1948

Manchester Baby

Manchester Baby runs its first program from read-write memory

The machine was built to test a memory tube, and its first job searched for a factor by doing subtraction over and over. Its whole memory held 1,024 bits.

11 a.m., 21 June 1948

At about 11 a.m. on 21 June 1948, the Small-Scale Experimental Machine — nicknamed the Manchester Baby — successfully ran a program stored electronically in its cathode-ray-tube memory at the University of Manchester.

Tom Kilburn wrote the 17-instruction program, which searched for the highest proper factor of an input number by repeated subtraction. The first program was entered by hand, one bit at a time, before the start switch was pressed.

A computer built around a memory tube

Frederic C. Williams and Kilburn needed a rigorous, sustained test of their cathode-ray-tube random-access memory, so they built the smallest practical computer around it. Geoff Tootill assisted in construction and later documented the early program.

The successful run showed that the Williams–Kilburn store could hold both instructions and data reliably enough to control an electronic computer. Its capacity was 32 words of 32 bits — 1,024 bits in total.

The 52-minute run came later

The input used in the first successful run was comparatively small. The celebrated calculation for 2^18 — 262,144 — which took 52 minutes, was performed within the following days rather than necessarily on 21 June.

That benchmark involved about 2.1 million instructions and approximately 3.5 million store accesses; the often-quoted 3.5-million figure refers to the accesses, not the instructions.

The word “first” is where the dispute sits. Computer historians have shown that a modified ENIAC ran a coded program from switch-based read-only memory in April 1948, so whether Manchester was first depends on whether the memory must be read-write.

Sources

Researched 23 Aug 2026 6 sources 2 audit passes Date disputed

How this was checked

Researched from the web into a fact sheet, rewritten from that sheet, then audited against it by a different model. How the pipeline works →

Date disputed. Sources disagree; the research places it at 1948-06-21.

What the sources leave uncertain

The date and successful run are well documented; the dispute concerns the word “first.”
Manchester was the first machine generally credited with executing instructions from the same electronic read-write memory used for data.
Computer historians have shown that a modified ENIAC ran a coded program from switch-based read-only memory in April 1948. Whether that counts as the first “stored-program” execution depends on the definition used.