Aucbvax.5536 fa.works utzoo!decvax!ucbvax!works Thu Dec 24 00:30:52 1981 WorkS Digest V1 #47 >From JSol@RUTGERS Thu Dec 24 00:21:16 1981 Works Digest Thursday, 24 Dec 1981 Volume 1 : Issue 47 Today's Topics: Administrivia Paging On Other Hardware National Semiconductor NS16032s - Availability Commentary - Contents Of This Digest ---------------------------------------------------------------------- Date: 20 Dec 1981 1924-EST From: Jonathan Alan Solomon Subject: Administrivia An old bug involving truncated messages just cropped up while sending Monday's digest. There was no digest yesterday. Let me know if you did not receive a complete digest on Monday, and I will resend one to you. This is the final digest of the year. I want to wish everyone a very Merry Christmas, and a Happy New Year. Next digest is scheduled for the 4th of January. Cheers, JSol ------------------------------ Date: 21 December 1981 23:17 est From: SSteinberg.SoftArts at MIT-Multics Subject: stacks on the 68000 Sender: COMSAT.SoftArts at MIT-Multics I thought that 68000 stack references went to memory. If so, the two CPU paging kludge would work with no modifications. Is the stack register a really weird kludge? [FLAME ON] The VAX paging system appears to offer no improvements in basic design since the work done on the ATLAS in the early 60's. Granted the hardware support is much better but I get the impression that the VAX paging scheme is needlessly complex! I also noticed that people cannot figure out the second step of the reasoning chain which begins: The reason for paging is to provide a large memory space in a small physical memory. The next step is: The reason to have a large memory space is to provide a large [object] name space. Large name/address spaces eliminate the need for vasty crocks such as overlay managers, data object paging managers and explicit user disk I/O. A great deal was written about this in the early sixties but as usual the industry is still 20-30 years behind the software technology. ------------------------------ Date: 22 Dec 1981 02:27:29-PST From: decvax!microsoft!gordon at Berkeley There are currently 'several' National Semiconductor NS16032s which are primarily functional. A completly functional chip is expected soon, the MMU chip will be available around mid-year. It is interesting to note that the 16032's MMU is a dual-level virtual memory system, with an associative LRU memory to hold the mapping elements. The MMU also has lots of fancy debugging go-fast in it, such as hardware breakpoints, and instruction back-tracing. The architecture is very good, and it looks like the initial 10mhz part will run 'c' a little bit faster than the 8mhz 68000. The hardware people claim that the internal layout of the chip will allow a 2 to 1 improvement in speed, without requiring faster clocks and memorys, when they do their next design itteration. ------------------------------ Date: 22 December 1981 22:06-EST From: Brian P. Lloyd Subject: Content of this digest C'mon gang. I thought that this digest was created to discuss workstations and not chips. I accept the fact that the 68000 has demand paging problems, but I also feel that the hardware hacking sould be discussed in Info-Micro. I personally would rather talk about making useful workstations for PEOPLE and what approaches other manufacturers have taken. I guess I am interested in ideas rather than engineering. Brian [I agree that the discussion of paging hardware for a specific microcomputer should be best discussed on INFO-MICRO, but as the discussion has started (with today's digest) to branch out to a more general paging discussion, there is really no better forum to discuss it (hint: The VAX is NOT a microcomputer). I don't advocate creating a INFO-PAGING-HARDWARE digest, but messages about any other topic will take precedence. Anyone wanting to discuss this, or any other policy or mechanical issue of the WorkS digest, please send mail to WORKS-REQUEST@MIT-AI, not to the whole list. -JSOL] ------------------------------ End of WorkS Digest ******************* ------- ----------------------------------------------------------------- gopher://quux.org/ conversion by John Goerzen of http://communication.ucsd.edu/A-News/ This Usenet Oldnews Archive article may be copied and distributed freely, provided: 1. There is no money collected for the text(s) of the articles. 2. The following notice remains appended to each copy: The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996 Bruce Jones, Henry Spencer, David Wiseman.