Efficient channelization on a graphics processing unit

We present an implementation of a channelizer (F-engine) running on Graphics Processing Unit (GPU). While not the first GPU channelizer, we have put significant effort into optimizing implementation. are able to process four antennas each with 2 Gsample/s, 10-bit dual-polarized input and 8-bit output, single commodity GPU. This fully utilizes available PCIe bandwidth The system is as optimized for high-bandwidth antenna, but handles 6.2 limited by single-core CPU performance.