Talk:VAX-11

Latest comment: 2 years ago by Guy Harris in topic Genesis of 11/782 ?

VAX-11/788

edit

I'm guessing the 11/788 was related to the other 78x models, although there are very few references to it on the web. Letdorf (talk) 12:51, 27 October 2008 (UTC).Reply

Google Books can only locate four books that mention the VAX-11/788, and only one of them provides a preview which states:
Page 632: "If we allocate one unit to a microvax S. processor CPU time, then we find that a VAX 11/788 (a 788 is similar to that of a 780) is 1.2 times of a microvax ..." (Emphasis is mine).
The book is "International Symposium Digest, Antennas and Propagation," by Institute of Electrical and Electronics Engineers, International Union of Radio Science, IEEE Antennas and Propagation Society (1976).
So we know its similar, but how different is it? Rilak (talk) 05:35, 28 October 2008 (UTC)Reply

Did the 11/788 even exist? Despite a lot of familiarity with VAX, I've never heard of it. Maybe a proposed configuration that was never built? — Preceding unsigned comment added by 173.76.121.238 (talk) 15:35, 2 February 2020 (UTC)Reply

I see a few mentions in other documents, generally referring to printer support of all things. However, it may be that even those references are bad OCRs of "780", which will turn up many hits on "788". I cannot find any references at all to VISQ except wiki related pages. 20:53, 27 October 2020 (UTC)Maury Markowitz (talk)

Serial I/O

edit

Would it be worth mentioning many serial ports were supported? Kevink707 (talk) 20:51, 10 April 2011 (UTC)Reply

The serial ports were typically implemented via DH-11 or DZ-11 cards on Unibus(es), so at least the hardware was virtually unlimited. BMJ-pdx (talk) 18:36, 26 April 2022 (UTC)Reply

Role in BSD Unix and UUCP

edit

I think some mention of VAX-11 being the host hardware for BSD 4.x is warranted. Also, many of the major UUCP mail relay points were Vax-11 hardware (i.e. ucbvax, mcvax). Kevink707 (talk) 20:51, 10 April 2011 (UTC)Reply

3BSD and 4BSD were originally done on a VAX-11/780, but does that belong here or on VAX?
And, by the time DEC dropped the "-11", were most of those hosts no longer VAXes? Guy Harris (talk) 07:53, 29 May 2021 (UTC)Reply

Bit posions

edit

In the infobox (added on 2013-08-10), the bit positions typography — 31, …, 01, 00 — don't make much sense. According to the manual, these are 32 bits, denoted decimally 31, …, 01, 00. —Mykhal (talk) 07:02, 29 May 2021 (UTC)Reply

"these are 32 bits" Yes, they are. In what fashion does what's in the infobox not reflect that? It doesn't include all values from 31 to 00, but it presumably doesn't do so in order to limit the width of the infobox.
"the bit positions typography — 31 ... 00 — don't make much sense" In what fashion? They're showing the denotations 31, etc., they just happen to do it with small numbers on a diagonal.
But that infobox doesn't belong in this article in any case. It's already in VAX, which is the article about the instruction set architecture; this page isn't about the instruction set, it's about the machines implementing that instruction set that had names beginning with "VAX-11/" followed by a number. It should probably be removed, and further discussion of the infobox should be done on Talk:VAX, as a similar infobox appears in VAX#Processor architecture. Guy Harris (talk) 07:48, 29 May 2021 (UTC)Reply
The horizontal space-saving argument is rather funny... why not rather remove more numbers? There might be some other reason, why to have the digits on diagonals, which I don't yet understand; however, because of last eight bit group, it resembles bytebit notation with intended typography: …, 07, (…), 00, which would be correct but only up to that 07. —Mykhal (talk) 08:24, 29 May 2021 (UTC)Reply
"why not rather remove more numbers?" E.g., only showing 31, 23, 15, and 00? And in what fashion is 07 correct but 00 not correct, if that's what you mean by "correct but only up to that 07"? Guy Harris (talk) 08:47, 29 May 2021 (UTC)Reply
I mean up to, in numerical sense, from right. I.e. 07 correct, but (hidden) 08 wrong., as well as all higher indexed numbers from left. —Mykhal (talk) 08:52, 29 May 2021 (UTC)Reply
So why is 07 correct, but 08 to the left of it, if visible, would be wrong? Guy Harris (talk) 10:59, 29 May 2021 (UTC)Reply
Only if meant as byte number, and its bit index in superscript, and it's obvious. —Mykhal (talk) 06:45, 30 May 2021 (UTC)Reply

For now, fixed at least on cswiki, where the infobox was already wider. —Mykhal (talk) 06:58, 30 May 2021 (UTC)Reply

Why is it not obvious that, for example, 31 is the number "31" written diagonally? If there's a way to render it vertically, that would be better, as it would match the way bit numbers are shown in the 1982-03-20 VAX-11 Architecture Reference Manual on pages 2-2 through 2-6. The original 1977 VAX-11/780 Architecture Handbook writes them horizontally; in both of those manuals, the bit numbers do not have leading zeroes. Guy Harris (talk) 07:28, 30 May 2021 (UTC)Reply
As one of the editors who put a lot of time into making the CPU architecture boxes (several years back now, ca. May 2013), I intended the boxes to look pretty much the same across all of the CPU articles. The bit positions were an attempt to keep an already cramped horizontal space as small as possible while still showing useful register bit widths. See the articles for the 8080, 6502, Z-80, 8086, 68000, IBM 360, PDP-10, PDP-11, and several others. Note that some of the manufacturers number the bits from left to right, instead of the more usual right to left order. I'd really like to see every CPU architecture article feature such a box, but I'll leave that noble task for others to complete. I can't take credit for the original design of the CPU box (I don't even remember which article first featured it), but I did add improvements to it. If you think changing the layout of the bit numbers will improve things, that's fine, but please be sure you make the changes across all of the other existing CPU articles to maintain a consistent look within WP. — Loadmaster (talk) 23:32, 10 June 2021 (UTC)Reply

Genesis of 11/782 ?

edit

When the 11/782 was first created, I remember hearing/reading that it was a project of UCB. They basically replaced the terminator of the MA780 bus with a second CPU (the first CPU was at the other end of the bus, providing termination at that end). One CPU was dedicated to system mode, the other to user mode (that differs somewhat from the description of the production version in the article).

Can any UCB old-timers confirm this, particularly, the roles of each CPU?

BMJ-pdx (talk) 18:24, 26 April 2022 (UTC)Reply

There was a dual-11/780 system - developed at Purdue University, not the University of California, Berkeley - whose hardware description sounds like that[1]: 292–293 , but whose software description is similar to that of the 11/782, with both CPUs running user-mode code and one designated CPU running kernel-mode code.[1]: 293–294 
Is that what you're thinking of? Guy Harris (talk) 20:08, 26 April 2022 (UTC)Reply

References

  1. ^ a b Goble, George H.; Marsh, Michael H. (April 1982). A dual processor VAX 11/780. 9th Annual Symposium on Computer Architecture. IEEE Computer Society Press. pp. 291–298. doi:10.1145/1067649.801738.