Silicon Mechanics now offers the Intel Xeon Phi coprocessor, based on Intel Many Integrated Cores technology. According to the company, the coprocessors deliver higher aggregate performance for highly parallel applications, and eliminate much of the need for dual-programming architecture.
"We're proud to be offering the Intel Xeon Phi coprocessor in our hybrid computing server line," said Ken Hostetler, director of product management at Silicon Mechanics. "We know that the programming model supported by Intel Many Integrated Cores architecture will be right for many implementations. Our customers are already excited about it."
Intel Xeon Phi coprocessors are designed to provide efficient performance for highly parallel applications that benefit from numerous mathematical calculations performed at once. They have many more, smaller cores, many more hardware threads, and wider vector units than multi-core Intel Xeon processors. The high degree of parallelism compensates for the lower speed of each individual core to deliver higher aggregate performance for highly parallel applications.
For more information, visit Silicon Mechanics.
Sources: Press materials received from the company and additional information gleaned from the company's website.