Qaqao: Implicitly Introducing Flexibility into Parallelism
Computer Sciences | Physical Sciences and Mathematics
J. Andrew Holey, Computer Science
In this paper, a method is suggested which facilitates a move back toward implicit parallel programming using the language of Qaqao. The model, in particular, is designed to be very flexible in order to allow for future developments in parallel language research. In addition, the model seeks to remain flexible, even at execution time, to allow for conservative but effective performance improvements by taking advantage of shared-memory, multi-core systems. This is particularly important in the consumer market, as software is not usually designed with the intention of having full access to all system resources. The initial design of the model does not aim at peak parallel performance, but it is designed to facilitate future improvements easily.
Though an implementation of the model is provided, it is by no means meant to be a simple solution to this radically complex problem, but, rather, another communication on the topic and a starting point for investigation. In addition, this paper implicitly makes the argument that the choice of parallel model is largely impacted by the design language. To begin with, then, it is important to look at object-oriented design as it has been approached thus far.
Ohmann, Peter, "Qaqao: Implicitly Introducing Flexibility into Parallelism" (2010). Honors Theses, 1963-2015. 164.