DAQV is a project attempting to address the general problem of providing high-level semantic-based support for query and visualization of parallel distributed data structures.
High-level access to parallel distributed data objects is not commonly supported by parallel debuggers. This is due, in part, to the fact that debuggers are not always aware of how parallel data structures are distributed.
There are similar needs for distributed data structure access, query, and visualization found in performance evaluation, for instance, to enable capabilities such as computational steering. Computational steering allows scientists to guide intensive, long-running programs toward optimal solutions or higher performance by altering program variables interactively.
A tool that leverages parallel language compilers and runtime systems can provide these capabilities in a component form that can be applied in parallel debuggers, performance evaluation tools, and computational steering environments.