Find a new advisor. Having a body of dishonest work is problematic. I am a Full Professor. Before we submit a paper, I do my best to make sure our code does

Find a new advisor. Having a body of dishonest work is problematic. I am a Full Professor. Before we submit a paper, I do my best to make sure our code does what we say it does. We make sure we understand the results as a second check on our implementation. Then we hope we caught all the important bugs that might invalidate our results. Now to be clear, we are not building industrial grade systems and so new benchmarks might expose bugs we don't know about. I am pretty sure that we are not the only group that does this in the field of Programming Languages & Compilers.