Silent Data Corruption: Gradual Parameter Drift
A permanent hardware fault causes monotonically increasing divergence of parameters from a baseline once injection begins, even while loss curves look identical to a healthy run.
A permanent hardware fault causes monotonically increasing divergence of parameters from a baseline once injection begins, even while loss curves look identical to a healthy run.
What this failure is
Silent Data Corruption: Gradual Parameter Drift is a Data Integrity failure seen during ML training runs. A permanent hardware fault causes monotonically increasing divergence of parameters from a baseline once injection begins, even while loss curves look identical to a healthy run. Common tags: Sdc, Parameter Drift, Data Integrity, Permanent Fault.
Is this what broke your run? Paste your log.
You're reading about Silent Data Corruption: Gradual Parameter Drift. Paste your own crash log or traceback below and get the real root cause for YOUR run, not this generic entry. No account, no card. Logs are masked at ingress and never saved to account history.
Want 14 days on the Scale plan?
Request an evaluation code. A verified workplace organization activates up to 50 diagnoses a day, alerts, history, and follow-up questions. No credit card or automatic subscription.
Why it happens (the mechanism)
A permanent fault perturbs computation every step. Errors accumulate as steadily growing parameter divergence. Loss is insensitive enough to mask the drift. 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
- Loss curve looks indistinguishable from baseline
- Yet the model's parameters silently diverge
- Discovered only via cross-run comparison
Common symptoms and what they mean
| Symptom | Why it happens |
|---|---|
| Monotonically increasing parameter L2 distance from a reference run | A permanent fault perturbs computation every step |
| Loss visually identical to baseline | Errors accumulate as steadily growing parameter divergence |
| No NaN, no error | Loss is insensitive enough to mask the drift |
Which systems are affected
- Nodes with a permanent (non-transient) hardware fault
- Long pretraining/fine-tuning runs
- Any precision regime
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: Monotonically increasing parameter L2 distance from a reference run
- ✓Verified signal present: Loss visually identical to baseline
- ✓Verified signal present: No NaN, no error
- ✓A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.
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.
Sign up free. Unlock the full analysisNo 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 extensionRelated failures to investigate next
Root cause
- A permanent fault perturbs computation every step
- Errors accumulate as steadily growing parameter divergence
- Loss is insensitive enough to mask the drift
The fix and how to prevent it
Evaluate Denpex on your own logs
Request a Scale evaluation code. A verified workplace organization activates 14 days with up to 50 diagnoses a day. Every account keeps its current diagnosis allowance and gets a verification path. No card or automatic subscription.
Don't just read the fix, diagnose your run
The encyclopedia tells you what went wrong. Denpex tells you what went wrong in YOUR training run. With your logs, your config, and your stack.