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.
Just as a carpenter needs accurate measurement tools to build a quality piece of furniture, processor and subsystem designers need good measurement tools in order to produce superior products. By serving as a proxy for the target applications, benchmarks reveal critical insights regarding the design and its application suitability. This was vividly illustrated recently by a semiconductor vendor in the course of a major new processor development program.
The chip vendor originally selected multiple BDTI benchmark suites for use in marketing activities. To ensure that benchmark results would be available at product launch, the company’s engineering team began implementing the benchmarks while the processor design was being finalized. This early start proved fortuitous: in the course of carefully implementing the benchmarks, the engineers discovered that some of the features of their design did not deliver the expected capabilities. In addition, the designers realized that a few simple additional features would yield significant performance gains in a wide range of applications.
Because these changes were identified early, the design team was able to implement them relatively quickly and easily, resulting in a much stronger product, and saving millions of dollars compared to the cost of making such changes after the product was launched.
To learn how to use benchmarks for engineering purposes, contact
Add new comment