Multi-threading is quite possibly the most misunderstood concept from a consumer’s stand-point in computer science. I can’t even count the number of times I’ve heard statements such as “what’s the point with 4 cores, when current programs only supports two cores?”. So I thought I’d take it upon myself to clear this up, once and for all. No (sane) application is written for two, four or <insert arbitrary number> cores. Either it supports multiple cores, or it supports one. There’s no in-between. A well-written multi-threaded application will have a significant performance gain for each core you toss at it! Why? Well this is very simple; from an application stand-point, the CPU is completely abstract. The OS handles all the task scheduling, including threading. So in the program, I would just say “hey, I need another thread for doing some fancy-pansy time consuming quantum physics math”, and the OS will open up a new thread which may run on another core than the main application (this is entirely up to the OS).

So to clarify even further, you don’t program directly for a specific core, you just open a thread which will be allocated somewhere on either the core on which the program is running, or on a new core. If I fire off 6 threads in my application, and it’s running on a quad-core, it is very likely that I’ll be using all cores. If I run the same program on a dual-core, it will most likely run on both cores, and so on.

The main advantage in multi-threading on PC, still lies in the enhanced multi-tasking capabilities though, especially when we’re starting to talk about 12 core CPU’s. But either way, a program is either multi-threaded, or it’s not. A well-written application will utilize the cores you have available to you.