Intel has shipped its 60-core Intel Xeon Phi coprocessor 5110P to selected customers. First unveiled in June this year, the coprocessor comes in the form of a PCIe expansion card and operates independently of the host operating system, courtesy of its own Linux operating system that manages each x86 core.
The Xeon Phi series is part of Intel's High Performance Computing (HPC) program and it can be considered the successor to Intel's defunct Larrabee project that attempted to compete with GPUs in the realm of highly-parallelized streaming processing. The coprocessor is used to boost calculations used for mathematics, graphical tasks and scientific experiments that are performed on servers and supercomputers, alongside their regular processors.
The high-end chip in the 3100 family will offer more than 1,000 gigaFLOPS (1 TFLOPS) double-precision performance, while the Intel Xeon Phi coprocessor 5110P provides additional performance at a lower power envelope of 225W TDP. It is capable of reaching 1,011 gigaFLOPS (1.01 TFLOPS) double-precision performance, and supports 8GB of GDDR5 memory with a memory bandwidth of 320 GB/s.
According to Intel, the Intel Xeon Phi coprocessor 5110P has shipped last week, with general availability on January 28, 2013 at a recommended customer price of US$2,649. The Intel Xeon Phi coprocessor 3100 product family will be available during the first half of 2013 with SRPs of below US$2,000. Additional information on availability and ordering Intel Xeon Phi coprocessor 5110P can be found at here.