EXP06: Big (use all files)

Practical-Formal-Methods / adiff

Tool for differentially testing soundness and precision of program analyzers

MIT License

11 stars 6 forks source link

EXP06: Big (use all files) #85

Closed chkl closed 5 years ago

chkl commented 6 years ago

all files,

budget="min 100 (0.2*reads)"
timeout=20
strategy=random-uniform
batch-size=4

on all 2550 files.

chkl commented 6 years ago

Restarted the experiment today after adding +RTS -M500M, so at least it will continue.

chkl commented 6 years ago

restarting the experiment. Before that I cleared all results of smack and all results of klee where the result was != Sat. (That's because the old experiment ran with too small loop-bounds for smack and without some necessary definitions in klee.h)

But the actual reason for the restart is of course crab-llvm.

chkl commented 6 years ago

Fixing the typechecker has led to a semi-unexpected outcome: It's taking much longer now. So of the 2550 relevant files, it has so far processed 1322 files while having skipped 444 files already. Which means there are in the worst-case 789 files left. In other words, it might only be at 62% completion, meaning it probably takes about 3-4 more days. (fwiw: the database is 850MiB big)