An FPGA implementation for a high-speed optical link with a PCIe interface

To achieve speedup for multi-node, multi-GPU computing platforms, it is necessary to overcome performance bottlenecks in networks based on Ethernet or Infiniband. This paper describes an FPGA implementation of a custom network interface for an optical link between PCIe buses of compute nodes. The implementation uses an Altera Stratix IV chip with integrated PCIe interface logic and high-speed input/output for connecting optical fiber interfaces. The interface is designed with control and buffering for concurrent data transfers. A software driver enables application programs on the host computer to use the high-speed link. A bandwidth of 8.5 Gbit/s was achieved between software applications, exceeding bandwidth reported in recent work [7].