The following message is a courtesy copy of an article that has been posted to bit.listserv.ibm-main,alt.folklore.computers as well.
"Bill Ogden" <[EMAIL PROTECTED]> writes: > The statements about the 360/67 are correct. It was a little ahead of > its time in several ways. The 67's DAT design was a bit different than > the later S/370 DAT that was used by MVS, and is typically not > considered in the history lines for MVS. re: 2007p.html:Subject: Re: GETMAIN/FREEMAIN and virtual storage backing up 2007p.html:Subject: Re: GETMAIN/FREEMAIN and virtual storage backing up 2007p.html:Subject: Re: GETMAIN/FREEMAIN and virtual storage backing up other than original os/vs2 prototype implementation was done with mvt kernel modified with a lot of code borrowed from cp67 running on 360/67 i had done a lot of work with virtual memory as an undergraduate http://www.garlic.com/~lynn/subtopic.html#wsclock and then later after joining the science center http://www.garlic.com/~lynn/subtopic.html#545tech and in the early 70s several of us would make frequent sojourns to pok (out the mass pike and down the taconic) for architecture meetings (virtual memory, multiprocessing, etc) ... including architecture meetings where several features were pulled from 370 virtual memory architecture in order to buy 370/165 engineers six month schedule in their hardware implementation. there were other issues in the os/vs2 virtual memory implementation (spanning both svs and mvs) ... one had to do with the page replacement algorithm implementation ... the standard is LRU (least recently used) or various approximations related of LRU. The pok performance modeling group had discovered that (at a micro-level) that if a non-changed page was selected for replacement ... that the latency to service a page fault was much less than if a changed page was selected for replacement (non-changed pages could be immediately discarded, without needing to write, relying on copy already out on disk). However, i repeatedly pointed out to them that weighting the replacement algorithm based on changed bit as opposed to the reference bit ... severely negated any recently used strategy. They went ahead with it anyway (possibly they didn't have very good macro-level simulation capability and stuck with just the micro-level simulation could make informed judgement). in any case, it was well into a number of MVS release before somebody got an award for improving MVS performance by changing to give more weight to the reference use in replacement decisions (example was that under the earlier strategy, the replacement algorithm was selecting high-use, shared, executable linklib virtual pages for replacement before private, lower-use application data virtual pages). another influence of cp67 and the science center was a joint project between endicott and the science center to do custom modifications to cp67 to provide "370" (virtual memory architecture) virtual machines. For instance, this required cp67 simulating 370 architecture hardware format virtual memory tables ... rather than 360/67 architecture hardware format virtual memory tables ... internally, this was comingly referred to as "cp67h" system. After that was done, there were modifications to cp67 to make it run on 370 hardware ... building 370 format tables ... rather than 360/67 format tables. Internally, this was comingly referred to as cp67i. The first operational 370 hardware supporting virtual memory was a 370/145 engineering processor. However, cp67h with cp67i running in a 370 virtual machine was in regular operation a year before the 370/145 engineering box was operational. In fact, cp67i system was used as initial software brought up on the 370/145 engineering box. One of the complexities in the cp67h & cp67i development was it was all done on the science center cp67 timesharing service. Information about virtual memory for 370 was an extremely tightly held corporate secret ... and there were a variety of non-employees (from numerous education institutions in the cambridg area) with regular access to the science center timesharing service. As a result ... nearly all of the cp67h work went on in a 360/67 virtual machine (not on the bare hardware) to isolate it from any non-employee prying eyes. lots of past posts about use of cp67 for timesharing service ... both internally and externally (including mentioning it being used to address various security issues) http://www.garlic.com/~lynn/subtopic.html#timeshare misc past posts mentioning cp67h and/or cp67i systems: http://www.garlic.com/~lynn/2002j.html#0 HONE was .. Hercules and System/390 - do we need it? http://www.garlic.com/~lynn/2004b.html#31 determining memory size http://www.garlic.com/~lynn/2004h.html#27 Vintage computers are better than modern crap ! http://www.garlic.com/~lynn/2004p.html#50 IBM 3614 and 3624 ATM's http://www.garlic.com/~lynn/2005c.html#59 intel's Vanderpool and virtualization in general http://www.garlic.com/~lynn/2005d.html#66 Virtual Machine Hardware http://www.garlic.com/~lynn/2005g.html#17 DOS/360: Forty years http://www.garlic.com/~lynn/2005h.html#18 Exceptions at basic block boundaries http://www.garlic.com/~lynn/2005i.html#39 Behavior in undefined areas? http://www.garlic.com/~lynn/2005j.html#50 virtual 360/67 support in cp67 http://www.garlic.com/~lynn/2005p.html#27 What ever happened to Tandem and NonStop OS ? http://www.garlic.com/~lynn/2006.html#38 Is VIO mandatory? http://www.garlic.com/~lynn/2006e.html#7 About TLB in lower-level caches http://www.garlic.com/~lynn/2006f.html#5 3380-3390 Conversion - DISAPPOINTMENT http://www.garlic.com/~lynn/2006l.html#21 Virtual Virtualizers http://www.garlic.com/~lynn/2006m.html#26 Mainframe Limericks http://www.garlic.com/~lynn/2006o.html#19 Source maintenance was Re: SEQUENCE NUMBERS http://www.garlic.com/~lynn/2006q.html#1 Materiel and graft http://www.garlic.com/~lynn/2006q.html#45 Was FORTRAN buggy? http://www.garlic.com/~lynn/2006q.html#49 Was FORTRAN buggy? http://www.garlic.com/~lynn/2006w.html#3 IBM sues maker of Intel-based Mainframe clones http://www.garlic.com/~lynn/2007b.html#20 How many 36-bit Unix ports in the old days? http://www.garlic.com/~lynn/2007f.html#12 FBA rant http://www.garlic.com/~lynn/2007i.html#16 when was MMU virtualization first considered practical? ---------------------------------------------------------------------- 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
