REVIEW 5 cited by
FALCON: Pinpointing and Mitigating Stragglers for Large-Scale Hybrid-Parallel Training
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
FALCON: Pinpointing and Mitigating Stragglers for Large-Scale Hybrid-Parallel Training
read the original abstract
Fail-slows, or stragglers, are common but largely unheeded problems in large-scale hybrid-parallel training that spans thousands of GPU servers and runs for weeks to months. Yet, these problems are not well studied, nor can they be quickly detected and effectively mitigated. In this paper, we first present a characterization study on a shared production cluster with over 10,000 GPUs1. We find that fail-slows are caused by various CPU/GPU computation and cross-node networking issues, lasting from tens of seconds to nearly ten hours, and collectively delaying the average job completion time by 1.34%. The current practice is to manually detect these fail-slows and simply treat them as fail-stops using a checkpoint-and-restart failover approach, which are labor-intensive and time-consuming. In this paper, we propose FALCON, a framework that rapidly identifies fail-slowed GPUs and/or communication links, and effectively tackles them with a novel multi-level mitigation mechanism, all without human intervention. We have applied FALCON to detect human-labeled fail-slows in a production cluster with over 99% accuracy. Cluster deployment further demonstrates that FALCON effectively handles manually injected fail-slows, mitigating the training slowdown by 60.1%.
Forward citations
Cited by 5 Pith papers
-
StageFrontier: Synchronization-Aware Stage Accounting for Distributed ML Training
StageFrontier computes an exact additive accounting of exposed step time in distributed training by taking the frontier of per-rank coarse stage durations reported with unsynchronized CPU wall clocks.
-
From Detection to Recovery: Operational Analysis on LLM Pre-training with 504 GPUs
Production logs from a 504-GPU LLM training cluster show 100% failure detection via multi-metric analysis, NFS saturation limiting bandwidth to 1.4-10.4% of link speed, and auto-retry achieving 33.3% success versus 12...
-
From Detection to Recovery: Operational Analysis on LLM Pre-training with 504 GPUs
Production-scale empirical study of a 63-node 504-GPU cluster reports multi-signal failure detection needs, low checkpoint bandwidth utilization, heavy-tailed node exclusions, and 2.7x higher success for auto-retry chains.
-
PRISM: Probabilistic Runtime Insights and Scalable Performance Modeling for Large-Scale Distributed Training
PRISM introduces a probabilistic performance modeling framework that quantifies guarantees on training time for large-scale distributed systems under runtime variability.
-
TSGuard: Automated User-Centric Incident Diagnosis for AI Workloads in the Cloud
TSGuard builds domain knowledge bases offline from historical incidents and applies online multi-agent structured reasoning to diagnose AI workload failures, delivering 19.8% higher accuracy and 63.4% lower verificati...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.