Meteorologist, oceanographer, Python solvem probler, experimenter, Raspberry Pi tinkerer, geek dad, manager, space enthusiast, ham radio lover, science columnist, ADHDer, synthetic intellect curious, linux admin cosplayer.

Verified Facts: Post 31 - Tomasulo / IBM 360/91 / Conway / NMC

Post: 2026-05-06-The-algorithm-that-outlived-its-machine.md Status: In progress

Key Claims To Verify

Tomasulo & 360/91

  • Robert Tomasulo born 25 January 1934, died March 2008
  • “An Efficient Algorithm for Exploiting Multiple Arithmetic Units” - IBM Journal of R&D 11(1), Jan 1967, pp. 25-33
  • IBM System/360 Model 91 announcement / shipments
  • Tomasulo algorithm: register renaming, reservation stations, common data bus

Lynn Conway

  • Born 2 January 1938
  • Died 9 June 2024
  • 23 February 1966 DIS (Dynamic Instruction Scheduling) memo at IBM
  • Four-author cover sheet: Conway/Randell/Rozenberg/Senzig
  • Fired by IBM in 1968 for announcing gender transition
  • IBM’s 14 October 2020 apology by Diane Gherson

John Cocke

  • Born 30 May 1925
  • Died 16 July 2002
  • Son of Duke Power president (Norman A. Cocke)
  • Father of RISC

P6 / Pentium Pro

  • Launched 1 November 1995
  • $974 for 150 MHz model

Watson Jr.

  • “thirty-four people including the janitor” memo
  • 28 August 1963

CDC v IBM antitrust

  • Filed December 1968 in D. Minn.
  • Judge Edward Devitt or Neville? (verify)
  • $80M settlement (not $600M)

U.S. v IBM

  • Filed 17 January 1969
  • By AG Ramsey Clark (last day in office)
  • Dismissed 8 January 1982 by William Baxter

Verified Facts

Robert Tomasulo

  • Born 31 October 1934 (Wikipedia says 31 Oct 1934, NOT 25 January 1934 as the brief stated)
  • Died 3 April 2008 (NOT “March 2008” as the brief stated)
  • Manhattan College -> Syracuse University engineering degree (BS, then engineering degree, not confirmed if MS)
  • Joined IBM 1956; 25-year IBM career
  • 1997 Eckert-Mauchly Award - citation “For the ingenious Tomasulo algorithm, which enabled out-of-order execution processors to be implemented” CONFIRMED
  • Post says he attended Regis High School - confirmed
  • IBM Journal of R&D 11(1):25-33, January 1967, “An Efficient Algorithm for Exploiting Multiple Arithmetic Units” CONFIRMED

Lynn Conway

  • Born 2 January 1938, Mount Vernon NY CONFIRMED
  • Died 9 June 2024, Jackson MI CONFIRMED
  • Joined IBM Research Yorktown 1964 CONFIRMED
  • Fired by IBM 1968 CONFIRMED
  • IBM apology 14 October 2020 by Diane Gherson at virtual event “Tech Trailblazer and Transgender Pioneer” CONFIRMED (1200 IBM employees)
  • 23 February 1966 DIS memo: Conway/Randell/Rozenberg/Senzig four authors - confirmed by post’s URL (PDF at ai.eecs.umich.edu)

John Cocke

  • Born 30 May 1925 in Charlotte NC CONFIRMED
  • Died 16 July 2002 in Valhalla NY at age 77 CONFIRMED
  • Duke BS Mech Eng 1946, PhD Math 1956 CONFIRMED
  • IBM career 1956-1992 CONFIRMED
  • Turing Award 1987, Eckert-Mauchly 1985, National Medal of Tech 1991, National Medal of Science 1994 CONFIRMED
  • Father Norman Cocke president of Duke Power - NOT directly confirmed in Wikipedia; needs separate check (Norman Cocke article gives 404)

Pentium Pro

  • Launched 1 November 1995 CONFIRMED
  • $974 starting price (150 MHz model) CONFIRMED via Wikipedia secondary, Day in Tech History
  • 5.5 million transistors CONFIRMED
  • PROCESS DISCREPANCY: Wikipedia says 150 MHz used 0.50 micron BiCMOS (NOT 0.6 micron). Post says: “The launch part ran at 150 megahertz on a 0.6 micron BiCMOS process”. The 0.6 micron process was the 133 MHz prototype/initial version. The 150 MHz and 166 MHz versions were 0.35/0.50 micron BiCMOS. This may need correction or qualification.

