The profiling plots should include the distribution of the input layer / input data and the box for the input data type.

Here's a profiling plot for a "good" version of a model:

Then I changed config['LayerName']['fc1_input']['Precision']['result'] = 'ap_fixed<2,1>' (clearly not enough bits for the inputs), and get this profiling plot:

The plots look quite similar, but the accuracy of the first is 75% (as expected) while the second achieves 28%. It would be made really obvious if the profiling plots include the input data type. It should look something like this - but at the top of the existing activations plot:

The actual implementation needs a bit of care, since the input is not a real "layer" like the others.