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.

The People of ILLIAC IV

Research file for the forthcoming blog post on the ILLIAC IV at Urbana, Pasadena, and Mountain View. Cross-references the CDC 7600 post (Post 32, 2026-05-07) and the IBM 360/91 post (Post 31, 2026-05-06). The ILLIAC IV is the third great supercomputer of the 1968-1972 window, and unlike the other two it was not built by a commercial vendor as a product line. It was a research procurement, prototyped at one university, manufactured by a defence-electronics-and-banking conglomerate in Pasadena California, contested by the campus antiwar movement of 1969-1971, moved across the continent to a US Navy base on the south shore of San Francisco Bay, and used for ten years as the principal computational fluid dynamics machine of the United States space programme. The personnel are correspondingly varied. They include the von Neumann Institute alumnus who turned a 1952 idea into a 1972 machine; the principal Burroughs engineer who carried the construction project through Pasadena; the Burroughs B5000-derived architectural inheritance that put a B6500/B6700 in the front end; the Berkeley operating-systems engineer who became director at NASA; the German-born NASA Ames director who defended the project through political controversy; and the NASA Ames computational fluid dynamicist who became its principal scientific user.

Daniel L. Slotnick (12 November 1931 – 25 October 1985)

Daniel Leonid Slotnick was born on 12 November 1931 in New York City.1 The “Bronx” attribution that surfaces in some secondary sources does not appear to be confirmed in the primary biographical record on the IEEE Computer Society profile or in Prabook, both of which simply say “New York City.” He took a Bachelor of Arts in mathematics from Columbia University in 1951, a Master of Science in mathematics from Columbia in 1952, and a PhD in applied mathematics from the Courant Institute of Mathematical Sciences at New York University in 1956.2 The Courant connection is significant: it is the institution that, two decades later, would produce the Bauer-Garabedian-Korn airfoil codes that ran on his machine.

Immediately upon completing his MS in 1952, Slotnick joined the Electronic Computer Project at the Institute for Advanced Study at Princeton, under John von Neumann.3 This is the cohort to which the basic idea for parallel array processing traces back. Slotnick’s own account, in his 1982 retrospective “The Conception and Development of Parallel Processors: A Personal Memoir,” published in Annals of the History of Computing 4(1), pages 20-30, places the moment of conception in 1952 while he was a programmer at the IAS. Von Neumann himself, by Slotnick’s account, was sceptical of the idea – the IAS-machine architecture was a single-CPU design and von Neumann was committed to it.

Slotnick moved to IBM’s Development Laboratory in Poughkeepsie, New York, in 1957, where he stayed three years.4 During this period he assisted Rex Rice in the design of a small general-purpose computer, and crucially he had the conversations with John Cocke that produced the Cocke-Slotnick paper of 1958, the first published proposal for parallel processing in numerical computation. The Manhattan Project consultancy mentioned in some bibliographies appears to be a misattribution; the Manhattan Project formally ended in 1947, when Slotnick was sixteen, and there is no documented consultancy relationship.

In 1960 Slotnick was hired by Westinghouse Electric Corporation in Baltimore Maryland, in the air-defence and weapons-systems division.5 At Westinghouse he was given freedom to pursue parallel-computer architectures. The result was the SOLOMON project (“Simultaneous Operation Linked Ordinal Modular Network”), a 1024-element single-instruction multiple-data array processor that Slotnick designed between 1960 and 1962. The 1962 paper that announced the design publicly is the famous one: Slotnick, D. L., Borck, W. C., and McReynolds, R. C., “The SOLOMON Computer,” AFIPS Conference Proceedings 22, Fall Joint Computer Conference 1962, pp. 97-107. The paper won the AFIPS Prize for Slotnick in 1962. He filed the basic patent on centrally controlled parallel processors in 1962 and received it in 1966. Westinghouse built a partial prototype called SOLOMON I, and a more ambitious follow-on called SOLOMON II, but neither went into production. Slotnick stayed at Westinghouse through 1965, when ARPA funding for the Westinghouse parallel-processing programme dried up.

In May 1965 Slotnick was recruited to the University of Illinois at Urbana-Champaign as Professor of Computer Science by John R. Pasta, then chairman of the department.6 Pasta – the New York-born Los Alamos physicist of Fermi-Pasta-Ulam fame, whose own biography we have covered elsewhere – knew Slotnick from the early-1960s AEC committee on emerging computer technologies. Pasta’s recruitment of Slotnick was deliberate: the Department of Computer Science at Illinois had inherited the ILLIAC line (ILLIAC I, 1952; ILLIAC II, 1962; ILLIAC III, never completed), and Pasta wanted a parallel-machine project to follow them. Within months of Slotnick’s arrival, Ivan Sutherland of ARPA’s Information Processing Techniques Office visited Urbana to discuss the next ILLIAC. According to Pasta’s biographical retrospective, “A positive decision was made, and the ILLIAC IV project was born. Within a few months, a $10 million contract with ARPA was executed.”7 The contract was signed in 1965-1966 with Slotnick as principal investigator and Pasta as the departmental sponsor.

The construction subcontract with Burroughs (with Texas Instruments providing the integrated circuits) was signed in August 1966.8 Slotnick was responsible for the architectural specification; Burroughs was responsible for the physical construction in Pasadena, California; the University of Illinois was responsible for the operating system, the FORTRAN-derived parallel languages (IVTRAN, TRANQUIL, GLYPNIR), and the eventual operational support. The original specification called for 256 processing elements organised as four quadrants of 64 PEs each, all sharing a single instruction stream from a Control Unit. By 1969, with ARPA funding pressure, the design was scaled back to a single 64-PE quadrant. This is the version that was eventually built. Slotnick spent 1966-1971 dividing his time between Urbana and Pasadena, supervising the contract.

