Go backward to
Reducing Software Engineering Costs
Go up to
Go forward to
The Mapping Stage
Agglomeration Design Checklist
Has agglomeration reduced communication costs by increasing locality?
If agglomeration replicates computation, do benefits outweigh costs?
If agglomeration replicates data, is scalability not compromised?
Have tasks similar computation and communication costs?
Does the task number still scale with problem size?
Is ther still sufficient concurrency for future target computers?
Can task number be further reduced without introducing load imbalances?
How far must the code be modified?
Last modification: November 15, 1996