CPS 209 Lecture 11: Lecture 11 Advanced Concurrency in Java 8+
Document Summary
Lecture 11: advanced concurrency in java 8+ lecture. The famous moore"s law states: computing speed (or more accurately, frequency) doubles every 18 months. From its inception in 1965, this prediction has held up astonishingly well. However, this cannot possibly continue forever, since there are hard physical (speed of light, electronic noise, quantum uncertainty) and economic limits for this exponential growth. An easier way to speed up many computations is: parallelize them for multiple processors and processor cores. However, computational problems vary greatly in how much they can theoretically be parallelized. Amdahl"s law notes: that every computation consists of an inherently sequential part and parallelizable part. Using p processors, the total running time will be t = ts + tp / p. At one end of this continuum: the inherently sequential problems for which t = ts. These problems consist of a series of stages so that the stage n + 1 cannot begin until the stage n has been completed.