BOPS, Inc. quietly began auctioning off its patent portfolio last month, signaling the end of its operations as a vendor of high performance licensable DSP cores. Given the current business climate, the failure of another licensable core vendor is hardly surprising. (For more examples, see our story on Lexra.) However, BOPS had a few things working against it beyond the industry slow-down.
The BOPS architecture was based on a complicated, scalable array of powerful processing elements. At its 1997 introduction, this design was much faster than those offered by established processor and licensable core vendors. A 2 × 2 array of these elements, for example, could easily outperform the fastest mainstream DSPs of the day.
BOPS initially targeted this architecture at the “media processor” market, along with companies like Chromatic, MicroUnity, and Philips. These companies fielded high-performance VLIW architectures targeting multimedia-intensive applications. But the media processor market has been slow to develop, and media processor vendors have had difficulty finding volume applications for their architectures.
Later, BOPS repositioned its architecture as a licensable core for a wider range of signal processing applications, including voice-over-IP and wireless LANs. When BOPS hired Carl Schlachte as its CEO in 2000, much was made of his earlier tenure at ARM, Inc. Indeed, it was largely ARM’s success—coupled with the ready availability of venture capital—that attracted companies like BOPS to the processor core licensing business in the late 90’s. But while ARM cores focused on simplicity and low cost, BOPS targeted high performance with a complex architecture. Much like the media processor market, the market for high-performance licensable cores has yet to emerge.
Licensable cores are used in custom chips, which in turn are used mainly in high-volume applications. Because such applications tend to be extremely cost-sensitive, core licensees typically seek the smallest, simplest core that can get the job done—in short, the opposite of BOPS’ architecture. Hence, BOPS’ combination of a high-performance architecture and an IP-licensing business model severely limited its potential customer base.
The BOPS architecture was an impressive technical achievement. But with only two licensees announced over the course of the company’s five-year history, BOPS’ demise may prompt other licensors of high-performance processor cores, such as TriMedia and Siroyan, to reconsider their strategies.
Add new comment