Robert Wessel <[email protected]> writes: > VSE was similar (although probably DOS/VSE in that day). The VM "page > fault extensions" allowed a guest to be dispatched when VM handled a > page fault. > > And like VS1,VSE in PAGE=VM* mode generally ran better than VSE > native. Which really says something about the quality of virtual > memory support in VSE (and VS1).
part of the issue was i had a much better page replacement algorithm with a significantly shorter pathlength (not just VS1/VSE, but also MVS) ... originally done while i was undergraduate in the 60s for cp67. lots of past posts http://www.garlic.com/~lynn/subtopic.html#clock the other part was that DOS & VS1 used 2k virtual pages ... that allowed for for more efficient compaction in small real storage (compared to 4k virtual pages) ... but by the time of 138/148 nominal real storage was 1mbyte (and increasing) ... so it was no longer a useful trade-off. note one of the page replacement issues was global lru versus local lru. I had done global lru as undergraduate in the 60s about the same time there was some work on local lru published academically. At Dec81, Asilomar ACM SIGOPS, Jim Gray (I had worked with a ibm research, but had left for Tandem), asked me if i could help a co-worker get his Stanford PHD ... it involved work on global LRU similar to what i had done as undergraduate in the 60s ... but his thesis advisor was getting enormous pushback from person behind local lru. I had lots of data directly comparing global lru to local lru replacement algorithms, both implemented on cp67 running on 360/67, showing global lru significantly outperforming local lru. past post on the subject http://www.garlic.com/~lynn/2006w.html#46 unfortuantely ibm management blocked me sending a replay for nearly a year http://www.garlic.com/~lynn/2006w.html#email821019 I was being blamed for online computer conferencing on the internal network in the late 70s and early 80s ... folklore is that when the executive committee (chairman, ceo, pres, etc) was told about online computer conferencing (and the internal network), 5of6 wanted to fire me. In any case, hopefully IBM management blocking my sending a reply was somehow viewed as punishment for online computer conferencing (rather than taking sides in the academic dispute). trivia ... his thesis advisor went on later to be president of stanford. posts in thread: http://www.garlic.com/~lynn/2014d.html#54 Difference between MVS and z / OS systems http://www.garlic.com/~lynn/2014d.html#55 Difference between MVS and z / OS systems http://www.garlic.com/~lynn/2014d.html#56 Difference between MVS and z / OS systems http://www.garlic.com/~lynn/2014d.html#57 Difference between MVS and z / OS systems http://www.garlic.com/~lynn/2014d.html#59 Difference between MVS and z / OS systems http://www.garlic.com/~lynn/2014d.html#60 Difference between MVS and z / OS systems http://www.garlic.com/~lynn/2014d.html#61 Difference between MVS and z / OS systems http://www.garlic.com/~lynn/2014d.html#62 Difference between MVS and z / OS systems http://www.garlic.com/~lynn/2014d.html#63 Difference between MVS and z / OS systems http://www.garlic.com/~lynn/2014d.html#83 Difference between MVS and z / OS systems -- virtualization experience starting Jan1968, online at home since Mar1970 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
