FFT transforms are configured using the Configuration channel. The configuration information carried in this channel, and how it is packed, is discussed in more detail in Configuration Channel. When the core is ready to load a new frame for processing, the core checks if a new configuration has been supplied on the Configuration channel. If present, the FFT processing core is configured using that information before the frame is loaded.
In non-blocking configuration mode, if no new configuration information is supplied, the core processes the frame using the last configuration it had. If no configuration has ever been supplied, it uses the defaults described in Resets.
In blocking configuration mode, the core waits for a new configuration on the Configuration channel before loading a new frame. If no new configuration information is supplied, no new frame is loaded as discussed in Blocking Runtime Configuration.
The process of applying configuration data to a particular frame depends on the current status of the core:
- To apply a configuration to the very first frame after power on or after an idle period.
- To apply the configuration to the next frame in a sequence of frames.