AirStack Sandbox is an FPGA development kit for customers to deploy their own intellectual property (IP) on the AIR-T‘s FPGA. The Deepwave team has abstracted away all of the complications of writing custom firmware and has now enabled our customers to interact with the FPGA in an extremely simplified interface. Board specific system management (power, clocking, timing, and bus architecture) are handled by Deepwave developers in the Sandbox software. Customers can immediately utilize FPGA computational resources without the long initial development times typical of bringing up an entire system.
With Sandbox + AirStack, the full power of the AIR-T’s FPGA, GPU, and CPUs are now at your disposal. Customers are able to insert their own IP cores into the FPGA for their own application which may include
- Custom signal processing blocks,
- Communication physical layers,
- Pre-processing for deep learning applications, or
- Feature extraction.
The figure to the right shows the unique shared memory architecture of the AIR-T product line. With the AIR-T, the GPU and CPU share the same memory allowing for signal data to be equally accessible to both compute resources. The signal data is transferred to/from this shared memory using the DMA engine included in Sandbox.
This is in contrast to a traditional software defined radio shown to the left. In typical SDR systems, the GPU and CPU are only connected by a PCI express bus requiring an extra memory copy of the signal data to access the GPU compute resource. This extra copy adds significant latency to the application.
The presence of the GPU / CPU shared memory on the AIR-T allows for applications to operate with reduce latency while still leveraging the compute power of the GPU. By combining the AIR-T hardware, AirStack, and AirStack Sandbox, Deepwave offers the only SDR with combined FPGA, GPU, and CPU in a fully reconfigurable infrastructure.
AirStack Sandbox leverages IP cores that are licensed for free from Xilinx when using Vivado. The JESD204B license must be obtained from Xilinx; however, the evaluation version is compatible with AirStack Sandbox.