The campus crisis of 1969-1970 interrupted the project. (See “The 1970 University of Illinois crisis” below for the full story.) On 9 January 1971, with help from Hans Mark at NASA Ames, ARPA decided to remove ILLIAC IV from Urbana and install it at Ames Research Center on Moffett Field, a US Navy base on the south shore of San Francisco Bay. Slotnick remained at Urbana as Professor of Computer Science but was no longer responsible for daily operations of the machine. Burroughs delivered the machine to Ames in April 1972, and the first programs ran in summer 1973. Slotnick returned to Urbana to teach, supervise students, and deal with the ongoing software work that the Center for Advanced Computation conducted as the principal site for ILLIAC IV applications development through the 1970s.

Slotnick was elected an IEEE Fellow in 1976 “for contributions to the development of parallel array processors.”9 In 1983 he received the W. Wallace McDowell Award of the IEEE Computer Society “For his pioneering contributions to centrally controlled parallel computers and for his achievement in creating the parallel computer ILLIAC IV.” The Eckert-Mauchly Award and the IEEE Computer Pioneer Award that the user query mentioned do not appear in the primary record; the McDowell Award of 1983 is the principal recognition, and the Eckert-Mauchly Award of 1985 went to John Cocke, not Slotnick. (The Computer Pioneer Award page on the IEEE history site notes Slotnick under “Computer Pioneers” but does not list a year of award, suggesting this is the IEEE Computer Society’s general “computer pioneers” historical recognition rather than the formal Computer Pioneer Award.)

Slotnick died on 25 October 1985 at the age of 53, of an apparent heart attack while jogging in Baltimore, Maryland.10 (Some secondary sources report Champaign, Illinois; the IEEE-CS profile, Wikipedia, and Prabook all converge on Baltimore, where he had been visiting.) He had been Professor of Computer Science at Illinois for twenty years and had additionally held a joint appointment in the Department of Landscape Architecture from 1975 – a curiosity that reflected his interest in computer-graphics applications of array processors. He was survived by his widow, who in 1985 donated his professional library to the Institute for Defense Analyses Center for Communications Research and the National Security Agency’s supercomputing research facility in Bowie, Maryland; the donation indicates the practical computing community had taken Slotnick’s library as a resource for further parallel-machine work. The first issue of the Journal of Supercomputing in 1987 carried a dedicated tribute to him.

Slotnick’s own retrospective on the project, in the Annals memoir of 1982, was characteristically candid. The most-quoted passage of that paper, used as the standard summary of the ILLIAC IV verdict, runs: “I’m bitterly disappointed, and very pleased … delighted and dismayed. Delighted that the overall objectives came out well in the end. Dismayed that it cost too much, took too long, doesn’t do enough, and not enough people are using it.”11 On the question of Fairchild Semiconductor, which had supplied the integrated-circuit memories that replaced Burroughs’s failed magnetic thin-film memory effort and saved the project from a still worse delay, Slotnick was unreserved: “Fairchild did a magnificent job of pulling our chestnuts out of the fire. The Fairchild memories were superb and their reliability to this day is just incredibly good.”12

The Daniel L. Slotnick Scholarship at the Siebel School of Computing and Data Science at Illinois (formerly the Department of Computer Science) was endowed by his friends and colleagues after his death. It awards undergraduates demonstrating “academic merit, exceptional leadership qualities, and good citizenship.” The citation captures a piece of how Slotnick was remembered at Urbana: an architect of consequence, but also a teacher and a citizen of the department.

George H. Barnes – Burroughs principal engineer

George H. Barnes was the principal engineer at Burroughs Corporation responsible for the ILLIAC IV construction effort in Pasadena, California. He is the lead author of the original architectural paper: Barnes, G. H.; Brown, R. M.; Kato, M.; Kuck, D. J.; Slotnick, D. L.; and Stokes, R. A., “The ILLIAC IV Computer,” IEEE Transactions on Computers C-17(8), August 1968, pp. 746-757.13 The author list of that paper records the team. Barnes, Brown, and Stokes were the three Burroughs engineers; Kato (Maso Kato), Kuck (David J. Kuck), and Slotnick were the three University of Illinois principals. The paper was the published architectural specification that fixed the ILLIAC IV design after the 1966 contract signing and before the major Burroughs construction effort.

Barnes’s biographical record is sparse. He does not have a Wikipedia entry; he is not in the IEEE Computer Society’s Computer Pioneer files; he does not have a published obituary that surfaces in standard searches. What can be reconstructed: he was a senior staff engineer at Burroughs’s Pasadena facility (the former ElectroData site that Burroughs had inherited when it acquired ElectroData in 1956), and he was the principal Burroughs technical interface to Slotnick at Illinois through the construction phase 1966-1971. The Burroughs Pasadena facility’s main commercial product through this period was the B5500/B6500/B6700 family of stack-architected mainframes, originally designed under Robert S. Barton in 1961-1963 (see below). Barnes appears to have come from that engineering culture: the ILLIAC IV’s I/O subsystem and front-end controller, the B6500, were the same hardware that Burroughs Pasadena was shipping to commercial customers, and the engineering practices of the ILLIAC IV programme were the engineering practices of the B6500.

Barnes was a co-author of the 1972 Proceedings of the IEEE full-system paper: Bouknight, W. J.; Denenberg, S. A.; McIntyre, D. E.; Randall, J. M.; Sameh, A. H.; and Slotnick, D. L., “The Illiac IV system,” Proc. IEEE 60(4), April 1972, pp. 369-388.14 The 1972 paper is, for the historical record, the principal published description of the as-built machine. Note that the 1972 paper’s six co-authors are all from Urbana (“The authors are with the Center for Advanced Computation and the Illiac IV Project, University of Illinois, Urbana, Ill. 61801”); Barnes does not appear on it. By 1972 the Illinois group was reporting on the system as a whole, including the Burroughs hardware, with Burroughs absent from the author list. Barnes’s last published ILLIAC IV work appears to be the 1968 architectural paper. After ILLIAC IV he disappears from the public record. The Burroughs personnel records of that era are in the Charles Babbage Institute archives but have not been fully digitised.

