https://hgpu.org/?p=20025
Abstracting OpenCL for Multi-Application Workloads on CPU-FPGA Clusters