From the trace information in the run_summary in the Vitis Analyzer, navigate to the output port for which you want to calculate the throughput (
upscalekernel in this case). Add a marker at the start of the first output sample as highlighted as follows. Then click the
Go to last timeicon, and observe the cursor moves to the end of the last iteration. Now, click the previous transition icon to go the start of the last iteration. Add one more marker at the end, and observe the time difference as
The number of bytes transferred is
128 samples * 4 Bytes * 7 iterations=
Throughput = 3584/2282 * e-9 ~= 1.5 GBPS.
Theoratically, AI Engine can transfer four bytes per cycle (in this case, 0.8 ns). So, to transfer
3584bytes of data, it requires 896 cycles (896 * 0.8 = 716 ns). So, the theoratical througput is 5 GBPS.