Path: csiph.com!eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail From: Anne & Lynn Wheeler Newsgroups: alt.folklore.computers Subject: Re: Windows 11 is now available Date: Fri, 08 Oct 2021 08:35:34 -1000 Organization: Wheeler&Wheeler Lines: 48 Message-ID: <875yu74b61.fsf@localhost> References: <87zgrltu9u.fsf@quartz.kremvax.net> <20211007102537.52834272c0e1627486cf1d77@127.0.0.1> <2808cbbc-2334-4e34-a76c-46eee98e7526n@googlegroups.com> Mime-Version: 1.0 Content-Type: text/plain Injection-Info: reader02.eternal-september.org; posting-host="3f9b97ec28a8b50e365e18f610946b75"; logging-data="14968"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX182SXHFVclsExLMQ7vsMfw8SUaNLR2XH40=" User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/27.2 (gnu/linux) Cancel-Lock: sha1:Wq3WCDhttbuLpoZ5lXKIfr+1Ops= sha1:yP46toknLmGJrDFicVsWkwlf6XU= Xref: csiph.com alt.folklore.computers:219435 Branimir Maksimovic writes: > Well on unisys you could connect more then 100 terminals > on machine with only 4mb of ram, > because terminals had logic for input, and all that > was exchanged was just packets sent :P Original CP67 was installed at univ. in Jan1968 by three people from cambridge science center ... could play with it on weekends on 360/67 with 768kbytes memory. The TSS/360 IBM SE and I put together synthetic fortran, compile, and execute benchmark ... and CP67/CMS with 35 emulated users had better response time and throughput than TSS/360 (on same machine) with only four users. Later after rewritting a lot of CP67 (pathlengths, resource management, scheduling, "global LRU" page replacement, page I/O, ordered seek queuing, etc), graduating and joining the cambridge science center ... had 85 real users on 768kbyte (192 4k pages) 360/67 (104 4k pageable pages after fixed kernel and control block storage) had subsecond response and 100% CPU useage. This was important because the Grenoble Science Center had modified CP67 to have a "working set dispatcher" ("local LRU" page replacement, from 1968 academic papers) running on 1mbyte (256 4k pages) with 35 (real) users had worse response and lower CPU useage (aka "global LRU" page replacement easily beat "local LRU" page replacement). This came up more than decade later, Jim Gray and left IBM San Jose Research for Tandem and at SIGOPS (Asilomar, 14-16Dec81), he asked me if I could help co-worker at Tandem get his Stanford PHD (advisor was later president of Stanford) which involved "global LRU" (page replacement), and he knew I had down a lot of work on global LRU page replacement algorithm and had apple-to-apple comparison between local and global LRU (as undergraduate in the 60s). Some of the "local LRU" forces (dating back to 60s when I was doing "global LRU") were heavily lobbying Stanford to block awarding PHD involving global LRU. When I went to send information ... SJR management said that I wasn't allowed to (even tho none of the information involved anything after joining IBM). I've commented that I hoped that it was done as punishment for online computer conferencing ("Tandem Memos", in the late 70s I was blamed for online computer conferencing on the internal network, folklore is when corporate executive committee was told about it, 5of6 wanted to fire me) ... rather than they taking part in the global/local LRU academic dispute. I wasn't allowed to send reply for nearly a year http://www.garlic.com/~lynn/2006w.html#email821019 -- virtualization experience starting Jan1968, online at home since Mar1970