We want to know what performance cost we can expect to pay for tracing WT runs using ReproZip and extracting any additional system-level provenance after runs.
We can measure and compare the time required to execute run.sh directly, vs using reprozip trace, for a wide variety of examples, and separately measure the time required to extract facts from the trace.
This can be implemented as an optional target in the common Makefile.
We want to know what performance cost we can expect to pay for tracing WT runs using ReproZip and extracting any additional system-level provenance after runs.
We can measure and compare the time required to execute
run.shdirectly, vs usingreprozip trace, for a wide variety of examples, and separately measure the time required to extract facts from the trace.This can be implemented as an optional target in the common Makefile.