Ibm 7090
The IBM 7090 was a large, transistorized scientific mainframe computer announced by IBM in December 1958 and delivered to customers from 1959. It was the transistor-based successor to the vacuum-tube IBM 709, and it was in effect the 709's architecture rebuilt with solid-state circuits. It used 36-bit words, a magnetic core memory of 32,768 words, and was several times faster than the machine it replaced. By the early 1960s the 7090 was the standard tool for large scientific and engineering calculation, and its installations are one of the clearest signs of the moment when transistors moved from the laboratory and the portable radio into the heart of major computing.
The 7090 was a straightforward translation of a tube machine into transistors, and that is exactly why it mattered commercially. IBM kept the instruction set and programming model of the 709 so that existing software could carry over, and replaced the vacuum tubes with germanium transistor circuits. Eliminating tubes cut heat, power use and the failure rate, and allowed a faster cycle. Its speed is often quoted in the region of 100,000 additions per second, a loose figure that depends on the operation and the memory access involved, and the better comparison is that IBM described it as many times faster than the 709. The word length of 36 bits, rather than the 8-bit byte later standardized, reflected the scientific heritage of the line: a word could hold a floating-point number with good precision, or six 6-bit characters.
The programming language Fortran II made the 7090 usable by engineers rather than just by machine-code specialists. Fortran, created at IBM for the 704 in the mid-1950s, was a high-level language that translated formulas into machine code, and Fortran II added subroutines and separate compilation. On a 7090, a scientist could write a calculation in near-algebraic notation and rely on the compiler to produce efficient code, which made the machine far more productive than raw speed alone suggests. The 7090 also ran batch operating monitors that processed streams of jobs on cards and tape, and its descendants hosted early time-sharing experiments, including MIT's Compatible Time-Sharing System, which ran on a modified IBM 7090 series machine in the early 1960s.
Its most public applications tied it to the Cold War and the space program. The US government used 7090 systems for ballistic missile early warning, and NASA used 7090 machines at its centers to calculate trajectories and support the Mercury flights, where reliable and fast orbital computation was essential. In a quite different field, IBM's SABRE reservation system, developed with American Airlines, used a pair of 7090 computers at its core as it developed from a pilot system around 1960 into a full network, and is often cited as one of the first large online transaction systems. The same machine was therefore used both for missile defense and for taking airline bookings, an early illustration of how general-purpose computers spread through very different institutions.
A frequent misunderstanding is to treat the 7090 as the first transistorized computer, or as a microelectronic design. Neither is true. It followed experimental and smaller transistor machines, and it was built from discrete transistors, diodes and resistors mounted on circuit cards, with large cabinets, dedicated machine rooms and a purchase price in the millions of dollars. What it proved was that discrete transistor logic could be manufactured and maintained reliably by a large company at scale. Its limits, such as the count of connections and the cost of hand-wired assembly, were part of the reason the industry turned next to integrated circuits.
Today the 7090 is remembered through the software and ideas it carried rather than through its hardware. Fortran, compilers, batch operating systems and online reservation all matured on machines of this class, and the 36-bit word persisted in later architectures for years. Any history that moves directly from the 1947 transistor to the integrated circuit leaves out this middle chapter, in which transistors replaced tubes in the big, expensive machines that organizations depended on.
Read IBM 7090 as the moment the vacuum-tube mainframe became a transistor mainframe: the same programs ran, only faster, cooler and more reliably, and that was enough to carry computing into space tracking, early warning and airline booking.