W. Jack Bouknight – the Urbana systems engineer

W. Jack Bouknight was the principal University of Illinois systems engineer on ILLIAC IV in the period 1969-1972. He is the lead author of the 1972 Proc. IEEE paper. His listed affiliation in 1972 is “the Center for Advanced Computation and the Illiac IV Project, University of Illinois, Urbana.” Beyond the ILLIAC IV work, Bouknight was active in early computer-graphics research at Illinois – his 1970 paper “A procedure for generation of three-dimensional half-toned computer graphics presentations” is one of the founding papers of polygonal-shading computer graphics. He also contributed RFC 76 (“Connection by name: User-oriented protocol”) and other early ARPANET RFCs from his ILLIAC IV / Center for Advanced Computation desk; the ILLIAC IV had been connected to the ARPANET via the Center for Advanced Computation and was the first remotely-shared supercomputer in the world.15

Bouknight’s biographical record beyond the ILLIAC IV period is also sparse. He moved into private industry by the late 1970s and does not appear to have published extensively after 1975. He is one of several talented mid-career University of Illinois computer scientists whose careers tracked the trajectory of the Center for Advanced Computation: prominent through the 1970s, then absorbed into a generation of CAD-and-graphics start-ups in the 1980s.

The other co-authors of the 1972 paper were: Stewart A. Denenberg, who later took a faculty position at the State University of New York Plattsburgh and wrote on computer-science education; David E. McIntyre, also at Urbana through the early 1970s; J. M. Randall, the systems-software lead; and Ahmed H. Sameh, who would have a long subsequent career in numerical linear algebra and parallel matrix algorithms. Sameh is the most prominent of the co-authors – he stayed at Illinois through the 1970s and 1980s as a professor in the Center for Supercomputing Research and Development that David Kuck would direct, then moved to Purdue in the 1990s as the Samuel D. Conte Chair in Computer Science. The 2012 Festschrift volume Parallel Numerical Computing from Illiac IV to Exascale – The Contributions of Ahmed H. Sameh (edited by Olaf Schenk, Michael Hagemann, and Anshul Gupta, Springer 2012) is the principal source on Sameh’s career and includes an extensive overview of his ILLIAC IV-era work.16 Sameh’s 1968 ILLIAC IV Document 127 (Department of Computer Science, University of Illinois Urbana), “Eigenvalue problems,” with co-author Han L., is one of the earliest systematic treatments of parallel eigenvalue computation on a SIMD architecture. The paper formed the basis for what would become Sameh’s life’s work in parallel numerical linear algebra.

David J. Kuck – the lone software person

David J. Kuck, born 4 October 1937 in Muskegon Michigan, took his BS in electrical engineering from the University of Michigan in 1959 and his PhD from Northwestern in 1963. He joined the University of Illinois at Urbana-Champaign as a faculty member in 1965, the same year as Slotnick, and was, per his Illinois Hall of Fame citation, “the lone software researcher working on ILLIAC IV … the sole software person on the ILLIAC IV project in contrast to all the other hardware-oriented members.”17 Kuck was a co-author of the 1968 architectural paper. His subsequent career – the Parafrase auto-vectorising compiler of 1977, the founding of Kuck and Associates Inc. in 1979, the Center for Supercomputing Research and Development (CSRD) at Illinois that he directed from 1986 to 1993, the CEDAR hierarchical multiprocessor that CSRD finished in 1988, and his eventual move to Intel in 2000, where he led the development of the OpenMP and high-performance compilation infrastructure that propagated parallel-programming techniques into the commercial microprocessor world – all of it traces directly to his time as the lone ILLIAC IV software person at Urbana in the late 1960s and early 1970s.

Kuck’s PhD students at Illinois became one of the principal lineages of American parallel-computing research: Duncan Lawrie, Stott Parker, David Padua (a 1980 PhD who would later succeed Kuck as a leader of the Illinois parallelisation effort), Ron Cytron, Constantine Polychronopolous, Alex Veidenbaum, Michael Wolfe (a Pacific Northwest compiler-research figure who founded The Portland Group), and Utpal Banerjee (whose 1979 PhD on data-dependence analysis is foundational to all modern compiler vectorisation). Kuck supervised more than twenty-five students over his thirty years at Illinois; many of them are the founders of the modern academic discipline of parallel-computing compilers.

The reason Kuck rather than Slotnick is the principal name in the published record of Slotnick-supervised PhD students is that Slotnick concentrated on architecture and on machine-construction supervision, while Kuck concentrated on the software that made the architecture programmable. The two faculty members had complementary specialties. The students who came to Illinois between 1965 and 1980 with an interest in parallel computing tended to do their dissertations under Kuck if they were software-focused, and under Slotnick if they were hardware-focused. The Illinois Astronomy Building (the former Center for Advanced Computation) housed both groups physically.

The Burroughs B6700 lineage and Robert S. Barton

The B6500 (and its slightly later sibling the B6700) that served as the I/O front-end and operating-system host for ILLIAC IV was a product of an entirely separate Burroughs design tradition, traceable to one architect: Robert Stanley “Bob” Barton (13 February 1925 – 28 January 2009).18 Barton was born in New Britain Connecticut, took his BA (1948) and MS (1949) in mathematics from the University of Iowa, joined IBM’s Applied Science Department in 1951, moved to Shell Oil 1954-1958 to programme the Burroughs/Datatron 205, and was hired by Burroughs in 1958-1960 as chief architect of what would become the B5000. The B5000, designed 1961-1963 and shipped from 1964, was an architecturally radical machine: a stack-based computer that executed compiled ALGOL 60 programs directly, with a hardware-managed evaluation stack rather than the accumulator-and-index-register conventional model. The architecture survived through B5500, B6500, B6700, B7700, and into the modern Unisys ClearPath MCP systems. The architectural lineage was unbroken from 1963 to the present.

