Lesson 8 of 81967–today8 min
Buses, interrupts and parallel hardware
In 1967 Gene Amdahl, one of the architects of the IBM System/360, gave a short talk at the AFIPS Spring Joint Computer Conference that is still quoted in every course on parallel computing. His point was arithmetic: the part of a job that cannot be parallelised caps the speed-up, however many processors you add.
For decades the warning seemed academic, because single processors kept getting faster. In 2001 IBM's Power4 put two processor cores on one chip, a first by IBM's own account, and within a few years the whole industry had followed. Amdahl's arithmetic became everyday engineering.
The rest of this lesson is part of Pro
Create a free account to read the first lesson of every path, and go Pro for the rest.
Sources
- Validity of the Single Processor Approach to Achieving Large Scale Computing Capabilities (G. M. Amdahl, AFIPS 1967) — Stony Brook University (course reading)
- The IBM Power4 — IBM History
- A Brief History of CUDA — Beyond3D