Amdahl's Law¶

The maximum speedup achievable through parallelization is fundamentally limited by the sequential portion of the workload; if 10% must run serially, speedup is capped at 10x regardless of processor count

Tension Addressed¶

Generality vs. Optimization

Recognizable As¶

Diminishing returns from adding resources, serial bottleneck analysis, parallelization ceiling calculations, the "last 10%" problem

Canonical Sources¶

Amdahl, "Validity of the Single Processor Approach to Achieving Large Scale Computing Capabilities" (1967); Gustafson, "Reevaluating Amdahl's Law" (1988)

Transfers To¶

Project management (critical path), manufacturing (bottleneck theory), queuing theory, any resource allocation problem

Sub-concepts: - Gustafson's Law: Counter-perspective that as systems scale, the problem size often scales too, changing the serial/parallel ratio - Critical Path: The sequential dependency chain that determines minimum completion time