IBM/chopstix

Add SQL support to state extraction/tracer

rbertran opened this issue · 2 comments

The state should be saved in a database, instead of plain binary files.
Reasoning: This is both cleaner and a more portable approach.
Additionally we can implement versioning? Maybe this should be a separate issue..

Questions:

  • What should be the DB layout?
  • Should the information be saved in the same database as the rest of dynamic information?

We should have an interface to 'write' directly to a file (fast/minimum overhead) or insert into a database for offline analysis and dumping (good during exploration/research when looking for intelligent algorithms offline).

chop.db database is already big enough after sampling. Discard integrating more information into the database for now.