With more and more products incorporating computer vision functionality, semiconductor vendors are increasingly designing specialized processors for vision applications. For product developers, this means that identifying the best processor (whether a chip for use in a system design, or an IP core for use in an SoC) is challenging—for several reasons.
First, processors for vision vary greatly in their architectures, making them tough to compare. Second, there is no universal set of benchmarks to use in evaluation because vision applications and algorithms are quite diverse. Third, new processors are appearing rapidly, expanding the range of options. And fourth, due to intense competition, many processor vendors publish very little information about their products.
A company recently contacted BDTI for assistance in selecting a vision processor for a next-generation design. Its VP of engineering faced a classic dilemma: how to make a decision with too little time to examine all factors completely. For his design, he needed a processor that would give the best combination of performance, low power, low cost, good programmability, and low risk. But it was equally important to move quickly, due to an aggressive development schedule.
Fortunately, BDTI tracks processor suppliers on an ongoing basis and was able to move quickly and with confidence. Step one was to line up candidates; the client had identified some, but had overlooked other strong options. BDTI compiled a complete list, including a new processor under development which came to light through BDTI's trusted relationships with vendors.
Next BDTI compared the candidates. The schedule was so tight that it wasn’t feasible for BDTI to use its preferred approach—collecting software development tool suites, and implementing benchmarks. There wasn't even time for the vendors to implement benchmarks specified by BDTI. Instead, BDTI examined benchmarks that the vendors had previously implemented and selected a set of algorithms common to all which were available in optimized form. From this set, BDTI created benchmark metrics relevant to the client's application.
Leveraging existing code drastically reduced the work required but didn't completely eliminate it. On examination of the benchmark results, BDTI noticed discrepancies that required further analysis. For example, on what fabrication process node was the speed, power consumption, silicon area and other information gathered, and at what lot-to-lot variation "corner" of that process node? Did the power consumption include the entire on-chip clock distribution network, the external memory controller, the system bus interface, and other peripheral circuits? BDTI needed to answer these and other questions before a meaningful comparison between candidates was possible.
By leveraging existing code and independently analyzing data to ensure apples-to-apples assessments, BDTI was able to provide a comprehensive, detailed comparison of the processor options in less than 90 days. The client declared the project an unqualified success, saying "We are very happy with the way you handled the project and results we achieved. This gives us great ideas on how to move forward."
If you're selecting a processor, core, or SoC for a vision processing application, BDTI is your one-stop shop for advice, evaluation, and recommendations. Minimize your product development risk by taking advantage of BDTI's unique understanding of the vision processor market and its world-class skill in benchmarking. Contact Jeremy Giddings for a confidential consultation at +1 (925) 954-1411 or giddings@BDTI.com.