Kiev computer

The Kiev (Russian: «Киев», Ukrainian: «Київ») was a Soviet vacuum-tube digital computer developed in Kyiv at the laboratory that had built MESM — the first stored-program computer in continental Europe — and completed under the direction of Viktor Glushkov. Working from 1956 and fully finished in 1958, it was the flagship machine of the new Computing Center of the Ukrainian Academy of Sciences and a vehicle for some of the era's boldest ideas: a programming language with pointer arithmetic realized in hardware, image input and output devices for pattern-recognition experiments, and — in 1960 — remote control of a steel plant's converter shop from more than 500 kilometres away.^[1]

Inheritance: the Lebedev laboratory

The Kiev's lineage is direct. Sergei Lebedev's Kyiv laboratory built MESM (1951) — Europe's first continental stored-program machine — before Lebedev left for Moscow in 1952. The laboratory passed to the Institute of Mathematics of the Ukrainian Academy, whose director, the probabilist Boris Gnedenko, initiated a successor machine in 1954 with essentially the same engineering collective: senior scientists Lev Dashevsky and Kateryna Yushchenko, with engineers Kateryna Shkabara and Solomon Pogrebinsky.^[2]

Glushkov arrived in March 1956 from Sverdlovsk to head the laboratory — famously insisting on hiring Alexander Letichevsky over Gnedenko's objection — and supervised the final engineering, assembly, and debugging. In December 1957 the laboratory became the Academy's Computing Center, which completed the machine in 1958 and became, in 1962, the Institute of Cybernetics — today named for Glushkov.^[3]

The machine

The Kiev was an asynchronous, parallel-action computer: its functional blocks ran autonomously and passed control by handshakes rather than a central clock — a then-unusual architecture emphasizing the independence of its units.^[4]

┌───────────┐ ┌───────────┐ ┌───────────┐ ┌───────────┐
│Arithmetic │ │ Internal  │ │ External  │ │ Program-  │
│   unit    │ │  storage  │ │ storage   │ │ storage   │
│           │ │ (ferrite) │ │ (drums,   │ │ device    │
│ 16 ctrl + │ │ 1024 wd   │ │ tape)     │ │ (ferrite- │
│ 8 arith + │ │ ~10µs     │ │ ~6000–    │ │ xformer   │
│ 4 logical │ │ cycle     │ │ 9000 wd)  │ │ ROM,      │
│  ops      │ │           │ │           │ │ 512 wd    │
└───────────┘ └───────────┘ └───────────┘ └───────────┘
 ┌───────────┐ ┌───────────┐ ┌───────────┐
 │  Central  │ │   Input   │ │  Output   │
 │  control  │ │ cards /   │ │ paper /   │
 │  unit     │ │ tape /    │ │ photo-    │
 │           │ │ *images*  │ │ *plates*  │
 └───────────┘ └───────────┘ └───────────┘

Two Kiev machines were built: one at the Kyiv Computing Center (in service 1956–1962+), a second purchased by the Joint Institute for Nuclear Research at Dubna.^[5]

Pointers in hardware

The Kiev's most remarkable feature sits at the intersection of hardware and language theory. In 1955, Vladimir Korolyuk and Kateryna Yushchenko had created the Address programming language — one of the world's first high-level languages, built on two principles of machine operation: program control and addressing, including manipulation of addresses of higher ranks — what we now call pointers to pointers.^[6]

The Kiev implemented that idea in hardware: its instruction set included a unique group operation on addresses performing n-fold pointer dereference — repeated application of the dereference operator in one machine operation — dramatically accelerating work with abstract data types like lists. In the West, linked lists arrived with IPL in 1956 and an explicit pointer type with PL/I in 1964; the Kiev had n-fold dereference in hardware by 1958.^[7]

The feedback ran both ways: the address language shaped the machine (the "P" operation was included at Yushchenko's suggestion, alongside the NOO/OOO group operations), and the machine made the language fast. Programs were written in address language and translated by PP translator programs — an early "programming processor" pipeline.^[8]

What it computed

Beyond routine Academy work — structural mechanics, physics, the mathematical problems of a growing republic academy — the Kiev ran experiments that read like a survey of computing's future:

• Industrial remote control (1960): experiments in remote control of technological processes at the Dneprodzerzhinsk metallurgical combine's converter shop — over 500 km from the machine. Long before "teleoperation" was a word, a computer in Kyiv was steering a steel plant in the Dnipro region.^[9]
• Pattern recognition and "intellectual process" modeling: two custom peripherals — an image-input device reading from paper, photographic film, or plates, and an image-output device writing to paper or exposing photo plates — let Glushkov's group model learning of pattern recognition and goal-directed behavior: machine recognition of simple geometric figures, automata recognizing printed and handwritten letters, automatic synthesis of functional circuits. The institute's museum credits the Kiev as the first system of its kind for digital image processing.^[10]
• "Avtodirektor": one of the first database-management systems of relational type, written in the address language.^[9]

The book — and the American translation

In 1962 Glushkov and Yushchenko published Вычислительная машина «Киев» — математическое описание ("The Kiev Computer: A Mathematical Description") — a full engineering-level description: the basis for choosing the machine's parameters, its operation set, standard and test routines, and the address-language translation system. The U.S. Air Force's Foreign Technology Division found it important enough to translate in 1966 (FTD-TT-65-972, 227 pages) — the Kiev documented for American readers as a sample of Soviet design.^[11]

Lineage

The Kiev sits in the middle of a Kyiv line that runs from MESM (1951) through the Kiev (1956–58) to Dnipro (1961 — among the USSR's first semiconductor machines), Promin (1963 — stepwise microprogram control), and the MIR series (1965 onward — engineering workstations that are arguably ancestors of the personal computer). Its laboratory became the Institute of Cybernetics in 1962; Glushkov went on to design the doomed OGAS national computer network — and the institute still bears his name.^[12]

Timeline

References

• V. M. Glushkov & Ye. L. Yushchenko, Вычислительная машина «Киев» — математическое описание, Kiev 1962 — primary source; USAF FTD translation FTD-TT-65-972 (1966), available at icfcst.kiev.ua / infoua.net
• Glushkov Institute of Cybernetics museum pages — Kiev spec sheet, "Theoretical Programming in Ukraine," "Founder of IT in pictures" — icfcst.kiev.ua/MUSEUM
• Glushkov Institute museum — Kiev spec sheet; Dubna, remote-control and pattern-recognition experiment accounts
• Memoirs of Victor Glushkov (DataArt edition) — the 1956 arrival and the lab's formation
• B. Malinovsky, histories of computing in Ukraine — lineage context

See also

• MESM — the Lebedev predecessor in the same laboratory
• Viktor Glushkov · Kateryna Yushchenko · Boris Gnedenko
• Address programming language — the software half of the story
• Dnipro computer · MIR computer — the successors