I developed matrix component using HLS. It takes 6 floating point input, 6x6 matrix and sends 6 floating point output. The input arrives every 25 microseconds and the output should be available every 10 microseconds or less. The input is available through the Avalon memory mapped interface. I want to know whats the fastest way for the accelerator to read these inputs every sample? For eg., should the inputs be stored in a BRAM/register and streamed to the matrix component?
Please suggest some ideas.
Thanks in advance!
If I understand correctly, you are trying to receive an input at 25us, perform some computations and send the signal out at 10us. To determine how fast the input can be received or output can be sent, it depends on how fast you design can run.