Barton himself left Burroughs for the University of Utah in 1965, where he was Computer Science faculty 1965-1973 alongside David Evans, Ivan Sutherland, and Thomas Stockham. His Utah PhD students – among them Alan Ashton (later co-founder of WordPerfect), James H. Clark (founder of Silicon Graphics), John Warnock (Adobe co-founder), and Ed Catmull (Pixar co-founder) – formed the principal lineage of West Coast computer-graphics entrepreneurship of the 1980s and 1990s. After 1973 Barton returned to Burroughs full-time at the La Jolla Systems Research office. He won the W. Wallace McDowell Award in 1977 and was the first recipient of the ACM/IEEE Computer Society Eckert-Mauchly Award in 1979, “For his outstanding contributions in basing the design of computing systems on the hierarchical nature of programs and their data.”19 The 1979 Eckert-Mauchly is one of the most consequential awards in the history of the field; Barton received it for the architectural philosophy that the B5000 series embodied.

The B6500 that served ILLIAC IV was therefore a consequential architectural object in its own right, even apart from its ILLIAC IV role. It ran the Master Control Programme (MCP), which was the first commercial operating system written entirely in a high-level language (the Burroughs ESPOL extension of ALGOL); it pioneered the descriptor-based memory addressing that allowed all memory references to be checked at hardware level; and it executed compiled higher-level-language code as a first-class instruction format. The B6500 was the third generation of the Barton stack-machine lineage: B5000 (1964), B5500 (1966-67), B6500 (1969). Burroughs Pasadena began the B6500 production line in 1969, the same year ILLIAC IV’s hardware design was fixed.

The architectural quirk of the ILLIAC IV system as a whole was therefore that two radically different computer architectures sat next to each other on the same machine-room floor: the SIMD array processor (Slotnick’s, Burroughs-built, with hand-routed ECL logic) and the stack-architected mainframe (Barton’s, Burroughs-built, with conventional commercial discrete-transistor logic). The B6500 was the operating-system host; the array was the special-purpose peripheral. The 1972 Proc. IEEE paper makes the point explicitly: “From a total system standpoint, the Illiac IV array can be considered as a special-purpose peripheral device of the B6500 capable of solving certain classes of problems with extremely high speed.”20

NASA Ames replaced the B6500 with a DEC PDP-10 in 1972-73 because the PDP-10 was the standard time-sharing computer at Ames at the time, was already on the ARPANET, and was the natural connection point for the ILLIAC IV’s planned remote-access service. The replacement required a complete rewrite of the operating system, the compilers, and the I/O subsystem; Burroughs’s original B6500-based software stack was abandoned. The PDP-10-based system that emerged in 1974-75 was the version actually used in production. The original Slotnick-Barnes-Bouknight architecture, with the B6500 in front, ran briefly at Burroughs Pasadena during integration testing 1971-72 and then was retired.

Hans Mark (17 June 1929 – 18 December 2021)

Hans Michael Mark was born on 17 June 1929 in Mannheim, Baden, Germany, son of the prominent polymer chemist Herman Francis Mark. The family fled Nazi persecution in stages: Mannheim to Vienna, Vienna to Switzerland, Switzerland to London, London to Canada in late 1939, and Canada to Brooklyn New York in 1940 when his father took a professorship at the Polytechnic Institute of Brooklyn. Hans Mark became a US citizen in 1945, graduated from Stuyvesant High School in New York in 1947, took his BA in physics from UC Berkeley in 1951, and his PhD in physics from MIT in 1954.21

Mark was a research physicist at Lawrence Radiation Laboratory in Livermore from 1955 to 1958, an assistant professor of physics at MIT 1958-1960, and back at Lawrence Livermore (specifically at the Experimental Physics Division) from 1960 to 1964. In 1964 he became chairman of the Department of Nuclear Engineering at UC Berkeley. His scientific work was in experimental nuclear physics and reactor engineering.

On 20 February 1969 Mark became Director of NASA Ames Research Center, at the age of thirty-nine – the youngest director in Ames’s history.22 His Ames tenure ran until 15 August 1977, eight and a half years. The NASA Ames Hall of Fame retrospective records a vivid management style:

“It was obvious from the very beginning of Hans Mark’s tenure as Ames Director that he was a man who liked to shake things up. As the first director who came to Ames from outside rather than inside the Center’s ranks, he brought a new perspective to Ames’s mission and to the task of managing its work. … Chief among his innovations was the Ames ‘murder board’: a ruthless, uncompromising (but fair!) board of individuals who grilled those proposing new projects or lines of research. The objective was not to denigrate or humiliate, but rather to toughen a proponent’s skin and nerve for the tough presentation of their project to headquarters.”23

Mark’s role in ILLIAC IV was decisive. In late 1970 and early 1971, with the campus crisis at Urbana making continued construction at the University of Illinois increasingly untenable, ARPA needed a new home for the machine. Mark, working with Pasta (still at Illinois at that point) and with the ARPA leadership, proposed Ames Research Center as the alternative site. Ames had three advantages: it was located on Moffett Federal Airfield, an active US Navy base protected by the US Marines, so the security concerns that had driven the move were addressed; it had a strong computational fluid dynamics research programme that could provide the principal scientific workload for the machine; and Mark himself, as Director, would champion the project through the inevitable schedule slips and budget overruns. The ARPA decision to move ILLIAC IV to Ames was finalised in January 1971, and the machine was delivered in April 1972.

Mark defended ILLIAC IV through six years of difficult operation. The machine’s first operational date at Ames was 1975 – three years after delivery – and even then it was unreliable. The political pressure on the project from Congress, NASA Headquarters, and the broader supercomputing community (which by 1975 was using the much more reliable CDC 7600 and was about to adopt the Cray-1) was intense. Mark’s defence was that ILLIAC IV was specifically designed for computational fluid dynamics, that it was performing scientific work no other machine could perform, and that the experience of running it was producing parallel-computing expertise that would be essential to the next generation of NASA computing. He won that argument; the machine ran until 7 September 1981, and the parallel-computing expertise that Ames developed around it migrated, after 1981, to the Cray-2 and Cray X-MP era and ultimately to the Numerical Aerodynamic Simulation (NAS) facility that Ames opened in 1987.

