Measure time to achieve a given coverage (not vice versa)

For each new branch to be covered (or anything new to be [dis]covered, really), it takes exponentially more time (write up with experiments to be published shortly). This is why a comparison of the *coverage achieved by each fuzzer within a fixed time budget* gives very small deltas. This relationship might result in potentially spurious differences when ranking fuzzers in terms of their efficiency.

Instead, I would suggest to measure *the time each fuzzer takes to achieve a given coverage*. The *given* coverage could be determined manually or using the coverage achieved in 24h by that fuzzer which achieved the smallest coverage in 24h. This should give larger differences across fuzzers and might facilitate a better assessment of relative fuzzer efficiency.