perf issues with entropy calculation
Open
- Dominant language
- C
- Stars
- 3.4k
- Forks
- 539
- PR merge metrics
- No merged PRs in 30d
Description
when time/exec is small (to the order of 10s of microseconds), targets end up underutilizing their CPU cores because honggfuzz is spending 85% of its time calculating entropy, we should skip-sample bytes and/or cache the entropies for corpus items
Contributor guide
Assessment
This issue has not been assessed yet.