Week 1 Codefest - zanzibarcircuit/ECE510 GitHub Wiki

The first prompt gave me code. I followed up asking for a test bench. I ran the testbench in EDA playground, and the initial code failed to spike. I asked for code to fix this, and it worked. I'm just going to assume this works for now.

Then I prompted it to make the network and create a testbench. It also seems to work based on the provided output, but I have no idea. I also wonder if EDA playground is the best tool since the file management seems to be nonexistent, and I had to copy all modules into a single file. I think I should download the other suggestion: Icarus Verilog + GTKWave. I'm assuming it would be better.

After creating the SPI interface and top-level module, I downloaded the file. I have no real way to test this.

To compare mine to the paper, I'd need to understand more about the spiking neuron itself. It seems as though a few differences are that they use protection for when the potential goes less than 0 or above 255.

I still need to try OpenLane, do ReLU, and think about improvements.