Return to BSD News archive
Newsgroups: comp.unix.bsd Path: sserve!manuel.anu.edu.au!munnari.oz.au!spool.mu.edu!wupost!zaphod.mps.ohio-state.edu!darwin.sura.net!paladin.american.edu!news.univie.ac.at!hp4at!mcsun!Germany.EU.net!rrz.uni-koeln.de!unidui!flyer!flatlin!bad From: bad@flatlin.ka.sub.org (Christoph Badura) Subject: Re: cache terms (was Adding Swapspace ??) Organization: Guru Systems/Funware Department Date: Tue, 27 Oct 1992 02:44:37 GMT Message-ID: <BwrEAE.8vB@flatlin.ka.sub.org> References: <Bw7H4L.LLB@cosy.sbg.ac.at> <1992Oct16.162729.3701@ninja.zso.dec.com> <1992Oct16.201806.21519@fcom.cc.utah.edu> <Bw8Mw5.IFC@pix.com> <1992Oct18.082017.22382@fcom.cc.utah.edu> <BwLLxp.7Bt@flatlin.ka.sub.org> <1992Oct25.111525.25782@fcom.cc.utah.edu> <26965@dog.ee.lbl.gov> <1992Oct25.224950.3098@fcom.cc.utah.edu> Lines: 43 In <1992Oct25.224950.3098@fcom.cc.utah.edu> terry@cs.weber.edu (A Wizard of Earth C) writes: >Of course, this still leaves the issue of `write behind' unresolved; is it >synonymous with `write through' or `write back'? The latter. >Since I think >we can all agree that a block of data has to go to memory before that memory >is written to disk, then the operation is either `write through' or `write >behind' based on your definition of simultanaity (something can't be going >into a bounce buffer and written at the same time). Bounce buffers don't come into play. A cache is "write through", iff every write to the cache _immediatly_ solicites a write to secondary storage. It is "write behind" or "write back" if the flush tu secondary storage is triggered by other events (e.g. cache coherency protocols, periodic flushes, space shortage in the cache itself, etc.). I really don't understand your remars about buffer cache, swapping and swapping to files. Clearly swapping to files is better than no swapping at all or insufficient swap space when you can't allocate an additional swap partition. On the other hand, the swapping to devices can make potential use of the buffer cache. In that case "swap buffers" should have a different priority than "file buffers." Which should be higher I don't know. But I would guess that "file buffers" should have priority over "swap buffers" to improve file system performance when memory performance can't be improved. There are some caveats about unifying VM and buffer cache (as in SunOS 4.x), as large memory requirement can have undesirable performance drops in file system throughput and vice versa. -- Christoph Badura --- bad@flatlin.ka.sub.org AIX is a better... is a better... is a better... OpenSystem. IBM Rep at GUUG Symposium '92