ftp.nice.ch/peanuts/GeneralData/Usenet/news/1989/CSN-89.tar.gz#/comp-sys-next/1989/Dec/64MB-in-cubes

This is 64MB-in-cubes in view mode; [Up]


Date: Sun 01-Dec-1989 05:11:54 From: Unknown Subject: 64MB in cubes In article <9374@batcomputer.tn.cornell.edu> rogerj@tcgould.tn.cornell.edu (Roger Jagoda) writes: >Lets all be sure we're right on this one. As a test, our group >bought 20 4 MB SIMM boards from the Chip Merchant. They worked >fine in a MACIIci (030, 25Mhz) but the Cube froze up solid and >refused to boot. 64MB of ram definitely works in a cube using 4 megabit DRAMS. We have several 4 MB SIMMs that we have used here in software engineering and most definitely tested a 64MB configuration. I have no idea why the SIMMs from the Chip Merchant didn't work. Perhaps there was a timing problem? The ones we have were built in our factory and work fine (I don't remember who the actual chip vendor was though). Although 4 MB SIMMs are not on our current price lists I'm sure if there was enough interest (and enough people willing to spend the still high costs for the 4 MB parts) we would probably make them available.
Date: Sun 01-Dec-1989 12:16:28 From: Unknown Subject: Re: 64MB in cubes If the 4 MBit SIMMS are freezing up the machine, you might be interested in knowing about the ROM-monitor 'm' command. It displays the current memory configuration. You can generally get into the ROM monitor no matter what happens. It is documented in chapter 17 of the System Reference Manual, and is available on-line. You should also be certain that you are running System version 1.0 and have correctly installed the system and hardware. As far as the utility of 64 MB with the 68030... it depends on the breakdown of memory usage. If 80% of the memory is used by one process (say a large data processing application), than you are probably best off with a single processor (unless you have a parallel app). If you are loading 100 small processes that split up that memory and are not communicating with eachother too much, then you probably want to split the memory among multiple processors using, say, symmetric multiprocessing facilities, and, one would hope, a fast inter-CPU channel (or truly shared, low-overhead memory) for Mach ports to utilize. At the level of 64 MB, however, RAM probably ought to be a shared resource since it is unlikely to be used at capacity most of the time (for most people today). For MOST people. Note also that Mach is an unusually memory conserving Unix.
Date: Sun 02-Dec-1989 17:23:01 From: Unknown Subject: Re: 64MB in cubes In article <7163@pt.cs.cmu.edu> avie@wb1.cs.cmu.edu (Avadis Tevanian) writes: >64MB of ram definitely works in a cube using 4 megabit DRAMS. We have several >4 MB SIMMs that we have used here in software engineering and most definitely >tested a 64MB configuration. I have no idea why the SIMMs from the Chip >Merchant didn't work. Perhaps there was a timing problem? The ones we have >were built in our factory and work fine (I don't remember who the actual >chip vendor was though). Very interesting. Some questions: - Do all the SIMMs in a NeXT have to be the same size? Or can we stick 4 meg SIMMs into our remaining slots and leave the current 1 megs? - What precisely are the chips that work? - What are good prices for these chips? Thanks, Roger Rosner Lighthouse Design ...!uunet!lighthouse!rock rock@lighthouse.com >From: rock@lighthouse.com (Roger Rock Rosner)

These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Marcel Waldvogel and Netfuture.ch.