History of the ISA: Digital Equipment Corporation | Hacker News Reader

History of the ISA: Digital Equipment Corporation

allaboutcircuits.com

23 pointsrbanffy14 comments

This article really barely touches on the issue, even with regard to earlier influential architectures by DEC alone (plenty of others contributed as well, though in this regard the most influential were DEC, Intel and IBM in no particular order). It also ignores Gordon Bell, possibly the inventor of the ISA as an idea with his profound design choices in the early DEC architectures.

The most important machines to DEC were PDP-1->PDP-6->PDP-10->DEc-20", PDP8, and PDP-11->VAX, and from an ISA perspective the -10 and -11/Vax sequences had pervasive influence until today.

The 36-bit PDP6 architecture (unfairly called a "dead end" by another poster on this discussion) was highly influential: it was highly orthogonal and what we would consider today almost RISC, but was easy to program in both assembly and with a compiler. This is the architecture that gave us the Internet, Lisp machines, the bitblt instruction (ever used a bitmapped display?) and other technologies that led to today. The PDP-11 gave us C[+] -- there's a little PDP-11-like machine in every C and C++ compiler, and pretty much every other CPU-targeting compiler these days, and as a result, almost every modern ISA looks like a PDP-11 if you squint a little.

(The article also ignores what a barn burner Alpha was. Its biggest misfortune was to be born at the wrong time into a family that had fallen on hard times).

The VAX 11/780 we had at the university was a dream machine for programming.

>The article also ignores what a barn burner Alpha was.

Indeed. 300mhz, dual-issue...in ~1991! Of course, that was before low-power was "cool" and the first chips were very nice space heaters.

Source: worked on the Alpha team before it was "Alpha" ("EVAX").

Oh yeah, I'd forgotten the name EVAX!!

dont you mean 150MHz in 1992-3?

Ah, you are right. 300 “mips” (peak), 150mhz.

The cryo guys had something faster running in the back lab, but I don’t think they got 2x.

Nothing about the PDP-11 (direct ancestor of the VAX, which the article starts with). No details of how one ISA influenced another. I guess we can forgive the relative dead-end of the PDP-8 and PDP-10 architectures.

Good picture of Ken Olsen, though :-)

(I'd forgotten about PRISM, which was canceled through some nasty politics and caused Dave Cutler to move to Microsoft -- and Cutler's Windows NT was a heavy contributor to DEC's failure).

Hmm...is that one article in a series on ISAs, with this article the one on DEC. Or is it an article on ISAs, using DEC as the example?

The thing is, I can't find an article on IBM's 360, and while I like DEC's ISAs, the 360 is really the seminal example.

The article states:

>"VAX was a 32-bit Complex-Instruction-Set-Computer (CISC) ISA designed for DEC’s VMS operating system."

Is it unique that they designed the ISA for OS? Was this common back then? It seems the norm for some time now been to write an OS that targets an ISA and then possibly port it to other ISAs.

EDIT I guess its easy to forget that virtual memory was new and so you needed to design an architecture that had the concept of paging before you could have virtual memory support in your OS(VMS)

The VAX and VMS were designed for each other. Many features in the VAX were created specifically to support VMS. That said, just as the PDP-11 was the parent of the VAX, the RSX-11 operating system was the parent of VMS.

BTW, the original article is annoying unsubstantial.

Thanks.

I agree that the original article here is annoying and unsubstantial. Might you have some better references or links on the the history of VAX/VMS development and some of the lineage you mentioned such as the RSX-11? Cheers.

Search for a copy of "VAX VMS Internals and Data Structures". I have the version 4.4 book, but it appears later editions are available at Amazon and online as pdf. This is pretty dense reading and really goes into detail on the design and rationale for the design. I'd say it's a whole CS education in one volume. DEC had some pretty clever people and it shows in this book.

You might find some more historical information at https://williambader.com/museum/vax/vaxhistory.html for VMS or for RSX-11 and PDP-11 here https://williambader.com/museum/vax/pdphistory.html. Wikipedia has a good article on RSX-11 too.

Apparently you can even legally run RSX-11M on a simulator. Check http://simh.trailing-edge.com/ and http://rsx11m.com/.

I loved PDP-11s, particularly to write assembly, they were elegant. But it was always a struggle to fit a larger application into a 16 bit address space.

I'd be tempted to try the simulator, but my current hobby computing is doing embedded firmaware on the STM8 which satisfies my desire to puzzle solve fitting a lot of function into too small resources.

Thank you these are great links, I appreciate it. Cheers.

This article is fluff. If you're looking for substantive information on the earlier half of DEC's history, I recommend this book.

Computer Engineering: a DEC View of Hardware Systems Design C. Gordon Bell, J. Craig Mudge, John E. McNamara

http://www.computerhistory.org/collections/catalog/102630383

Reply on news.ycombinator.com