Education

1996 – 2002

• PhD Student in Parallel Computation at the Department of Computer Science, Faculty of Mathematics and Computer Science, Babes-Bolyai University of Cluj-Napoca
  • Ph.D. Thesis “Design Models for Parallel Algorithms”
    Scientific Adviser: Prof. Gheorghe Coman

1988 – 1994

• Student at the Department of Computer Science, Faculty of Mathematics and Computer Science, Babes-Bolyai University of Cluj-Napoca
  • Analyst-Programmer Diploma (MSc equivalent)
  • Average Final Grade: 9,92 (1-10 scale)
  • B.Sc. Thesis “Object Oriented Programming. Application: An Interactive Environment for Solving Geometrical Problems. ”
    Scientific Adviser: Prof. Bazil Parv

Ph.D. Thesis: “Design Models for Parallel Algorithms”

The thesis deals with the problem of construction of parallel algorithms using some formal methods that assures the algorithms correctness and their formal derivation. The main idea that governs the thesis is that is easier and better to build a parallel program correct by construction, than prove its correctness after the development process is finished.

Some models are analyzed and used for formal derivation of different parallel algorithms for different numerical problem. These also represent an important contribution of the thesis.

Two paradigms are considered: imperative and functional.

The presented imperative model uses a technique for deriving parallel programs from specifications, which is similar with the technique for deriving sequential programs. Parameterized specifications and programs are used. The distributions may lead the derivation to different programs. In order to increase the generality and the flexibility, a new kind of distributions, named set-distributions, are defined based on set-valued mappings.

A functional model based on Bird-Meertens formalism is also analyzed. The advantages of using functional programming are emphasized.

Some special data structures: PowerList, ParList, and PList are used, with important advantages, for building some parallel numerical algorithms; Fast Fourier Transform is a very eloquent example.

Finally, it is proposed a model that tries to combine the advantages of the technique of deriving from specifications and the advantages of functional models. The model is based on recursion, but it is not a really functional model. Parallel, sequential and alternative compositions are used. The model uses definition variables, which are abstractions of messages.

 

B.Sc. Thesis: “Object Oriented Programming. Application:
An Interactive Environment for Solving Geometrical Problems. “

The application defines a language for specifying geometrical problems, and builds for any geometrical problem the
associate figure. The application is suitable especially for problems that need to compute geometrical place; many figures for the same problems can be overdrawing, in order to emphasize the geometrical place. Also, some interrogations may be made about the properties of the figure: parallelism,
perpendicularity, equality, etc.