Alasir Enterprises
Main Page >  Articles >  Alpha: The History in Facts and Comments  

 
Main Page
 
 
Reviews
 
Articles
 
Software
 
Reference
 
Motley
 
 
About Us
 
 
Alpha: The History in Facts and Comments
Alpha Powered Dig my grave both long and narrow
Make my coffin neat and strong 

(from an old American song)

Paul V. Bolotoff
 
Release date: 14th of April 2005
Last modify date: 8th of April 2010

Contents:
in Russian

 
Foreword

This work is dedicated to Alpha processors, system logic sets and the architecture in whole. It's more than just an overview, and it doesn't pretend to be just a technical reference. Frankly speaking, the complete history of the Alpha architecture from the very beginning to our days is taken into consideration. Nevertheless, there is a real lot of various reference information included because this work would be vastly incomplete and difficult to understand otherwise. There is no serious way to talk about history without having facts in hand and analysing them. Maybe this article would be looking better if written and published several years ago, when Alpha systems have been real kings in means of performance and their future has been expected to be bright very much. Who knows. However, it seems to be the right time nowadays to draw the final line, to explain what has happened, and why one of the most interesting and promising computer architectures has been thrown into oblivion. Anyway, this article isn't a necrologue or a funeral prayer definitely. Many facts have been interpreted accordingly to the author's knowledge and experience. No one is perfect, so take it or leave it.

DEC's logotype

PDP and VAX

Digital Equipment Corporation (abbreviated to DEC) was founded in 1957 by two engineers, Kenneth Olsen and Harlan Anderson, graduates of the Massachusetts Institute of Technology, and was one of the oldest and most known companies of the world computer industry.
 
Before founding, Olsen worked for Lincoln Laboratory at the institute mentioned above and participated in development of one of the world's first transistor based computers called TX-2. By the way, that laboratory was sponsored those days by the Department of Defense of the United States. The company started business by producing and selling backplane modules for computers, but in 1960 it offered the first computer of its own — an 18-bit PDP-1 (Programmable Data Processor - 1) capable of about 100 thousand operations per second. That machine was used eventually to run the first computer game in known history, Spacewar by Steven Russell. A 12-bit PDP-8 was introduced in 1964. Sized like a small wardrobe or refrigerator, it deserved an honour to be called the first "minicomputer" manufactured in quantity. In addition, the list price was quite attractive, about 18000 USD (1965). In fact, PDP-8 offered an excellent price/performance ratio which allowed to compete successfully with those famous mainframe systems by IBM. There were about 1450 machines produced until 1968 not counting numerous modifications which followed the original design, when a 36-bit PDP-10 was introduced. It was based upon the design of experimental PDP-6 and was targeted at data processing centres, research laboratories and military needs. Certain development attempts were taken towards this 36-bit architecture which were organised within the Unicorn project under supervision of Leonard Hughes and David Rogers. However, the project was closed in June of 1975 with all its resources transferred to support a new 32-bit architecture. Nevertheless, different versions of PDP-10 remained in production until 1983.
 
A 16-bit PDP-11 appeared somewhere in the beginning of 1970's. It descended from the PDP-8 model line, though there were numerous improvements made. For instance, it was the first computer by DEC to make use of 8-bit bytes. A simple yet effective Unibus based architecture which was upgraded upon Q-bus later, a considerably effective instruction set and low manufacturing costs turned the PDP-11 model line into a real success on the market. No wonder that PDP-11 became a subject of cloning attempts all over the world including even those "countries of people's democracy": SM-4 (the USSR, Bulgaria, Hungary), SM-1420 (the USSR, Bulgaria, the German Democratic Republic), SM-1600 (the USSR), IZOT-1016 (Bulgaria), DVK (the USSR). There were many operating systems developed for PDP-11: P/OS, RSX-11, RT-11, RSTS/E and several derivatives of DOS. In addition, the first release of UNIX operating system was completed at Bell Laboratories in 1971 on PDP-7 and PDP-11 machines using their assembly languages. PDP-11 left the market during 1980's because of one but inevitable reason: lack of address space. A new architecture, 32-bit though still CISC, was promoted to the market.
 
So, that architecture was called VAX (Virtual Address eXtension) officially after a VAX Architecture Committee session in April of 1975. The new architecture was a logical conclusion to several months of development while the Star project was operational under supervision by Gordon Bell. This project took a run in parallel with the Unicorn project mentioned above, but was considered a more promising. In brief, the Star project was to evaluate performance improvements which could be achieved by increasing PDP-11 integer registers' width from 16 to 32 bits and their number from 8 to 16, also with a significant redesign of the instruction set. The first VAX machine, model 11/780, was announced in October of 1977. A few months later, in February of 1978, a new operating system for VAXen (the plural for VAX) was released — VMS (Virtual Memory System) v1.0. It was a multiuser and multitasking OS supporting up to 64Mb of operating memory, networking functions (DECnet), adaptive task scheduler, extended process management and many more innovations hard to be seen before. Renamed to VAX/VMS and carrying numerous improvements over the previous release, v2.0 was released in April of 1980. In addition, the classical UNIX OS was also ported to VAX. Overall, VAXen were manufactured and sold with a success during 1980's and were shipped in limited quantities under special contracts even when close to the end of the century. The whole model line included several dozens of machines ranging from compact workstations (MicroVAX) to 6-processor mainframe-class servers. There were several generations of VAX processors manufactured throughout 1980's using DEC's proprietary CMOS technological processes: CVAX in 1985 (12MHz maximum, 0.2 mln. transistors, 2.0µ CMOS1), Rigel in 1987 (37MHz maximum, 0.4 mln. transistors, 1.5µ CMOS2), Mariah in 1989 (63MHz maximum, 0.8 mln. transistors, 1.0µ CMOS3). The last two processors of this architecture called NVAX and NVAX+ were presented in 1992. Both of them carried 1.3 mln. transistors internally and were manufactured using a proprietary 3-layer 0.75µ CMOS4 process. The maximal clock speed of NVAX was defined at 83.3MHz while of NVAX+ — at 90.9MHz. Although they shared the same functional design, NVAX supported a legacy system bus interface while NVAX+ was made bus compatible with Alpha 21064 (EV4). Even nowadays, thousands of VAXen keep working at subdivisions of the Department of Defense and the NSA (National Security Agency), also at numerous commercial organisations.
 
VAX 11/780 brochure cover VMS sales update cover
  Next page >>

Copyright (c) Paul V. Bolotoff, 2005-10. All rights reserved.
A full or partial reprint without a permission received from the author is prohibited.
 
Designed and maintained by Alasir Enterprises, 1999-2010
rhett from alasir.com, walter from alasir.com