After Ames, Mark moved through a remarkable sequence of senior government positions: Under Secretary of the Air Force July 1977 to July 1979 (concurrently Director of the National Reconnaissance Office August 1977 to October 1979); Secretary of the Air Force May 1979 to February 1981 under Carter and the start of Reagan; NASA Deputy Administrator July 1981 to September 1984 under Reagan; Chancellor of the University of Texas System 1984-1992; Professor of Aerospace Engineering at UT Austin 1992-2014; and Director of Defense Research and Engineering 1998-2000 under Clinton. He retired as Professor Emeritus on 1 July 2014. He died from progressive dementia at the Westminster senior home in Austin Texas on 18 December 2021 at the age of 92.24 (The user query said “1 December 2021”; the correct date is 18 December.) He was survived by his wife Marion “Bun” Thorpe (married seventy years), his children Jane Mark Jopson and James Randall “Rufus” Mark, and five grandchildren.

The Hans Mark legacy at Ames – the institutional consolidation of computational fluid dynamics as a NASA core competence, the Pioneer Venus and Pioneer Saturn missions, the cooperation with Santa Clara University, the matrix management style – was the institutional precondition for ILLIAC IV’s home there. Without Mark’s championing through 1971, the machine would not have been moved to Ames; without his defence through 1975-1977, the schedule slips might have killed it before it produced useful science.

Mel Pirtle – the day-to-day director at Ames

Mel Pirtle was Director of the ILLIAC IV programme at NASA Ames from approximately 1971 through the mid-1970s.25 Pirtle came from the University of California Berkeley computer-science programme and had been a principal at Berkeley Computer Corporation, the time-sharing-systems start-up that had spun out of Project Genie at Berkeley in the late 1960s. Project Genie had been one of the canonical 1960s time-sharing experiments (alongside MIT’s CTSS and Multics, and Dartmouth’s DTSS); Berkeley Computer Corporation had been the commercial vehicle. When BCC folded in 1971, Pirtle was hired at Ames to provide operational direction for the ILLIAC IV programme. His expertise in time-sharing operating systems was directly relevant: NASA wanted the ILLIAC IV to be remotely accessible to multiple users via the ARPANET, which required a serious time-sharing front end. Pirtle led the team that built the PDP-10-based time-sharing system that replaced the original Burroughs B6500 front end.

Pirtle’s biographical record beyond the ILLIAC IV period is also limited. He moved into private industry in the late 1970s and worked through the Silicon Valley supercomputing-and-networking ecosystem of the 1980s. The ILLIAC IV is the project for which he is principally remembered.

Paul Kutler – the principal CFD user at Ames

Paul Kutler is the most prominent of the ILLIAC IV’s NASA Ames scientific users. He graduated from Central High School Omaha in 1961, took his BS in mechanical engineering from Iowa State University in 1965 through a NASA Ames/Iowa State work-study programme, completed his MS and PhD in aerospace and mechanical engineering at Iowa State in the late 1960s, and was hired by NASA Ames as a research scientist directly out of his doctoral studies.26 He spent the next three decades at Ames in computational fluid dynamics. His career titles included Chief of the Applied Computational Aerodynamics Branch, Acting Director of the Aerodynamics Division, Manager of High Performance Computing and Communications, Director of the Consolidated Supercomputing Management Office, Deputy Director of the Information Sciences and Technology Directorate, and Director of West Coast Programs. He took a Master of Science in Management from Stanford in 1985 alongside his NASA work. After NASA he founded Flow Simulations Inc. and taught at San Jose State University. He was a Fellow of the American Institute of Aeronautics and Astronautics.

Kutler’s ILLIAC IV-era work was principally on hypersonic and transonic flow over arbitrary geometries – the kind of computation that the SIMD architecture of ILLIAC IV was particularly suited for, because the underlying numerical methods (finite-difference solutions of the Euler and Navier-Stokes equations) were highly regular and parallelisable across the geometric grid. His co-authors at Ames included Harvard Lomax (the senior CFD figure at Ames through the 1970s and 1980s), Thomas H. Pulliam (who took his PhD at Iowa State in 1979 and was, with Kutler, one of the principal Ames CFD researchers from the late 1970s onward), Joseph Steger, and Denny Chaussee. The development of the three-dimensional implicit codes by Lomax and Pulliam for ILLIAC IV is one of the principal computational achievements of the machine’s run; the codes that Pulliam wrote in the 1970s for ILLIAC IV evolved into the ARC2D and ARC3D codes that are still used in NASA aerospace simulation today, decades after the machine itself was decommissioned. The Lomax-Pulliam code lineage is, in CFD historiography, the principal scientific legacy of the ILLIAC IV.

Frances Bauer (5 July 1923 – 1 November 2022) and the Bauer-Garabedian-Korn airfoil codes at Courant

Frances Renee Brand Bauer was born on 5 July 1923 in New York City, took her AB at Brooklyn College in 1943, an MS at Brown University in 1945, and a PhD in applied mathematics at Brown in October 1948 under William Prager.27 Her dissertation, “An extension of the Prandtl-Reuss theory of plastic flow,” was an early woman’s PhD in American applied mathematics. She joined the Courant Institute of Mathematical Sciences at NYU in the early 1950s and remained there for the next sixty-plus years; she died on 1 November 2022 in New York at the age of 99.

The Bauer-Garabedian-Korn airfoil method (often called the BGK method or the supercritical-wing-section method) was a Courant Institute project of the early 1970s. Frances Bauer, Paul Garabedian, and David Korn developed a series of two-dimensional FORTRAN codes for the analysis and design of transonic airfoils, exploiting the hodograph transformation and analytic continuation into the complex domain. The principal codes, often referred to as Programs A, B, and H, were widely distributed in the CFD community and became the standard tools for transonic-airfoil design through the late 1970s and early 1980s. The published works that codified the method were Bauer, Garabedian, and Korn, A Theory of Supercritical Wing Sections, with Computer Programs and Examples (Springer, 1972), and the Supercritical Wing Sections II: A Handbook (1975) and III (1977) follow-ups.

