Sign up for our newsletter and get the latest big data news and analysis.

A Passing Lane for the Scalability Limit

Over at Admin HPC, Dell’s Jeff Layton writes that parallelized applications eventually hit a scalability limit where adding processors no longer increases performance. The good news is that you can use Amdahl’s law as a guide to where to concentrate your resources to improve performance.

The first explanation for this behavior is described by Amdahl’s Law. This law illustrates the theoretical speedup of an application when running with more processes and how it is limited by the serial performance of the application. At some point, your application will not run appreciable faster unless you have a very small serial portion. One of the big sources of serial bottlenecks in an application is I/O. Although I didn’t spend much time discussing it, examining the amount of time an application spends on I/O is an important step in understanding the serial portion of your application. Several tools and libraries can help you parallelize the I/O portion of your application; however, at some fundamental level, the application has to read input files and write output, limiting the scalability of the application (the scalability limit).

Read the Full Story.

Resource Links: