Step 1: Sync Scalar - 2022.2 English

Vitis Tutorials: AI Engine (XD100)

Document ID
XD100
Release Date
2022-12-01
Version
2022.2 English

Kernel execution depends on the availability of windows of data on their inputs, and the space to write windows of data on their outputs. This example illustrates a complementary method, whereby a kernel will only get triggered to run after a write of data from another processor. This can be an ARM® processor or another AI Engine.

In this example, a write from an Arm processor causes a partial sine wave to be generated using the direct digital synthesis (DDS) kernel on the AI Engine. The Arm processor can control the frequency of the sine wave by writing different values to the runtime parameter. In the HW and HW cosim flow, the AI Engine output is streamed to the PL kernels, and the PS controls the running AI Engine and PL. The following figure shows this example.

design structure and flow

Note: The default working directory in this step is “step1”, unless specified explicitly otherwise.