(Originally posted 2005-11-11.)
This probably counts as my contribution to RMF for this decade. 🙂
RMF has long had the Software Model
for a processor in SMF 70. Until z900 this was the same as the Hardware Model
. With z900 you could have processors that were there but you hadn’t bought yet.
With z990 it got a lot worse
as you could have between 1 and 4 books on the processor.
In z990 (and z9-109) studies we would know, for example, that the processor was a z990 Model 324 but wouldn’t be absolutely sure whether it was a 3-book or a 4-book machine. With a 324 there’s nowhere to go without installing a fourth book – if you needed to add processors. And we, frustratingly wouldn’t know that.
Now, with Release 7 RMF we get the hardware model in addition to the software model. So we expect in future engagements to be a little more sensitive to what it would take to upgrade.
Actually here’s what you could do today – on Release 6 or prior…
The Logical Processor Sections in the SMF 70 record do record under the PHYSICAL
LPAR the processor numbers. Here’s what I saw for the above-mentioned z990 Model 324:
- 00,02,05,07,08,0A,0C,0E
- 20,22,25,27,28,2A,2C,2E
- 30,32,35,37,38,3A,3C,3E
I wrote those in 2-digit hex format for a reason: The first digit is the book number.
So where’s Book 1, then?
It doesn’t exist. So I conclude this is a C24 (3-book) machine.
If this was your machine you’d probably be saying Well, I know I bought a 3-book machine
but, in my consultancy, I think I’d like to know that, too.
One other thing: You can tell what the engine type is from these sections. In this case they’re all General-Purpose CPs (GCPs) and that corresponds to what we happen to know about the client: 2 12-way dedicated LPARs and no ICFs, IFLs or zAAPs.
Now none of that is an “intended interface” but it is fascinating. 🙂 But you are much better off with the 1.7 support for Hardware Model.
Mainframe, Performance, Topics 