LARC Scientific Compiler(ID:2193/)
Fortran for the LARC
Compiler developed by CSC for the LARC from FORTRAN II. Legendary for its power and expressivity as well as its speed.
Had advanced array sorting features - perhaps a benefit from CSC having just finished written FACT?
Betty Snyder expressed disatisfaction with the delivery of the compiler because the source was assembler patched and the documentation poor even though the compiler was outstanding
Q: Would you describe some of the major contracts which CSC has obtained?
A: One of our first projects included development of the FACT language, the design of its compiler, and implementation of most of the FACT processor.
Later, CSC developed the LARC Scientific Compiler, an upgraded variant of the FORTRAN II language, and designed the ALGOL/ FORTRAN compiler for the RCA 601.
Recently, we have finished other complete systems for large scale computers such as the UNIVAC III and 1107. By complete systems, I mean the algebraic and business compilers, executive system, assembly program and other routines.
For the Philco 2000, we developed ALTAC IV, COBOL-61, and a very general report generator.
An assembly program for a Daystrom process control computer, a general sort-merge for another computer, several simulators, and many others at this level of effort have been accomplished.
In areas concerned with applications, we developed two PERT/COST systems for aircraft manufacturers and are responsible for the design and implementation of a command and control system at Jet Propulsion Lab for use by NASA in many of our space probes scheduled later this year and beyond.
In the scientific area we have worked on damage assessment models, data acquisition and reduction systems, re-entry and trajectory analysis, maneuver simulation, orbit prediction and determination, stress analysis, heat transfer, impact prediction, kill probabilities, war gaming, and video data analysis.
In the commercial applications area, we have produced systems for payroll and cost accounting, material inventory, production control, sales analysis and forecasting, insurance and banking problems, etc.
Right after getting my Ph.D. in 1963, I started to work hard preparing the chapter on sorting. I knew hardly anything about sorting, but I thought it would be nice to read up on the subject and to toss a chapter about sorting into a book about compilers, especially because the LARC scientific compiler had just come out, and it was reputedly based on the idea of sorting the data in unusual ways. I found that sorting was really interesting and pretty soon I found myself digging into lots of technical articles.
in Annals of the History of Computing 4(3) July 1982 view details
The original design envisioned 35,000 transistors, and even before construction began, the estimated cost had grown to double the contract price. To help out, the Navy agreed to buy a second LARC with 30,000 words of memory for its David Taylor Model Basin, outside of Washington, D.C.
By late 1957, the designers estimated that the LARC would use 70,000 transistors. The Philadelphia group began building the LARCs in the summer of 1958, and the Livermore machine was finished in May 1960, 27 months behind schedule. It was shipped to California, where it passed its acceptance test in August. The Navy's LARC was completed in October and passed its acceptance test in February 1961. Much of the delay arose from the difficulties inherent in building such an innovative machine; however, other priorities of the company also intruded. For example, Herman Lukoff, the LARC project director, had to go to St. Paul for three months in the summer of 1957 to help out with the problems of getting the Univac II to work. Programmers in Philadelphia wrote a simple operating system called L'Opera, which handled queuing programs for the CU. The provision of a Fortran compiler was beyond the resources of the company, so that project was farmed out to Computer Science Corporation (CSC). Betty Snyder Holberton, one of EMCC's original employees who had left Remington Rand to work at the David Taylor Model Basin, was not happy with the CSC compiler: "They delivered the compiler, which was beautifully designed, very poorly documented, and patched at the machine level so we couldn't recompile it or reassemble it." Once these initial problems were overcome, the LARCs gave excellent service. The machine at Livermore was retired in December 1968, while the one at David Taylor continued in service until 1969. Within Sperry Rand, the picture was not so bright: The LARC project had cost $19 million, of which not quite $6 million had been recovered. While the LARC was being developed, there had been advances in transistor design, and no additional customers could be found at the asking price of $6 million for a machine with older transistors. Extract: Origins of LARC
In 1954, the University of California Livermore Radiation Laboratory, which was funded by the Atomic Energy Commission, requested proposals for a high-speed scientific computer. Its goal was to have a computer that was significantly faster than any machine currently in production or development, something that would stretch the limits of technology. Both St. Paul and Philadelphia were interested in responding, but Eckert obtained corporate authorization for Philadelphia to submit Sperry Rand's only bid, and it was presented in April 1955. The Livermore laboratory chose Sperry Rand's bid over a proposal by IBM (which went on to have its high-speed machine, nicknamed Stretch, funded by the rival Los Alamos Laboratory). A contract for the Livermore Automatic Research Computer (LARC) was awarded to Sperry Rand in September 1955 at a price of $2.85 million (see Fig. 3). The scheduled completion date was September 1957. The design, a joint effort by the Livermore laboratory and Eckert's group, called for a computing unit (CU) and a separate input/output processor (IOP), with the capability of later adding a second CU.
in IEEE Annals of the History of Computing, 20(3) July-September 1998 view details