exiv2 - redqueen behavior

[CStriker]

Hello,

I'm trying to reproduce the best fuzzer choice for the 'exiv2' binary as described in the autofz paper.

According to the paper, redqueen should produce coverage of around 8000 within one round after fuzzing starts.

I manually overrode the cpu_assign dictionary in the one_round() function and tried two scenarios: running the preparation phase and not running the preparation phase. In both cases, redqueen appears to be almost ineffective for fuzzing 'exiv2' in the early stages.

Is there something else I need to modify to override the fuzzer choice for the focus phase? Or is there an additional condition required to make redqueen work effectively?

Thank you.