My view of CRPC started substantially before its official start 11 years ago. I can recall, as if it were yesterday, when I met Ken Kennedy during the fall of 1982. Ken was in the Mathematical Science Department at Rice, and Convex Computers had 20 people and no product. We shared a vision of easy-to-use supercomputers, ultimately moving to the desktop and becoming commonplace technology. We envisioned that compilers and tools for vector and multiprocessor systems would be as easy to use and as efficient as existing single-processor scalar systems. Many people doubted this vision would ever occur. But we both realized that this objective had to be met if supercomputing was to become mainstream.

We began working together and with various students of Ken's, within an ad hoc structure that existed in the U.S. supercomputing research community. While this worked, it became clear that a formal center was needed to help focus the many diverse efforts toward this common vision. Also, this center had to be geographically dispersed. We needed to "follow the people."

Fortunately for the industry and what would become the CRPC, NSF began an aggressive funding of both supercomputer centers and supercomputing research. Ultimately, this led to the Science and Technology Center effort. Ken and many of the members of the community rallied. This was needed to help move supercomputing software to the next generation. Ken was asked to head this effort. After many discussions, he agreed. This meant a change in both his own research efforts and his role at Rice. During many, many meetings with the various people at Rice (president, provost, etc.), the Rice External Engineering Advisory Board gave its complete support for the project.

The CRPC, in its first meetings, reflected the enthusiasm of its members and excitement filled the air. But various organizational impediments remained. With the continued efforts of Ken, the CRPC moved from an organization in transition to a center inhabited with world-class researchers, with world-class research and worldwide impact. This could not have occurred without the extraordinary leadership in both research and management by Ken.

So sadly, the CRPC is coming to an end, after 11 years of hard work and significant accomplishments. Its legacy will last well into the next millennium. I cannot help but think of the recently announced Vector-based home computers, with ads in consumer-oriented publications, depicting comparative performance for numerical algorithms like fft, convolution, etc. Thinking back to the Fall of 1982, our shared vision has been met and this is the legacy of Ken Kennedy and the CRPC that will remain.