Feature: 'Writing Efficient Game Code for Next-Gen Console Architectures'

In today's main Gamasutra article, SN Systems' Andy Thomason analyzes the key concepts needed in order to write efficient code for all next-gen consoles, in an important article for any programmers working or considering working on next-gen titles. In the introduction to his piece, Thomason suggests: "Code performance is vital to a successful game title, and poor performance can lead to shipping delays and cutting of content as well as a much more expensive game development process. Getting code generation right has to start from the beginning of a project, as the architecture of a game system determines the performance as much as anything else. A naïve approach to architecture design will spell disaster for a project. Things are only getting more complicated, since next-gen consoles have introduced us to multiple cores, longer pipelines, Level 2 caches, and multi-thread issues. Most next-gen consoles have chosen to pack more computing elements onto the chips than most workstations. This comes at the price of the omission of instruction renaming and large caches intended to make legacy applications run faster. The plus side is that next-gen consoles are capable of hundreds of gigaflops, or about a thousand Cray-1s, the most powerful supercomputer available in 1976 when I was first learning to code. This article illustrates a few of the key elements that lead to high performance code in the system as a whole." You can now read the full Gamasutra feature on the subject, including details on technical specifics aimed to help next-gen development (no registration required, please feel free to link to the article from external websites).