CDC v IBM

  • December 1968 filing, D. Minn., Judge Philip Neville CONFIRMED
  • $80 million settlement value CONFIRMED
  • Settlement January 1973 CONFIRMED

IBM System/360 Model 91

  • 15 produced, 4 internal, post says 14 with PJ&P canonical – both numbers in tension; Wikipedia says 15 explicitly. Post hedges this correctly (“fourteen to fifteen built, ten outside customers”).
  • First shipped October 1967 to NASA Goddard CONFIRMED
  • First implementation of Tomasulo’s algorithm CONFIRMED
  • Two Model 95 thin-film variants to NASA CONFIRMED

Watson “thirty-four people including the janitor” memo

  • Dated 28 August 1963 CONFIRMED (Wikipedia CDC 6600, CHM Revolution exhibit)
  • Cray’s reply “It seems like Mr. Watson has answered his own question” CONFIRMED

Cocke’s father

  • Norman Atwater Cocke (1884-1974) CONFIRMED
  • President of Duke Power Company CONFIRMED
  • Lake Norman named for him CONFIRMED
  • Source: amturing.acm.org, IBM history page, multiple genealogy sources

Pentium Pro process (NUANCED)

  • ISSCC paper Feb 1995 by Colwell & Steck: “A 150 MHz 0.6 µm BiCMOS Superscalar Microprocessor” (the prototype/research die) - this is what Colwell published
  • Actual launched 150 MHz part on 1 November 1995: 0.50 micron BiCMOS per multiple primary sources (Wikipedia Pentium Pro, MPR Nov 1995)
  • 166/180/200 MHz parts: 0.35 micron BiCMOS
  • The post’s claim “The launch part ran at 150 megahertz on a 0.6 micron BiCMOS process” is INACCURATE for the launched product; it conflates the ISSCC research-paper process with the production-launch process.
  • Possible fix: change to “0.50 micron BiCMOS” or hedge as “early production parts moved from 0.6 to 0.5 micron BiCMOS by launch”

U.S. v IBM filing details

  • Filed 17 January 1969 by AG Ramsey Clark (last day before Nixon admin) - WIDELY ATTESTED in secondary sources, Wikipedia 404 on direct article
  • Judge David Edelstein SDNY CONFIRMED via secondary sources
  • 8 January 1982 dismissal by Baxter CONFIRMED via secondary sources
  • 104,000-page transcript figure cited in Levy & Welzer 1985 - reasonable

Items NOT directly verified but consistent with sources

  • Tomasulo’s first name “Robert Marco” - Wikipedia just says “Robert Tomasulo”; the middle name “Marco” is in the post but not in the Wikipedia article. Should be flagged as not directly confirmed.
  • “W. D. Silkman” implementer named in Tomasulo 1967 acknowledgements - taken from the paper text (Tomasulo’s own acknowledgements). Verified by post via quoted footnote.
  • Conway’s date of arrival at Yorktown — post says 1965 per her 2012 memoir. Wikipedia says joined 1964. Both can be correct (hired 1964, arrived 1965).

Sources

  • https://en.wikipedia.org/wiki/Robert_Tomasulo
  • https://en.wikipedia.org/wiki/Lynn_Conway
  • https://en.wikipedia.org/wiki/John_Cocke_(computer_scientist)
  • https://en.wikipedia.org/wiki/Pentium_Pro
  • https://en.wikipedia.org/wiki/Tomasulo%27s_algorithm
  • https://en.wikipedia.org/wiki/Control_Data_Corporation
  • https://en.wikipedia.org/wiki/IBM_System/360_Model_91
  • https://wraltechwire.com/2020/11/22/ibm-apologizes-52-years-later-for-firing-transgender-pioneering-engineer/
  • Justia: Control Data Corp. v. IBM, 306 F. Supp. 839 (D. Minn. 1969) https://law.justia.com/cases/federal/district-courts/FSupp/306/839/1794359/
  • https://dayintechhistory.com/dith/november-1-1995-pentium-pro-ibm-pcjr-introduced/