Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uunet!brunix!twl From: twl@brunix (Ted "Theodore" (W) Leung) Newsgroups: comp.sys.next Subject: Re: Will the Next sell? (max memory size) Message-ID: <21909@brunix.UUCP> Date: 30 Nov 89 19:26:48 GMT References: <4283@helios.ee.lbl.gov> <21301@uflorida.cis.ufl.EDU> <21736@brunix.UUCP> <21795@brunix.UUCP> Sender: news@brunix.UUCP Reply-To: twl@boojum.UUCP (Ted "Theodore" (W) Leung) Distribution: na Organization: Brown University Department of Computer Science Lines: 29 In article dennis@yang.cpac.washington.edu (Dennis Gentry) writes: > > I'd be all for a 25+MIP machine with >32MB of physical memory..... > >It's not clear that 64Mb is really a reasonable amount of RAM >for a 68030. Even though 64Mb costs twice as much as 32Mb, the >performance gain you'll achieve by going from 32Mb to 64Mb will >almost certainly be less impressive than the performance gain >from 16 to 32Mb. I wasn't trying to debate the feasibility of installing that much memory in a machine. Our SPARCstations have the same capability to accept the 4MB devices. What I was saying was that for some of the applications that I am running, the code wants to allocate or preallocate sufficiently large amounts of memory that 32+MB of real memory would cut down on the paging substantially, especially when running a number of these large programs concurrently. Ideally, I'd like to be able to load up my set of working applications, get them all into main memory, and page as little as possible. Why is it unclear whether 64MB of memory is reasonable for use with a 68030? I'm not intimately familiar with the details of the architecture, but it seems to me that this is more a function of the operating system and the set of applications that are being run. -------------------------------------------------------------------- Internet/CSnet: twl@cs.brown.edu | Ted "Theodore" Leung BITNET: twl@BROWNCS.BITNET | Box 1910, Brown University UUCP: uunet!brunix!twl | Providence, RI 02912