The relationship of the Bauer-Garabedian-Korn codes to ILLIAC IV is indirect. The BGK codes themselves were developed and primarily run on the CDC 6600/CDC 7600 series at Courant and at NYU’s ERDA-funded Mathematics and Computing Laboratory, which was a serial-machine environment, not a SIMD-machine environment. The BGK method was a complex-domain method requiring sequential evaluations that did not parallelise naturally onto a SIMD array. The CFD work that ran on ILLIAC IV at Ames was instead the explicit and implicit finite-difference method work of the Lomax-Kutler-Pulliam group, which did parallelise onto a regular geometric grid. The two CFD code-lineages – the Courant-NYU complex-method lineage of Bauer, Garabedian, and Korn, and the Ames-NASA finite-difference lineage of Lomax, Kutler, and Pulliam – ran in parallel through the 1970s on different machines for different purposes. Both contributed to the broader CFD literature; the BGK codes were used at Ames for inverse airfoil-design problems (where the desired pressure distribution is given and the airfoil shape is to be found) while the ARC2D/ARC3D codes were used for direct flow simulations on given geometries.

Bauer’s work is therefore relevant to the ILLIAC IV story principally as background on the broader CFD environment of the 1970s, not as direct ILLIAC IV computation. Her significance in her own right is independent: she was one of the principal women applied mathematicians of the postwar generation, a senior mathematician on Project Cyclone (the missile-systems computer laboratory at Reeves Instrument Corporation) in the 1950s, a six-decade member of the Courant Institute, and an advocate for women in computing whose Bauer Outreach Program at NYU’s Women in Technology group continues today.

The 1970 University of Illinois crisis and the move to Ames

The user query asked specifically about “the Christmas Eve 1970 bombing” of the Digital Computer Laboratory at the University of Illinois. After extensive search of the Daily Illini archive, the News-Gazette (Champaign-Urbana), the University of Illinois Alumni Association’s “Flash Point” retrospective, the University Archives’ “Birth of the Computer Age” account, the Student Life and Culture Archives timeline, and the LibGuides archive of the U-of-I in the Cold War era, I can find no reference to a Christmas Eve 1970 bombing of the Digital Computer Laboratory or any other University of Illinois building. The user’s query appears to combine several facts incorrectly. What did happen, and what is documented:28

6 January 1970: The Daily Illini publishes an article describing the ILLIAC IV project, revealing that approximately two-thirds of the machine’s time would be reserved for Department of Defense use including nuclear-weapons-related computations. The article triggers organised opposition.

24 February 1970: The ROTC lounge in the University of Illinois Armory is firebombed. No injuries; minimal damage. The firebomb does not damage ILLIAC IV (which was at Burroughs Pasadena, not at Urbana, at that time – only design work was at Urbana).

Spring 1970: Multiple firebombs are thrown at the Champaign Police Department, the Air Force Recruiting Station in Urbana (which suffers $100 000 in damage), and another is found unexploded in Lincoln Hall. Two student suspects, including Larry Allan Voss (a sophomore in a wheelchair), are arrested. The University of Illinois is in active and substantial campus turmoil.

30 April 1970: Edgar Hoults, a 23-year-old Black security guard at the Illinois Central Railroad, is shot in the back and killed by Champaign Police Officer Fred Eastman during a chase. The killing intensifies campus tensions and generates additional protests.

4 May 1970: Kent State University: four students killed by Ohio National Guard. Nationwide campus protests, including a “day of Illiaction” student strike at Urbana on 9 May 1970.

24 August 1970: The Sterling Hall bombing at the University of Wisconsin Madison destroys most of Sterling Hall and kills the postdoctoral physicist Robert Fassnacht, who was working on a CDC 3600 in the basement. Sterling Hall housed the Army Mathematics Research Center, the principal target. The bombing is the single most consequential event in the ILLIAC IV-relocation decision: it makes clear that violent attacks on military-funded research computing on US university campuses are no longer hypothetical. Slotnick, by Wikipedia’s reconstruction, is “increasingly concerned that the machine would be subject to attack by the more radical student groups.” This is the position that drives the relocation request to ARPA.

6 January 1971: The Daily Illini publishes the announcement that the ILLIAC IV will be moved to NASA Ames Research Center rather than installed at the University of Illinois. Hans Mark of NASA Ames and the ARPA leadership are credited with the decision.

April 1972: Burroughs delivers ILLIAC IV to Ames. Installation in the Central Computer Facility, building N-233.

Christmas Eve 1970: Nothing of consequence. There was no December bombing at Urbana that I can locate in the documentary record. The user may be conflating the 4 May 1970 Kent State events, the August 1970 Sterling Hall bombing, the various Urbana firebombings of February-April 1970, and the January 1971 announcement of the move into a “Christmas Eve 1970” event. The reader-friendly summary, for a blog post, is that the violence at Wisconsin in August 1970 plus the threat of similar violence at Illinois in 1970-1971 drove the ARPA decision, finalised in January 1971, to move ILLIAC IV from Urbana to NASA Ames. No one was hurt in any incident specifically tied to the ILLIAC IV machine at Urbana; the machine itself was at Burroughs Pasadena throughout the period of the campus crisis.

The names of the Sterling Hall bombers (relevant for context on the violence Slotnick feared) were the four members of the New Year’s Gang: Karl Armstrong, Dwight Armstrong (his brother), David Fine, and Leo Frederick Burt. Karl, Dwight, and Fine were eventually arrested and convicted; Burt has remained a fugitive since August 1970 and is one of the longest-running FBI Most Wanted entries in American history. The Wisconsin connection mattered to the ILLIAC IV decision because of geographic proximity (Madison and Urbana are both Big Ten universities in the upper Midwest, six and a half hours’ drive apart); because of institutional similarity (both were land-grant universities with Department-of-Defense-funded research centres); and because the New Year’s Gang was a recognisable model that Urbana-based imitators could plausibly follow.

