Editing PDP-11 (section)

== History ==
===Previous machines===
In 1963, DEC introduced what is considered to be the first commercial minicomputer in the form of the [[PDP-5|PDP–5]]. This was a 12-bit design adapted from the 1962 [[LINC]] machine that was intended to be used in a lab setting. DEC slightly simplified the LINC system and instruction set, aiming the PDP-5 at smaller settings that did not need the power of their larger 18-bit [[PDP-4]]. The PDP-5 was a success, ultimately selling about 1,000 machines. This led to the [[PDP-8|PDP–8]], a further cost-reduced 12-bit model that sold about 50,000 units.

During this period, the computer market was moving from [[computer word]] lengths based on units of 6 bits to units of 8 bits, following the introduction of the 7-bit [[ASCII]] standard. In 1967–1968, DEC engineers designed a 16-bit machine, the PDP–X,<ref>{{cite web|url=http://www.bitsavers.org/pdf/dec/pdp-x/|title=PDP-X memoranda|website=bitsavers.org|access-date=2017-07-13|archive-date=2017-09-23|archive-url=https://web.archive.org/web/20170923042101/http://bitsavers.org/pdf/dec/pdp-x/|url-status=live}}</ref> but management ultimately canceled the project as it did not appear to offer a significant advantage over their existing 12- and 18-bit platforms.

This prompted several of the engineers from the PDP-X program to leave DEC and form [[Data General]]. The next year they introduced the 16-bit [[Data General Nova]].<ref>{{cite web|url=https://archive.computerhistory.org/resources/access/text/2012/07/102702207-05-01-acc.pdf|title=Oral History of Edson (Ed) D. de Castro|access-date=April 28, 2020|archive-date=2016-03-05|archive-url=https://web.archive.org/web/20160305040632/http://archive.computerhistory.org/resources/access/text/2012/07/102702207-05-01-acc.pdf|url-status=live}}</ref> The Nova sold tens of thousands of units and launched what would become one of DEC's major competitors through the 1970s and 1980s.

===Release===
[[Ken Olsen]], president and founder of DEC, was more interested in a small 8-bit machine than the larger 16-bit system. This became the "Desk Calculator" project. Not long after, [[Datamation]] published a note about a [[desk calculator]] being developed at DEC, which caused concern at [[Wang Laboratories]], who were heavily invested in that market. Before long, it became clear that the entire market was moving to 16-bit, and the Desk Calculator began a 16-bit design as well.<ref name=McGowan98>{{cite web |last=McGowan |first=Larry |url=http://hampage.hu/pdp-11/birth.html |title=How the PDP-11 Was Born |date=19 August 1998 |access-date=2015-01-22 |archive-date=2015-06-17 |archive-url=https://web.archive.org/web/20150617022503/http://www.hampage.hu/pdp-11/birth.html |url-status=live }}</ref>

The team decided that the best approach to a new architecture would be to minimize the memory bandwidth needed to execute the instructions. Larry McGowan coded a series of [[assembly language]] programs using the instruction sets of various existing platforms and examined how much memory would be exchanged to execute them. Harold McFarland joined the effort and had already written a very complex instruction set that the team rejected, but a second one was simpler and would ultimately form the basis for the PDP–11.<ref name=McGowan98/>

When they first presented the new architecture, the managers were dismayed. It lacked single instruction-word immediate data and short addresses, both of which were considered essential to improving memory performance. McGowan and McFarland were eventually able to convince them that the system would work as expected, and suddenly "the Desk Calculator project got hot".<ref name=McGowan98/> Much of the system was developed using a [[PDP-10]] where the SIM-11 simulated what would become the PDP–11/20 and Bob Bowers wrote an assembler for it.<ref name=McGowan98/>

At a late stage, the marketing team wanted to ship the system with 2K of memory{{efn|It is not clear in the document whether this is 2k bytes or 2k words – 4k in modern terms.}} as the minimal configuration. When McGowan stated this would mean an assembler could not run on the system, the minimum was expanded to 4K. The marketing team also wanted to use the forward slash character for comments in the assembler code, as was the case in the PDP–8 assembler. McGowan stated that he would then have to use semicolon to indicate division, and the idea was dropped.<ref name=McGowan98/>

The PDP–11 family was announced in January 1970 and shipments began early that year. DEC sold over 170,000 PDP–11s in the 1970s.<ref name=Ceruzzi03>Paul Cerruzi, ''A History of Modern Computing'', MIT Press, 2003, {{ISBN|0-262-53203-4}}, page 199</ref>

Initially manufactured of small-scale [[transistor–transistor logic]], a single-board [[Integrated circuit#LSI|large-scale integration]] version of the processor was developed in 1975. A two- or three-chip processor, the [[DEC J-11|J-11]] was developed in 1979.

The last models of the PDP–11 line were the single board PDP–11/94 and PDP–11/93 introduced in 1990.<ref name="16-bit Timeline">{{cite web |url=http://research.microsoft.com/~gbell/digital/timeline/16-bit.htm |title=16-bit Timeline |work=microsoft.com |access-date=November 8, 2016 |archive-date=December 8, 2008 |archive-url=https://web.archive.org/web/20081208063850/http://research.microsoft.com/~gbell/digital/timeline/16-bit.htm |url-status=live }}</ref>