Why does the stack trace point at the wrong line?

CUDA kernels launch asynchronously. The Python trace shows the next operation that synchronizes with the GPU, not the kernel that actually asserted. Re-run with CUDA_LAUNCH_BLOCKING=1 so the trace points at the faulting kernel.

What does `srcIndex < srcSelectDimSize` mean?

It is the internal bound check in an index/gather kernel. It fails when an index (a label or token id) is greater than or equal to the dimension size, i.e. a label equals num_classes, or a token id exceeds vocab_size. Bound-check your labels and input_ids.

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 a label or index is outside the valid range for an embedding/loss layer (e.g. label == num_classes, or pad id >= vocab_size). Validating against the symptom list in this entry is the fastest way to confirm the diagnosis.

Is this a hardware or a software issue?

Training Stability 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 training stability 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 Classification heads with label index >= num_classes 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-side-assert, #cuda, #index out of bounds. 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-side-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 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.

CUDA Device-Side Assert Triggered

Q: What does `srcIndex < srcSelectDimSize` mean?

It is the internal bound check in an index/gather kernel. It fails when an index (a label or token id) is greater than or equal to the dimension size, i.e. a label equals num_classes, or a token id exceeds vocab_size. Bound-check your labels and input_ids.

Q: 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 a label or index is outside the valid range for an embedding/loss layer (e.g. label == num_classes, or pad id >= vocab_size). Validating against the symptom list in this entry is the fastest way to confirm the diagnosis.

A CUDA kernel hit a device-side assertion. Almost always an out-of-bounds index into an embedding, loss, or gather/scatter op. Because CUDA is asynchronous, the reported stack trace points at an unrelated later line, so the real cause is hidden unless you force synchronous execution.

Quick answer

A CUDA kernel hit a device-side assertion. Almost always an out-of-bounds index into an embedding, loss, or gather/scatter op.

Symptom: RuntimeError: CUDA error: device-side assert triggered
Root cause: A label or index is outside the valid range for an embedding/loss layer (e.g. label == num_classes, or pad id >= vocab_size).
Recommended fix: CUDA_LAUNCH_BLOCKING=1
How Denpex helps: Denpex investigates CUDA Device-Side Assert Triggered 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.

Training Stability#device-side-assert#cuda#index out of bounds#embedding#cross-entropy#training-stability

What this failure is

CUDA Device-Side Assert Triggered is a Training Stability failure seen during ML training runs. A CUDA kernel hit a device-side assertion. Almost always an out-of-bounds index into an embedding, loss, or gather/scatter op. Because CUDA is asynchronous, the reported stack trace points at an unrelated later line, so the real cause is hidden unless you force synchronous execution. Common tags: Device Side Assert, Cuda, Index Out Of Bounds, Embedding.

Live diagnosis, no signup

Is this what broke your run? Paste your log.

You're reading about CUDA Device-Side Assert Triggered. 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)

A label or index is outside the valid range for an embedding/loss layer (e.g. label == num_classes, or pad id >= vocab_size). CUDA executes asynchronously, so the trace points at the next synchronizing call, not the faulting kernel. Target tensor contains -1 or a sentinel id the loss does not ignore. Vocabulary/num_classes drifted from the data after a config change. 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'
The reported line is unrelated to the real fault
Subsequent CUDA calls all fail until the process restarts

Common symptoms and what they mean

Symptom	Why it happens
RuntimeError: CUDA error: device-side assert triggered	A label or index is outside the valid range for an embedding/loss layer (e.g. label == num_classes, or pad id >= vocab_size)
Assertion `srcIndex < srcSelectDimSize` failed	CUDA executes asynchronously, so the trace points at the next synchronizing call, not the faulting kernel
nll_loss / index_select / embedding op in the trace	Target tensor contains -1 or a sentinel id the loss does not ignore
Error disappears on CPU but the loss is wrong (label out of range)	Vocabulary/num_classes drifted from the data after a config change

Which systems are affected

Classification heads with label index >= num_classes
Embedding lookups with token id >= vocab_size
Models after a tokenizer/vocab change
Any gather/scatter with a stale index tensor

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
✓Verified signal present: Assertion `srcIndex < srcSelectDimSize` failed
✓Verified signal present: nll_loss / index_select / embedding op in the trace
✓Verified signal present: Error disappears on CPU but the loss is wrong (label out of range)
✓A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.

Root cause

A label or index is outside the valid range for an embedding/loss layer (e.g. label == num_classes, or pad id >= vocab_size)
CUDA executes asynchronously, so the trace points at the next synchronizing call, not the faulting kernel
Target tensor contains -1 or a sentinel id the loss does not ignore
Vocabulary/num_classes drifted from the data after a config change

The fix and how to prevent it

Searchable error signature

search key

RuntimeError: CUDA error: device-side assert triggered

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