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 single-bit or double-bit ecc error in gpu memory corrupts tensor data during a kernel operation, causing an assertion in downstream operations (e.g., index out of bounds in embedding lookup). Validating against the symptom list in this entry is the fastest way to confirm the diagnosis.

Is this a hardware or a software issue?

Hardware 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 hardware 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 NVIDIA GPUs with hardware faults (ECC errors, NVLink issues, memory channel failures) 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 #device-assert, #cuda, #hardware. 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 device-assert is upgraded. Pin compatible versions and rerun the prevention smoke test on every dependency change.

Is there a fingerprint I can monitor?

Yes. RuntimeError: CUDA error: device-side assert triggered at /pytorch/aten/src/ATen/native/cuda/.. 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.

CUDA Device Assertion Failure

Q: 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 single-bit or double-bit ecc error in gpu memory corrupts tensor data during a kernel operation, causing an assertion in downstream operations (e.g., index out of bounds in embedding lookup) 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.

CUDA device assertion failures indicate GPU hardware problems that crash training with ambiguous error messages. Denpex maps the device assertion to the specific GPU and operation, distinguishing hardware faults from driver issues.

Quick answer

CUDA device assertion failures indicate GPU hardware problems that crash training with ambiguous error messages.

Symptom: RuntimeError: CUDA error: device-side assert triggered at /pytorch/aten/src/ATen/native/cuda/..
Root cause: Single-bit or double-bit ECC error in GPU memory corrupts tensor data during a kernel operation, causing an assertion in downstream operations (e.g., index out of bounds in embedding lookup).
Recommended fix: Immediately run nvidia-smi -q -d MEMORY and nvidia-smi -q -d ECC on the failing GPU to check for hardware errors
How Denpex helps: Denpex investigates CUDA Device Assertion Failure 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.

Hardware#device-assert#cuda#hardware#gpu#ecc#xid

What this failure is

CUDA Device Assertion Failure is a Hardware failure seen during ML training runs. CUDA device assertion failures indicate GPU hardware problems that crash training with ambiguous error messages. Denpex maps the device assertion to the specific GPU and operation, distinguishing hardware faults from driver issues. Common tags: Device Assert, Cuda, Hardware, Gpu.

Live diagnosis, no signup

Is this what broke your run? Paste your log.

You're reading about CUDA Device Assertion Failure. 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)

Single-bit or double-bit ECC error in GPU memory corrupts tensor data during a kernel operation, causing an assertion in downstream operations (e.g., index out of bounds in embedding lookup). NVLink transaction failure corrupts data transferred between GPUs in the same node, detected by a device-side bounds check. GPU memory channel degradation causes intermittent bit flips that only manifest under specific memory access patterns. GPU clock or voltage instability from aging hardware causes computation errors in CUDA cores during sustained high utilization. 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

Training crashes with 'CUDA error: device-side assert triggered' and no further context
The error is non-reproducible: restarting on different GPUs may succeed
The assertion provides no information about which kernel or tensor triggered it

Common symptoms and what they mean

Symptom	Why it happens
RuntimeError: CUDA error: device-side assert triggered at /pytorch/aten/src/ATen/native/cuda/..	Single-bit or double-bit ECC error in GPU memory corrupts tensor data during a kernel operation, causing an assertion in downstream operations (e.g., index out of bounds in embedding lookup)
The same training config crashes on some GPUs but not others in the same cluster	NVLink transaction failure corrupts data transferred between GPUs in the same node, detected by a device-side bounds check
nvidia-smi shows Xid errors (14, 48, 64) associated with the failing GPU	GPU memory channel degradation causes intermittent bit flips that only manifest under specific memory access patterns
CUDA kernel launches succeed but produce garbage output before triggering the assertion	GPU clock or voltage instability from aging hardware causes computation errors in CUDA cores during sustained high utilization

Which systems are affected

NVIDIA GPUs with hardware faults (ECC errors, NVLink issues, memory channel failures)
Training with CUDA graphs that reuse cached kernel launches across faulty hardware
Long-running training jobs that accumulate hardware errors over days or weeks
High-utilization clusters where GPUs are not regularly stress-tested

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: RuntimeError: CUDA error: device-side assert triggered at /pytorch/aten/src/ATen/native/cuda/.
✓Verified signal present: The same training config crashes on some GPUs but not others in the same cluster
✓Verified signal present: nvidia-smi shows Xid errors (14, 48, 64) associated with the failing GPU
✓Verified signal present: CUDA kernel launches succeed but produce garbage output before triggering the assertion
✓A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.

Root cause

Single-bit or double-bit ECC error in GPU memory corrupts tensor data during a kernel operation, causing an assertion in downstream operations (e.g., index out of bounds in embedding lookup)
NVLink transaction failure corrupts data transferred between GPUs in the same node, detected by a device-side bounds check
GPU memory channel degradation causes intermittent bit flips that only manifest under specific memory access patterns
GPU clock or voltage instability from aging hardware causes computation errors in CUDA cores during sustained high utilization

The fix and how to prevent it

Searchable error signature

search key

RuntimeError: CUDA error: device-side assert triggered at /pytorch/aten/src/ATen/native/cuda/..
nvidia-smi shows Xid errors (14, 48, 64) associated with the failing GPU

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

CUDA errors in context

CUDA reports errors asynchronously, so the traceback usually points at whatever line synchronised next rather than the one at fault. The hub covers every common CUDA error and how to make it report honestly.

Compare every cuda error side by side