What Is Amdahl's Law?

When it comes to computers, no one likes waiting. Amdahl's law shows how much latency you can take from a performance task by introducing parallel computing. Amdahl's law is named after Gene Amdahl, who introduced it in his work in the 1960s. It's used primarily in predicting the theoretical maximum speedup for program processing using multiple processors, and it's also known as Amdahl's argument. The law states that if the task comprises two parts, one serial port and one parallel part and if the serial part takes up 90% or more of the total time required to complete the task, then adding more processors will not help increase its overall speed. So, how much faster will your computer be once you buy that new graphics card? That's a good question, and there's a simple answer: Amdahl's law. You can apply it to find out how much your computer will speed up by adding more processors, GPUs, or whatever else you've got in mind. The concept is simple: multiple processors can work on a given task simultaneously, speeding up a charge by a particular factor. That's where Amdahl's law applies in finding that factor's benchmark. In a perfect world, we'd have the time to do everything. We could wake up whenever we enjoyed ourselves and sleep when needed. We could work on whatever project we wanted, at any time of day or night, and if it was a tricky one, we could take as long as we needed to figure out how to solve it. That's not how life works. In the real world, we're always on a deadline: an assignment due at school or a report that needs to be sent tomorrow afternoon. If one thing helps you meet deadlines more efficiently, it's having multiple processors working for you simultaneously! That's why multicore processing is so necessary: it helps us get more done in less time by allowing us to split tasks among multiple cores simultaneously.