Conclusions

The following questions should be answered:

1. Which computer projects are in progress?

2. Are the proposed models and methodologies sufficient to develop efficient algorithms?

3. Is there a potential in developing completely new algorithms or is it sufficient to provide methodologies to migrate standard algorithms (bsp-algorithms) to clusters of multiprocessors?

4. Is it necessary to design a new programming paradigm or are the common paradigms sufficient (perhaps with certain optimizations)?

5. Where are the problems of mixed-mode programming?

6. Will there be enough algorithms and libraries to use the clusters efficiently in the future? Which projects are in progress.

7. Where are lacks of suitable solutions?