What does this failure look like in production?

In production this failure typically appears during the first tens-to-hundreds of training steps: loss curves become non-monotonic, validation metrics plateau above the baseline, and downstream checkpoints either diverge or fail to improve. Engineers frequently misread it as data drift when the underlying mechanism is one rank stopped progressing (data stall, slow i/o, faulted gpu). Validating against the symptom list in this entry is the fastest way to confirm the diagnosis.

Is this a hardware or a software issue?

Distributed Training failures can look like either layer depending on the symptom. The discriminator is the resolution of the symptom to the recommended fix: if applying the steps under "How to fix it" eliminates the symptom, this is a software/configuration class issue regardless of the hardware it surfaced on. Repeat the fix on a fresh node to confirm.

How long does the fix typically take?

Most resolutions from this entry take 10 to 30 minutes of engineer time once the diagnosis is confirmed. Long fixes come from skipping the diagnosis step and chasing apparent symptoms (validation drift, memory growth, slow throughput) instead of the actual root cause listed above. Apply the fix in a small experiment first, then roll it forward.

Does this affect inference or only training?

Primarily this is a distributed training failure observed during training, but the same misconfiguration can leak into downstream evaluation and serving. Treat the prevention checklist as a guardrail: any pipeline that touches a trained checkpoint should re-validate the failure mode before promoting the model.

What's the cheapest way to confirm this is the problem?

Cheapest signal: reproduce the symptom on a 1-GPU toy run with the same configuration. If the toy run exhibits the same behavior, the cause is the configuration rather than the cluster. If only the multi-GPU run is affected, the failure is a distributed interaction, so Large model+data parallel jobs with many comm groups is the most likely layer to inspect.

Can the Denpex platform detect this automatically?

Yes. Denpex correlates the symptoms in this entry against your training logs in real time and surfaces the failure taxonomy, root cause, and exact remediation command on the console. Three full diagnoses are included free during signup, and the encyclopedia entry you are reading is generated from the same signal library that powers the platform.

What related failures should I check next?

Closely related failures cluster around #nccl, #hang, #straggler. After applying this fix, validate against the related entries below to confirm no overlapping cause is masking as this one. Engineers routinely discover that the dominant failure mode is actually a second, sibling failure compounded on top.

Will this recur after the fix?

Applied once with the prevention checklist, this failure does not recur under normal operation. The recurring variant most engineers encounter is the re-emergence of the same symptom on a different model or stack after nccl is upgraded. Pin compatible versions and rerun the prevention smoke test on every dependency change.

How does this scale from 1 GPU to 1000 GPUs?

At single-GPU scale the symptom usually looks like slow convergence or a single anomalous loss value. At 64+ GPU scale, distributed interactions amplify one rank stopped progressing (data stall, slow i/o, faulted gpu) through collectives, and the failure can manifest as a full job abort. Always reproduce the fix at the smallest cluster configuration that still exhibits the symptom before promoting to production scale.

Is there a fingerprint I can monitor?

Yes. GPU utilization 100% but power draw low (~70W) on the stuck rank is the highest-signal fingerprint. Wire a metric or alert on that signal (or its absence after the fix) so you can detect the recurrence automatically. The Denpex platform encodes this exact fingerprint and surfaces it directly in the dashboard.

What is the official documentation that covers this?

Reference documentation is listed in the References section at the bottom of this entry. The most authoritative URLs are the upstream framework docs and the issue trackers cited; the Denpex entry synthesizes both into the executive, engineer, and on-call perspectives so each audience can act without reading every linked source.

Does this entry include working code?

Yes. The "Code Examples" section contains copy-paste-ready snippets that target the most common stacks (PyTorch, Hugging Face Trainer, DeepSpeed, Megatron-LM, vLLM/Transformers). Each snippet is annotated with why it works so you can adapt the pattern to your own framework without losing fidelity.

Finding the Culprit Rank in an NCCL Hang

When one rank stalls, NCCL's synchronous collectives freeze every rank, masking which GPU actually failed. Localizing the culprit needs divergence signals, not the NCCL stack trace.

Quick answer

When one rank stalls, NCCL's synchronous collectives freeze every rank, masking which GPU actually failed. Localizing the culprit needs divergence signals, not the NCCL stack trace.

