Editing Digital Equipment Corporation (section)

=== Early 1990s faltering and attempted turnaround ===
As microprocessors continued to improve in the 1980s, it soon became clear that the next generation would offer performance and features equal to the best of DECs low-end minicomputer lineup. Worse, the [[Berkeley RISC]] and [[Stanford MIPS]] designs were aiming to introduce 32-bit designs that would outperform the fastest members of the VAX family, DEC's [[cash cow]].<ref>{{cite book| author = John L. Hennessy|author2=David A. Patterson |author3=David Goldberg | title = Computer Architecture: A Quantitative Approach| url = https://archive.org/details/computerarchitec0003henn| url-access = registration| year = 2003| publisher = Morgan Kaufmann| isbn = 978-1-55860-596-1| page = [https://archive.org/details/computerarchitec0003henn/page/152 152] }}</ref>

The company recognized that RISC offered at least twice as many [[million instructions per second|MIPS]] per cost as VAX,<ref name="dtj1990spring_decstation3100">{{ cite journal | url=http://www.bitsavers.org/pdf/dec/dtj/dtj_v02-02_1990.pdf | first1=Thomas C. | last1=Furlong | first2=Michael J. K. | last2=Nielsen | first3=Neil C. | last3=Wilhelm | title=Development of the DECstation 3100 | journal=Digital Technical Journal | volume=2 | issue=2 | date=Spring 1990 | publisher=Digital Equipment Corporation }}</ref> but funded so many competing projects in reaction during the 1980s that it lost control of its strategy until the early 1990s.<ref name="bellstreckerpdp11vaxalpha">{{Cite report |url=https://archive.computerhistory.org/resources/text/DEC/alpha/dec.alpha.bell_stecker.pdp-11_vax_and_alpha.1998.102630388.pdf |title=What Have We Learned from the PDP-11 - What We Have Learned from VAX and Alpha |last1=Bell |first1=Gordon |author-link1=Gordon Bell |last2=Strecker |first2=W.D. |access-date=2025-06-26}}</ref> Constrained by the huge success of their [[VAX]] and [[OpenVMS|VMS]] products, which followed the proprietary model, the company was very late to respond to these threats. In the early 1990s, DEC found its sales faltering and its first layoffs followed. The company that created the minicomputer, a dominant networking technology, and arguably the first computers for personal use, had abandoned the "low end" market, whose dominance with the PDP-8 had built the company in a previous generation. Decisions about what to do about this threat led to infighting within the company that seriously delayed their responses.

One group suggested that every possible development in the industry be poured into the construction of a new VAX family that would leapfrog the performance of the existing machines.{{r|bellstreckerpdp11vaxalpha}} This would limit the market erosion in the top-end segment, where [[profit margin]]s were maximized and DEC could continue to survive as a minicomputer vendor. This line of thought led, eventually, to the [[VAX 9000]] series, which were plagued with problems when they were first introduced in October 1989, already two years late.<ref>{{cite news|first=John|last=Markoff|title=Market Place; Digital Finally Follows a Trend|newspaper=[[The New York Times]]|date=July 16, 1990|url=https://www.nytimes.com/1990/07/16/business/market-place-digital-finally-follows-a-trend.html}}</ref> The problems took so long to work out, and the prices of the systems were so high, that DEC was never able to make the line the success they hoped.

Others within the company felt that the proper response was to introduce their own RISC designs and use those to build new machines.{{r|bellstreckerpdp11vaxalpha}} However, there was little official support for these efforts, and no less than four separate small projects ran in parallel at various labs around the US. Eventually these were gathered into the [[DEC PRISM|PRISM]] project, which delivered a credible 32-bit design with some unique features allowing it to serve as the basis of a new VAX implementation.<ref>{{cite conference|first1=Dileep|last1=Bhandarkar|first2=D.|last2=Orbits|first3=Richard|last3=Witek|first4=W.|last4=Cardoza|first5=Dave|last5=Cutler|author-link5=Dave Cutler|url=https://ieeexplore.ieee.org/document/63667|title=High performance issue oriented architecture|book-title=Proceedings of Compcon Spring '90|pages=153–160}}</ref> Infighting with teams dedicated to DEC's [[Mainframe computer|big iron]] made funding difficult, and the design was not finalized until April 1988, and then cancelled shortly thereafter.<ref>{{cite web|first=Mark|last=Smotherman|url=http://www.cs.clemson.edu/~mark/prism.html|title=PRISM (Parallel Reduced Instruction Set Machine)|website=Clemson University School of Computing|date=October 2009}}</ref> The PRISM project was accompanied by the [[DEC MICA|MICA]] project, which intended to consolidate VMS and ULTRIX into a single operating system.<ref name="mica-business-plan">{{cite web|url=http://www.bitsavers.org/pdf/dec/prism/mica/Mica_Software_Business_Plan_Mar87.pdf |archive-url=https://web.archive.org/web/20081201133102/http://bitsavers.org/pdf/dec/prism/mica/Mica_Software_Business_Plan_Mar87.pdf |archive-date=2008-12-01 |url-status=live|title=MICA Software Business Plan|author1=Catherine Richardson|author2=Terry Morris|author3=Rockie Morgan|author4=Reid Brown|author5=Donna Meikle|date=March 1987|access-date=January 4, 2021|website=bitsavers.org}}</ref>

