testsmt / yinyang

A fuzzing framework for SMT solvers
https://testsmt.github.io/
MIT License
185 stars 23 forks source link

High memory usage #47

Open mpreiner opened 2 years ago

mpreiner commented 2 years ago

yinyang (version 0.3.0 installed via pip) consumes ~40G of memory when running with the provided semantic-fusion-seeds/QF_SLIA seeds. Using more seeds will increase memory consumption (with the QF_SLIA logic from SMT-LIB yinyang runs out of memory on my machine with 128G memory). The memory consumption increases rapidly on startup and seems to be stable after yinyang actually starts fuzzing. The same behavior can be observed with the current git version on master.

The command I use to run yinyang is: yinyang ./solver.sh semantic-fusion-seeds/QF_SLIA/