Symptom: No NCCL error emitted, watchdog eventually times out
Root cause: One rank stopped progressing (data stall, slow I/O, faulted GPU). Synchronous collective semantics block all peers waiting on it. The faulted rank is in a different state than the rest.
Recommended fix: NCCL_DEBUG=INFO
How Denpex helps: Denpex investigates Finding the Culprit Rank in an NCCL Hang using the evidence you provide or your connected workload collects. Earlier rank, host or application evidence is needed to distinguish an initiating failure from a downstream report.

Distributed Training#nccl#hang#straggler#flight-recorder#debugging#distributed

What this failure is

Finding the Culprit Rank in an NCCL Hang is a Distributed Training failure seen during ML training runs. When one rank stalls, NCCL's synchronous collectives freeze every rank, masking which GPU actually failed. Localizing the culprit needs divergence signals, not the NCCL stack trace. Common tags: Nccl, Hang, Straggler, Flight Recorder.

Live diagnosis, no signup

Is this what broke your run? Paste your log.

You're reading about Finding the Culprit Rank in an NCCL Hang. Paste your own traceback and relevant evidence for an investigation of your workload, with a next action or a specific missing fact. A reference entry does not establish your cause. No account or card for the free diagnosis. Review data handling before submitting sensitive logs.

Before uploading, review cloud data handling and local options.

training_logs.txt

No log to hand? Try one:

Upload or drop a plain-text log

Upload your log file. Large files are processed in your browser. Only selected evidence is submitted when you run diagnosis.

3 free diagnoses/day

Why it happens (the mechanism)

One rank stopped progressing (data stall, slow I/O, faulted GPU). Synchronous collective semantics block all peers waiting on it. The faulted rank is in a different state than the rest. Taken together, these mechanisms explain why the failure is reproducible, why it tends to surface on specific workloads or scales, and why generic mitigation attempts often fall short without addressing the underlying cause.

What you'll observe

The whole job hangs with no error
Every rank shows the same NCCL wait, so the stack is unhelpful
Need to find which specific GPU/rank diverged

Common symptoms and what they mean

Symptom	Why it happens
GPU utilization 100% but power draw low (~70W) on the stuck rank	One rank stopped progressing (data stall, slow I/O, faulted GPU)
All ranks blocked in the same collective (AllReduce/ReduceScatter)	Synchronous collective semantics block all peers waiting on it
No NCCL error emitted, watchdog eventually times out	The faulted rank is in a different state than the rest

Which systems are affected

Large model+data parallel jobs with many comm groups
Long multi-node runs
Mixed-speed or partially faulted hardware

How to confirm this is the problem

Use this checklist to test the hypothesis against a small reproduction. No single line proves the root cause, so preserve the preceding events and compare one variable at a time.

✓Reproduce the failure from a clean checkpoint/seed: the symptom must appear without warm-up state from a previous run.
✓Verified signal present: GPU utilization 100% but power draw low (~70W) on the stuck rank
✓Verified signal present: All ranks blocked in the same collective (AllReduce/ReduceScatter)
✓Verified signal present: No NCCL error emitted, watchdog eventually times out
✓A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.

Root cause

One rank stopped progressing (data stall, slow I/O, faulted GPU)
Synchronous collective semantics block all peers waiting on it
The faulted rank is in a different state than the rest

The fix and how to prevent it

Searchable error signature

search key

No NCCL error emitted, watchdog eventually times out

Use this text as a lookup key in logs and upstream issue trackers. It is not presented as a captured customer log. Confirm the cause from your own preceding events, versions, configuration and the cited references.

The fix and the prevention pattern

The root cause is on this page and stays free. A free account adds the exact remediation steps, saved history, and the fix on every entry in the encyclopedia.

No credit card. Daily allowance follows verified trust tier. Instant access.

Diagnose this failure in VS Code

Select the traceback or open the failed terminal, then run Denpex locally to see the initiating rank, collateral failures, exact fix, and verification command without uploading the log.

Install the free VS Code extension

NCCL errors in context

NCCL is where a distributed job reports failure, which is not the same as where it failed. The hub lists every common NCCL error next to what it actually indicates, and the environment variables that tell them apart.

Compare every nccl error side by side