Slotnick’s later career and legacy

Slotnick’s last fifteen years – from the relocation of ILLIAC IV to Ames in 1971 to his death in 1985 – were spent teaching, supervising students, and writing.29 He held his Computer Science professorship at Urbana through this period, with the joint Landscape Architecture appointment from 1975. The 1972 Proc. IEEE paper was the last major architectural paper of his career. After 1972 he turned to historical and reflective writing on parallel computing, culminating in the 1982 Annals of the History of Computing memoir. He served as a consultant to ARPA, NASA, and a number of private industry clients. He was elected an IEEE Fellow in 1976 and won the McDowell Award in 1983.

The teaching at Illinois 1972-1985 was, by departmental memory, characteristically idiosyncratic. Slotnick taught both undergraduate and graduate courses on parallel computer architecture, and his graduate seminars were the principal channel through which the ILLIAC IV experience was communicated to the next generation of Illinois computer scientists. The students who passed through those seminars included many of the same figures who appear in Kuck’s PhD-supervision lineage (Lawrie, Padua, Polychronopolous, etc.), suggesting that Kuck-Slotnick was effectively a team and that students moved fluidly between the two. The memorial tribute in the first issue of the Journal of Supercomputing (1987) recorded that the influence of his graduate teaching extended well past his publication record.

The judgement on Slotnick’s career has shifted over time. The contemporary 1981-1985 view was unfavourable: ILLIAC IV was technologically interesting but commercially and operationally unsuccessful, and the more pragmatic alternatives (CDC’s CDC 7600 of 1968, Cray’s Cray-1 of 1976) had captured the supercomputing market. By 1985, with the Cray X-MP and Y-MP shipping and the parallel-vector-multiprocessor architecture entrenched, ILLIAC IV looked like a dead end. The retrospective view of the late 1990s and 2000s is more favourable: the SIMD architectural paradigm that Slotnick pioneered would, in different silicon and at different scales, become the basis for the Connection Machine (1985-1994), for the SIMD instruction extensions of every modern microprocessor (Intel MMX 1996, SSE 1999, AVX 2011, AVX-512 2017), and for the massively parallel architecture of the modern GPU. The compute capability of an Nvidia H100 in 2024 is descended, architecturally, from Slotnick’s 1962 SOLOMON sketch and his 1972 ILLIAC IV. Slotnick had been right about the architecture all along; he had been wrong only about the silicon era it would arrive in.

The Slotnick verdict, in the Annals memoir of 1982, is that the architecture was successful and the engineering was painful. Reading the memoir against the eventual SIMD-on-GPU outcome, the verdict is correct.

Summary of the team

The ILLIAC IV team for blog purposes can be summarised in seven figures:

  • Daniel L. Slotnick (1931-1985): Architect, principal investigator, idea-source, project champion. The intellectual centre of the project from conception (1952) through completion (1972) through retrospective (1982).
  • George H. Barnes: Burroughs Pasadena principal engineer, lead author of the 1968 architectural paper, the Burroughs interface to Slotnick. Largely undocumented after 1972.
  • W. Jack Bouknight: Urbana systems engineer, lead author of the 1972 Proc. IEEE paper, the principal Urbana team-lead.
  • David J. Kuck (born 1937): Lone software person at Urbana from 1965, Slotnick’s faculty colleague through the project, founder of the Illinois parallel-compilers tradition. Trained the students who carried the work forward.
  • Robert S. Barton (1925-2009): Designer of the Burroughs B6500/B6700 stack-architected mainframe that served as the original front-end controller. Eckert-Mauchly Award 1979.
  • Hans Mark (1929-2021): NASA Ames Director 1969-1977, the institutional advocate who got ILLIAC IV to Ames in 1971-72 and defended it through the difficult 1972-77 commissioning period.
  • Paul Kutler (born 1943): Principal NASA Ames CFD user, with Lomax and Pulliam, of the ILLIAC IV from 1972 onwards. The Ames CFD code lineage that ran on ILLIAC IV (ARC2D, ARC3D) is the principal scientific legacy of the machine.

The supporting cast: Mel Pirtle (Berkeley/BCC, day-to-day Director at Ames); Ahmed Sameh (Urbana, parallel numerical linear algebra, later Purdue); John Pasta (Urbana department chair who recruited Slotnick); Frances Bauer at Courant (parallel CFD lineage on serial machines); the University of Illinois antiwar movement of 1970-71 (the campus radicals whose threats motivated the move); the New Year’s Gang of Wisconsin (the bombers whose 24 August 1970 attack on Sterling Hall made the threat concrete).

Footnotes

  1. IEEE Computer Society Profile: “Daniel L. Slotnick,” https://www.computer.org/profiles/daniel-slotnick; Prabook World Biographical Encyclopedia: “Daniel Leonid Slotnick,” https://prabook.com/web/daniel_leonid.slotnick/230926; Computer History Museum, “November 12: ILLIAC IV Designer Slotnick Born,” https://www.computerhistory.org/tdih/november/12/. ↩

  2. IEEE Computer Society Profile, op. cit. The PhD year (1956) is consistent across the IEEE-CS profile, Prabook, and the Wikipedia article. The institution (Courant Institute at NYU) and the field (Applied Mathematics) are both confirmed. ↩

  3. Slotnick, D. L., “The Conception and Development of Parallel Processors: A Personal Memoir,” Annals of the History of Computing 4(1), 1982, pp. 20-30. IEEE Xplore link: https://ieeexplore.ieee.org/document/4392921/. The memoir is the principal primary source for Slotnick’s recollection of his time at the IAS. ↩

  4. IEEE-CS Profile, op. cit.: “1957-1960: IBM Development Laboratory, Poughkeepsie, New York.” The Cocke-Slotnick paper of 1958 is the principal output. ↩

  5. IEEE-CS Profile, op. cit.: “1960-1965: Westinghouse Electric Corporation, Baltimore”; the Wikipedia article reaches the same conclusion: “In 1960, Dr. Slotnick was employed by the Westinghouse Electric Corporation in Baltimore, Maryland, where he was given the opportunity to pursue his ideas on parallel computers.” ↩

  6. IEEE-CS Profile, op. cit.: “1965: Professor of Computer Science at University of Illinois at Urbana-Champaign.” Pasta’s recruitment of Slotnick is documented in the IEEE Computer Society Pasta biography page, “John R. Pasta,” https://history.computer.org/pioneers/pasta.html: “Pasta recruited Daniel L. Slotnick to the Computer Science Department in May 1965.” ↩

