ChipWhisperer GUI interfaces: First feedback after having tried both, the older v4 GUIs and the new jupyter v5 environment


Hi Colin, Hi all,

after now having worked a bit with the “give-away” CW-nano and the Jupyter environment and with the older CW4 GUI with our recently purchased CW-Lite-ARM, I am having the following preliminary feedback.

For the capture part, I do believe that a GUI-based tool might be very helpful also in the future. For instance for adjustment of gain and adjustment of H-field probes an interactive GUI in the style of an oscilloscope might be very helpful. Also for adjustment of the LNA-gains and other HW parameters a GUI-style environment might be desireable.

Possibly, once you have found the optimum parameters, setting up a notebook for a more automatic capture task is a good idea. However for the process of finding good settings kind of an “Oscilloscope view” might be best.

Maybe the most user-friendly option for the “capture” part might be to have a gui that spits out a python script with all of the ADC and clock configurations on request :-), for later use in a more automatic capture process …

On the other hand for the analysis part, I believe that the Jupyter environment makes it more transparent, what is actually happening. I did not run very complex attacks so far, but I did get quite well along with the jupyter notebooks. For the CW4 GUI, I had the impression that there seem to be some “magical scripts being running in the background” or automatic stuff that you need to re-engineer before you understand the logic of what is happening during the actual attack. The fancy features in CW4 where you see the correct AES key in red bubbiing up with increasing number of traces might not be essential. The attack logic was much more clear for me with the v5 environment, also because it is possible to actually describe the attack and give background in the notebook’s text.




Hi Björn,

Thank you for the feedback! Good to see that Jupyter makes what’s happening more transparent, as that was a big reason why we switched in the first place. IIRC, we do have plans to eventually implement a simple GUI for capture, so hopefully that will solve your concerns in that area.