ETTR · Effective Training Time Ratio
The share of wall-clock time a training job spends making forward progress rather than restarting, checkpointing, or waiting on failures; the honest utilization metric for training clusters.
Current numbers
≈10 s checkpoint writes, or failure rate ≈1 (vs 6.50) per 1,000 node-daysMeta's revised RSC-1 model: what a roughly 12,000-GPU run needs for ETTR of about 0.9