Path: csiph.com!x330-a1.tempe.blueboxinc.net!newsfeed.hal-mli.net!feeder1.hal-mli.net!news.glorb.com!dotsrc.org!filter.dotsrc.org!news.dotsrc.org!not-for-mail From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Newsgroups: comp.sys.dec,comp.os.vms Subject: Re: RB730 Integrated Disk Controller (R80/RL02) usable with VAX-11/750? Message-ID: Date: 3 May 2011 12:36:03 -0500 References: <5a076b79-20b8-461d-93ee-e7ee8b50e249@dn9g2000vbb.googlegroups.com> <2ecc8832-029c-41a9-9f97-e4da0238ca4d@l14g2000pro.googlegroups.com> Organization: Encompasserve Lines: 13 NNTP-Posting-Host: 67.53.90.116 X-Trace: news.sunsite.dk DXC=kfLTc6^hV81:YiKBaBLXD=YSB=nbEKnk;??bK;=RO3Q:T1og, Johnny Billquist writes: > > Not entirely true. Most the kernel needs to be in physical memory (even > if it is in virtual address space). It's just user processes that use > demand paging. (Virtual memory itself only means you are remapping one > set of addresses to another set, it does not reduce the memory > requirements at all, that's what the demand paging does for you.) Some of the kernel needs to be in RAM. Although it's not particularly great for performance, you can tune VMS to get non-0 kernel page faults for the rest. And not all of the VMS executive is in the kernel.