Designing a full architecture for software that runs on a GPU and FPGA is very different from CPU-software.
We are specialised and experienced in designing optimal data-flows for data to be processed on accelerators. We assist in porting the architecture for high-throughput and/or low-latency.