Another group concluded that new [[workstation]]s like those from [[Sun Microsystems]] and [[Silicon Graphics]] would take away a large part of DEC's existing customer base before the new VAX systems could address the issues, and that the company needed its own Unix workstation as soon as possible. Fed up with slow progress on both the RISC and VAX fronts, a group in [[Palo Alto, California|Palo Alto]] started a [[skunkworks project]] to introduce their own systems. Selecting the MIPS processor, which was widely available, introducing the new [[DECstation]] series with the model 3100 on January 11, 1989.<ref>{{cite journal|first1=Thomas C.|last1=Furlong|first2=Michael J. K.|last2=Nielsen|first3=Neil C.|last3=Wilhelm|url=http://www.dtjcd.vmsresource.org.uk/pdfs/dtj_v02-02_1990.pdf |archive-url=https://web.archive.org/web/20110128164108/http://www.dtjcd.vmsresource.org.uk/pdfs/dtj_v02-02_1990.pdf |archive-date=2011-01-28 |url-status=live|title=Development of the DECstation 3100|journal=Digital Technical Journal|volume=2|issue=2|date=Spring 1990|pages=84–88}}</ref> These systems would see some success in the market, but were later displaced by similar models running the Alpha.

==== 32-bit MIPS and 64-bit Alpha systems (1992) ====
[[File:Locale_RS6_AlphaServer-2100-guts.jpg|thumb|Inside view of AlphaServer 2100]]

The press described DECstation as a defensive product with which DEC would capture some of the inevitable migration of VAX customers to RISC, even as it and comparable products from rivals would greatly increase the migration.<ref name="marshall19890206">{{Cite magazine |last=Marshall |first=Martin |date=1989-02-06 |title=The Year of the Workstation |url=https://books.google.com/books?id=OzoEAAAAMBAJ&pg=PT42 |access-date=2025-06-27 |magazine=InfoWorld |pages=43–44 |volume=11 |issue=6}}</ref> Eventually, in 1992, DEC launched the [[Alpha 21064|DECchip 21064]] processor, the first implementation of their [[DEC Alpha|Alpha]] [[instruction set architecture]], initially named Alpha&nbsp;AXP; the "AXP" was a "non-acronym" and was later dropped. This was a [[64-bit computing|64-bit]] [[RISC]] architecture as opposed to the 32-bit [[Complex instruction set computer|CISC]] architecture used in the VAX. It is one of the first "pure" 64-bit [[microprocessor]] architectures and implementations rather than an extension of an earlier 32-bit architecture. The Alpha offered class-leading performance at its launch and was used in the massively-parallel [[Cray T3D]]. Subsequent variants continued that performance trend into the 2000s, along with the Alpha-derived Pentium Pro, II, and III CPUs.<ref>{{Cite magazine|url=https://www.wired.com/1997/10/intel-dec-settle-alpha-chip-dispute/|title=Intel, DEC Settle Alpha Chip Dispute|magazine=[[Wired (magazine)|Wired]]|last1=Levine|first1=Daniel S.|date=October 27, 1997}}</ref><ref>{{Cite web| last1=Shandor | first1=John |url=https://www.hpcwire.com/1997/05/16/dec-sues-intel-alleges-architecture-patent-infringement/|title = Dec Sues Intel, Alleges Architecture Patent Infringement| work=HPCwire |date = May 16, 1997}}</ref> An AlphaServer SC45 supercomputer was still ranked No.&nbsp;6 in the world in November 2004.<ref>{{cite web|url=http://www.top500.org/lists/2004/11|work=TOP500 Top 10 Supercomputing Sites|title=November 2004}}</ref> Alpha-based computers comprising the DEC AXP series, later the [[AlphaStation]], and [[AlphaServer]] series respectively superseded both the VAX and MIPS architecture in DEC's product lines. They supported [[OpenVMS]], DEC [[OSF/1]] AXP (later known as [[Tru64 UNIX|Digital Unix]] or Tru64 UNIX) and Microsoft's then-new operating system, [[Windows NT]], an operating system made possible by ex-Digital Equipment Corporation engineers.<ref>{{Cite web|url=https://www.itprotoday.com/compute-engines/windows-nt-and-vms-rest-story|title=Windows NT and VMS: The Rest of the Story|first=Mark|last=Russinovich|date=November 30, 1998|website=ITPro Today}}</ref>

In 1998, following the takeover by Compaq Computer Corporation, a decision was made that Microsoft would no longer support and develop Windows NT for the Alpha series computers, a decision that was seen as the beginning of the end for the Alpha series computers.

==== StrongARM (1995) ====
{{Main|StrongARM}}

In the mid-1990s, Digital Semiconductor collaborated with [[ARM Limited]] to produce the [[StrongARM]] microprocessor. This was based in part on ARM7 and in part on DEC technologies like Alpha, and was targeted at [[embedded system]]s and portable devices. It was highly compatible with the ARMv4 architecture and was very successful, competing effectively against rivals such as the [[SuperH]] and [[MIPS architecture]]s in the [[personal digital assistant]] market. [[Microsoft]] subsequently dropped support for these other architectures in their [[Pocket PC]] platform. In 1997, as part of a lawsuit settlement, the [[StrongARM]] intellectual property was sold to [[Intel]]. They continued to produce [[StrongARM]], as well as developing it into the [[XScale]] architecture. Intel subsequently sold this business to [[Marvell Technology Group]] in 2006.