Parallelism executes multiple computational activities simultaneously, typically on separate processors or cores. It differs from concurrency: concurrency manages several tasks at once, while parallelism performs them at the same time. Go supports this style through lightweight goroutines, channels for communication, and runtime-managed scheduling. section 9.0
Use parallelism when independent work can proceed together: CPU-bound computations can use multiple cores, while concurrent I/O work can reduce waiting during network, disk, or database operations. The same approach supports data pipelines, worker pools, background jobs, real-time feeds, device coordination, and responsive interfaces. section 9.6.2 section 9.1
Fetching RSS feeds concurrently
An RSS reader can launch one goroutine per feed instead of fetching sources sequentially. Calling printFeedTitle with the go keyword starts independent work for the New York Times and BBC feeds, allowing network waits to overlap. A production design should coordinate completion rather than rely on an arbitrary sleep. section 9.1
Practical parallelism patterns include:
- Run independent network, file, or database operations concurrently to reduce idle waiting. section 9.6.2
- Use worker pools to control parallelism while distributing queued jobs across goroutines. section 9.6.2
- Pass data between concurrent pipeline stages with channels. section 9.6.2
