ApronPrecCompare runs >= 35h and consumes 40GB of RAM

[michael-schwarz]

/home/goblint/michael-schwarz-dissertation/analyzer/apronPrecCompare /home/goblint/michael-schwarz-dissertation/bench/bench_result/wrk.box.prec /home/goblint/michael-schwarz-dissertation/bench/bench_result/wrk.oct.prec /home/goblint/michael-schwarz-dissertation/bench/bench_result/wrk.cluster12.prec

Does not terminate after more than 35h and gobbling up some 40GB of RAM, even though the individual analyses all terminated within 15min.

[sim642]

How large are the dump files themselves? It could either be that the input files are already overly massive or that the comparison somehow allocates too much something.

[michael-schwarz]

They are ~800 MB each.

[sim642]

That's quite significant, although I don't remember how large they were at the time (maybe I still have some laying around). There isn't much that can be done about it though I guess. We might've also gotten more precise which leads to more constraints in the output.

Is there any output over this comparison time or not? I wonder if there's slowdown due to just printing large states (e.g. in Pretty/Format) or it blows up before that point is reached.

[michael-schwarz]

Is there any output over this comparison time or not?

There's no output at this point.