Next: Simplest System
Up: Theory and Examples
Previous: Theory and Examples
In the following sections, very simple noemic algorithms are introduced and compared to their more traditional counterparts.
The comparisons are based on:
- Calculation: The cost of the computation.
- Latency: The time to receive the first results.
- Throughput: The cost of the computation once the system is primed.
- Resource: The number of elements.
- One Fault Tolerance: If one component fails in the system, the entire system fails.
- Fault Tolerance Threshold: The number of components that have to fail to make the total system fail.
- Elaboration Tolerance: The cost of a small change in the computation.
Ronald LeRoi Burback
Wed Jul 30 15:24:07 PDT 1997