Von Neumann Bottleneck

The von Neumann bottleneck looks at how to serve a faster CPU by allowing faster memory access. Part of the basis for the von Neumann bottleneck is the von Neumann architecture, in which a computer stores programming instructions, along with actual data, versus a Harvard architecture, where these two kinds of memory are stored separately. These types of setups became necessary as simpler, preprogrammed machines gave way to newer computers requiring better ways to control programming and information data.

Computer scientists have attempted to address the von Neumann bottleneck in various ways. One is to place critical memory in an easily accessible cache. There is also the idea of multithreading, or managing multiple processes in a triaged system. Other potential tools, like parallel processing, or changing the memory bus design, also work on the idea of decreasing this "bottleneck" or, in a phrase commonly used with this issue, increase the bandwidth for memory coming in and out of the processor.

Other ideas for "fixing" a von Neumann bottleneck are more conceptual. Experts have posited various "non-von Neumann" or "non-von" systems, some modeled around the biological world, which would allow for more distributed memory intake, versus the linear system used in conventional computing. Some ideas involve other emerging technologies, such as where a "memrister" or other nanoscale component could help with memory processing. The diversity of ideas around the von Neumann bottleneck show how integral this idea is to evaluating computing's potential as it has emerged over the last few decades.