Generated by DeepSeek V3.2| UNIVAC LARC | |
|---|---|
| Name | UNIVAC LARC |
| Manufacturer | Remington Rand |
| Designer | Seymour Cray |
| Release date | 1960 |
| Price | $6 million (approx.) |
| Operating system | Master Control Program (MCP) |
| Memory | 98,304 12-digit words (core memory) |
| Storage | Magnetic tape, IBM 729 tape drives, drum memory |
| Predecessor | UNIVAC 1103 |
| Successor | UNIVAC 1107 |
UNIVAC LARC. The UNIVAC LARC, formally the **Livermore Advanced Research Computer**, was one of the earliest supercomputers and a landmark in high-performance scientific computing. Designed primarily for the Lawrence Livermore National Laboratory, it represented a massive leap in processing power for nuclear weapon simulation and computational fluid dynamics. Only two systems were ever built, with one delivered to Livermore, California and the other to the United States Navy's David Taylor Model Basin.
The LARC project was initiated by Remington Rand's UNIVAC division in the mid-1950s to create a machine an order of magnitude faster than any existing computer. It was conceived specifically to meet the intense computational demands of the Atomic Energy Commission for weapons research at Lawrence Livermore National Laboratory. As a transistorized computer, it eschewed vacuum tube technology, utilizing over 60,000 germanium transistors and aiming for performance in the range of 250,000 to 1,000,000 operations per second. Its development was fraught with technical challenges and cost overruns, ultimately limiting its commercial viability but cementing its status as a pioneering system in the history of computing.
The LARC featured a highly innovative, though complex, multiprocessor design centered on a dual CPU configuration. The main computational section consisted of a high-speed arithmetic unit and a separate input/output processor, which could operate concurrently to manage data flow from peripheral devices. Its memory hierarchy was advanced, employing fast magnetic core memory for main storage, supplemented by larger-capacity drum memory and banks of IBM 729 magnetic tape drives. The system's instruction set was designed for scientific calculation, with hardware support for floating-point arithmetic and a 48-bit word length that provided high precision for Fortran programs.
Development was led by a team under the renowned computer architect Seymour Cray, who would later found Cray Research. The project faced significant hurdles, including the immaturity of transistor technology and the immense difficulty of cooling such a dense, powerful machine. These challenges caused substantial delays and escalated costs, prompting Remington Rand to cancel the commercial version. Ultimately, only the two contracted machines were completed: the first, known as the **Livermore Computer**, was delivered to Lawrence Livermore National Laboratory in 1960 after a strenuous acceptance test involving a complex neutron diffusion calculation. The second was delivered to the David Taylor Model Basin for naval architecture and hydrodynamics research.
At Lawrence Livermore National Laboratory, the LARC became a workhorse for classified nuclear weapons design and simulation, running critical codes for phenomena like fission and hydrodynamic shock waves. It also contributed to unclassified scientific research in areas such as astrophysics and magnetohydrodynamics. The system at the David Taylor Model Basin was used extensively for analyzing ship hull designs and propeller performance. Both machines were operated around the clock, with programming primarily done in Fortran and assembly language, and were supported by a sophisticated operating system known as the Master Control Program. They remained in service for nearly a decade before being superseded by newer, more cost-effective machines like the CDC 6600.
Although not a commercial success, the UNIVAC LARC had a profound impact on the trajectory of supercomputing. It served as a critical learning experience for Seymour Cray, directly influencing the streamlined, high-performance designs of his later machines at Control Data Corporation and Cray Research. The project demonstrated the feasibility and necessity of multiprocessing, advanced memory hierarchy, and dedicated input/output processors for scientific computing. Furthermore, its development for Lawrence Livermore National Laboratory solidified the partnership between national laboratories and computer manufacturers in pushing the boundaries of processing power, a model that would drive the supercomputer race for decades. The LARC stands as a monumental, if rare, first step into the era of transistorized supercomputers.
Category:Supercomputers Category:UNIVAC computers Category:One-of-a-kind computers Category:1960 introductions