AI RESEARCH
Modernizing Amdahl's Law: How AI Scaling Laws Shape Computer Architecture
arXiv CS.AI
•
ArXi:2603.20654v1 Announce Type: cross Classical Amdahl's Law assumes a fixed decomposition between serial and parallel work and homogeneous replication; historically, it bounds how much parallel speedup is attainable. Modern systems instead combine specialized accelerators with programmable compute, tensor datapaths, and evolving pipelines, while empirical scaling laws shift which stages absorb marginal compute. The central tension is. therefore.