How computational modeling and simulation took root in Texas.
Texas computing grew from targeted, practical applications. And the real story is a largely invisible, numerical one.
Advanced computing took root in Texas, especially UT Austin and Houston, to serve concrete needs: working the land, and defending it. Oil and aerospace.
The real story is the largely invisible world of numerical modeling and simulation, or what would later be called computational engineering and science. Popular science promoted rockets and thinking machines, but not the quiet work of solving huge systems of equations in hardware floating-point on the most advanced machines available.
The document argues this history is worth preserving through the code, the machines, and above all the linear algebra, because the methods themselves are what changed the course of whole industries and scientific fields.
Foundational algorithms carried the field from rooms of human computers with desktop calculators to automated, hardware-integrated code, from Los Alamos through Los Angeles and Houston to Austin.
Foundational methods, all built on basic linear algebra subprograms executed in hardware floating-point on the fastest machines available.
Linear Programming
George Dantzig on RAND's IBM CPC. Optimized refinery blending and logistics, and drove early industrial computing.
Successive Over-Relaxation
David Young at TRW, then UT. Made iterative solving practical for aerospace, especially heat and fluid flow problems.
Alternating Direction Implicit
Rachford, Peaceman, and Douglas at Exxon precursor Humble Oil, and later their colleague Mary Wheeler. Enabled subsurface reservoir simulation on early hardware, from the IBM CPC onward.
Satellite Geodesy · Utopia
Byron Tapley, Bob Schutz, and the UT Orbit Processor, using satellites as test particles to map the Earth's gravity field.
Together they moved the field from rooms of human computers with desktop calculators to automated, hardware-integrated computational engineering and science.
Five forces that turned Texas into a computational powerhouse.
Lyndon Johnson's influence and the 1958 Defense Education and NASA Acts channeled federal money into Texas, anchoring NASA in Houston and transforming UT Austin's engineering and computing programs within a decade.
Industry researchers spread real-world projects to the universities, bringing along Big Science methods and federal funding. David Young's move from TRW to found the UT Computation Center in 1958 is the archetype.
The oil industry's need to solve blending and reservoir problems made companies like Exxon early major mainframe buyers, driving the commercial market's development and a Houston to Austin research corridor.
The push for numerical fidelity drove the fitful, contested adoption of hardware floating-point, from Zuse's wartime machines to the IBM 704, the zoo of incompatible formats, and eventually IEEE 754. UT's long association with the Minnesota tradition traces this history.
CDC 1604 → CDC 6600 → CDC Cyber → Cray X-MP
The 60-bit CDC world of Fortran and linear algebra, embodied in Utopia, now thrived alongside a new interactive, exploratory culture that arrived with timesharing and the DEC-10: AI research using Lisp, machine translation building on the strong German heritage in Texas, automatic program verification, and a vibrant student hacker community embodied in the Star Trek and Decwar game lineage. UT became the first ARPANET node in Texas in 1977.
Waypoints in the mainframe era
Dantzig and Orchard-Hays implement Simplex on their IBM CPC at RAND in Los Angeles, birthing foundational linear algebra methods such as the Revised Simplex Method and the Product Form of the Inverse.
Rachford, Peaceman, and Douglas at Exxon precursor Humble Oil in Houston make digital reservoir simulation practical with the Alternating Direction Implicit method on their IBM CPC. David Young's SOR method is in use on aerospace projects at TRW precursor Ramo-Wooldridge in Los Angeles.
David Young leaves TRW to found the UT Computation Center. Humble Oil donates their IBM CPC, and Al Matsen and his students carry it into Welch Hall themselves.
UT becomes the first university with a CDC 6600, the world's first commercial supercomputer. 60-bit hardware floating-point in a purpose-built underground home beside the Tower.
The University of Texas Orbit Processor becomes operational on the CDC 6600, turning satellites into test particles to map the Earth's gravity field with centimeter precision.
The HRC DEC-10 brings modern interactive timesharing to campus. Pioneering work brings 18-player DECWAR to life and leaves life-long marks on those who experience it. UT becomes the first ARPANET node in Texas.
UT installs the first Cray supercomputer at a Texas university, pushing gravity models and subsurface simulation into the vector-computing era.
Because these models were fundamentally computational linear algebra, they can be faithfully reconstructed, run, and explored. The machines, tools, methods, and beginner's textbook-style problems are all well-known.
The document's forward-looking proposal is for Docker Time Capsules. Self-contained, version-controlled environments that bundle emulated hardware, operating systems, tools, code, and documentation with AI as an enabling partner.
Regardless of the source code's origin, the linear algebra, assembly code, and Fortran are effectively the same, so the models that changed whole industries and scientific fields can be reconstructed and run faithfully.
The aspiration is to become a kind of Feynman Lectures of modeling. First-principles clarity and minimalist purity, preserving the models that changed the course of industries and scientific fields.
Longer excerpts and a detailed section-by-section summary will be available here.