  7. IEEE-CS Pasta page, op. cit. ↩

  8. The August 1966 contract date is recorded in multiple sources including the Wikipedia ILLIAC IV article and the Good Science Project review. ↩

  9. IEEE-CS Profile, op. cit.: “1976: Elected IEEE Fellow.” ↩

  10. Wikipedia, “Daniel Slotnick,” https://en.wikipedia.org/wiki/Daniel_Slotnick: “died in Baltimore, Maryland, on October 25, 1985, age 53, from an apparent heart attack while jogging.” The IEEE-CS Profile and Prabook converge on the same date. ↩

  11. The “I’m bitterly disappointed” quote is most commonly cited from secondary syntheses including the Wikipedia ILLIAC IV article, but originates in Slotnick’s own Annals memoir of 1982 and in interviews of the 1981-1982 period. See Wikipedia, “ILLIAC IV,” https://en.wikipedia.org/wiki/ILLIAC_IV. ↩

  12. Same source. ↩

  13. Barnes, G. H.; Brown, R. M.; Kato, M.; Kuck, D. J.; Slotnick, D. L.; and Stokes, R. A., “The ILLIAC IV Computer,” IEEE Transactions on Computers C-17(8), August 1968, pp. 746-757. IEEE Xplore link: https://ieeexplore.ieee.org/document/1687448/. ↩

  14. Bouknight, W. J.; Denenberg, S. A.; McIntyre, D. E.; Randall, J. M.; Sameh, A. H.; and Slotnick, D. L., “The Illiac IV system,” Proc. IEEE 60(4), April 1972, pp. 369-388. PDF copy at the University of Auckland school of computer science: https://www.cs.auckland.ac.nz/courses/compsci703s1c/resources/Bouknight-ILIAC-IV.pdf. ↩

  15. Bouknight’s RFC contributions are listed at dblp: https://dblp.org/pid/124/5297.html. ↩

  16. Schenk, O.; Hagemann, M.; and Gupta, A., eds., Parallel Numerical Computing from Illiac IV to Exascale – The Contributions of Ahmed H. Sameh, Springer 2012. Springer link: https://link.springer.com/chapter/10.1007/978-1-4471-2437-5_1. ↩

  17. Wikipedia, “David Kuck,” https://en.wikipedia.org/wiki/David_Kuck; Grainger College of Engineering Hall of Fame: “David J. Kuck,” https://grainger.illinois.edu/alumni/hall-of-fame/9645. ↩

  18. Wikipedia, “Robert S. Barton,” https://en.wikipedia.org/wiki/Robert_S._Barton: “Robert Stanley ‘Bob’ Barton (February 13, 1925 – January 28, 2009) was the chief architect of the Burroughs B5000 and other computers such as the B1700, a co-inventor of dataflow architecture, and an influential professor at the University of Utah.” ↩

  19. Wikipedia, “Eckert-Mauchly Award,” https://en.wikipedia.org/wiki/Eckert-Mauchly_Award: “Robert S. Barton: 1979. The Burroughs B5000 and successor large systems.” ↩

  20. Bouknight et al. 1972, op. cit., section IV. ↩

  21. Wikipedia, “Hans Mark,” https://en.wikipedia.org/wiki/Hans_Mark; UT Austin Aerospace Engineering memorial: https://www.ae.utexas.edu/people/faculty/faculty-memorials/hans-m-mark. ↩

  22. NASA Ames History Office, “Hans Mark,” https://history.arc.nasa.gov/hist_pdfs/bio_mark.pdf: “Mark joined Ames as its Center Director in 1969, serving until 1977.” ↩

  23. NASA Ames Hall of Fame retrospective, op. cit. ↩

  24. UT Austin News, “University of Texas Mourns Loss of Former UT System Chancellor, Aerospace Engineering Professor and NASA Administrator Hans Mark,” 21 December 2021: https://news.utexas.edu/2021/12/21/university-of-texas-mourns-loss-of-former-ut-system-chancellor-aerospace-engineering-professor-and-nasa-administrator-hans-mark/; NASA, “Hans Mark,” https://www.nasa.gov/people/hans-mark/. ↩

  25. ed-thelen.org “ILLIAC IV” page, https://ed-thelen.org/comp-hist/vs-illiac-iv.html: “Mel Pirtle … was engaged as the ILLIAC IV’s director.” ↩

  26. Central High School (Omaha) Foundation Hall of Fame, “Paul Kutler,” https://chsfomaha.org/halloffame/paul-kutler/; Iowa State University Aerospace Engineering Hall of Distinguished Alumni, https://www.aere.iastate.edu/alumni/hall-of-distinguished-alumni/. ↩

  27. Women Becoming Mathematicians database, “Frances Renee Brand Bauer,” https://womenbecomingmathematicians.net/db/forties/brown/bauer1948/: “Born 5 July 1923, New York NY … PhD (Applied Math) October 1948, Brown University … adviser William Prager … Died 1 November 2022, New York NY”; NYU Women in Technology, “Bauer Outreach Program,” https://www.nyu.edu/life/information-technology/about-nyu-it/women-in-technology.html. ↩

  28. University of Illinois Alumni Association, “Flash Point,” March 2020: https://uiaa.org/2020/03/25/flash-point/; University of Illinois Library Student Life and Culture Archives timeline 1970-1979: https://www.library.illinois.edu/slc/research-education/timeline/1970-1979/; Wikipedia, “ILLIAC IV,” section on relocation; Wikipedia, “Sterling Hall bombing”: https://en.wikipedia.org/wiki/Sterling_Hall_bombing. ↩

  29. Daniel L. Slotnick Scholarship at the Siebel School of Computing and Data Science, https://siebelschool.illinois.edu/about/awards/undergraduate-scholarships-awards/daniel-l-slotnick-scholarship: “Slotnick served as a Professor in the Siebel School of Computing and Data Science until his death in 1985.” ↩