And z/VM has managed cases like this forever. If you want to learn more,
suggest you go to some z/VM sessions at SHARE. Lots of folks there who can
speak knowledgably about this.

 

In case this helps, here's a simplified version of how it works: there are
segment and page tables. These indicate what pages exist and where
(in-memory, on backing store, not at all), and what segments (1MB? 16MB
these days? I forget! Been too long!)-bigger "chunks" of memory-exist. Those
can be sparse as well (since 2**64 pages would itself take a lot of memory
to describe-let's see, if a PTE (Page Table Entry) was just 4 bytes
(probably optimistic), then 2**64 bytes of memory would take
2**(64-12+4)=2**56 bits of memory just to describe, if I've done the math
right. That would be a lot.

 

Does this help?

 

.phsiii


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to