Processor designers, marketers, and users with a sophisticated understanding of benchmarks know that raw benchmark results rarely give the most accurate picture of processor performance for a specific application scenario.  While useful for providing a general impression of processor capabilities, raw benchmark results must be adapted to give a clear sense of how processors will perform in a particular application.

One of the things I really love about digital signal processing technology is how, year by year, it gets easier to create new things.  Things like clever new audio or video compression algorithms, for example.

Forget about needing mainframes, minicomputers, or expensive engineering workstations to evaluate that new algorithm. Today we’ve got powerful PCs and easy-to-use simulation environments that make the whole process relatively painless.

Note: This article has been changed on 7/18/2007 from its original version.

MIPS has introduced the MIPS 74K, a new, high-performance synthesizable general-purpose microprocessor core. The 74K targets demanding multimedia and networking applications, such as H.264 and WiMaX, and according to MIPS, the core has already been shipped to initial licensees.

There’s been a lot of press lately about start-up companies offering multi-core DSP chips.  What’s less widely discussed is that large, established DSP chip vendors have been offering multi-core DSP chips for years.  These chips have been popular in “channelized” applications where workload partitioning is fairly straightforward.

Chip and IP vendors typically utilize benchmarks for marketing purposes—specifically, to demonstrate the capabilities of their products to prospective customers.  But processor vendors that use benchmarks for marketing purposes alone are missing half the picture:  during the design of a processor, subsystem, or chip, good benchmarks are invaluable for ensuring that the design is as good as it can be.

Introduction