----Original Message----- > Imanol Aguirre > > Hi all, > > I am just new to subcapacity this year so we begun to work in terms of > MSUs. > I'm using RMF monitor III in conjuction with the RMF performance reporter > connected to it. > > With these tools we know certain MSU cunsumption values in the RMF > interval (100 seconds in my case) > MSU 4h Avg > Actual MSU > > Let's say that The "Actual MSU" value for the last 100 seconds is 60 MSUs > for > the entire LPAR. > My question is: There is a way to know the MSUs consumed by INDIVIDUAL > Adress spaces?? or do I have to work it out myself with complicated > spreadsheets? > > ---------------------------------------------------------------------- I'm a little bit at a loss here as to why you would want to determine the software MSUs for an individual address space. The MSUs you're talking about are just a way for IBM to hold down the cost of its software as the size of the processors increase. The ratio between this MSU value and processor power changes with each new processor version. They aren't really a very good tool for performance determination or capacity planning. Having said that, I don't know of anyway to determine the MSUs used by an individual address space except to do your own calculations. For example, we have a z9-BC of 1004 MIPS or 138 MSUs. Therefore 100% of the machine is 138 MSUs. I'd have to determine the CPU percent used by the address space and then multiply the by 138 to determine the MSUs used. I think that's correct. However, I think I have to think this out again, as Fagin would say. Tom Kelman ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
