GPU MIG (Multi-Instance GPU) Mode
GPU MIG mode splits a GPU into multiple instances, which can cause issues with PyTorch and NCCL if not configured correctly.
GPU MIG mode splits a GPU into multiple instances, which can cause issues with PyTorch and NCCL if not configured correctly.
What this failure is
GPU MIG (Multi-Instance GPU) Mode is a Hardware failure seen during ML training runs. GPU MIG mode splits a GPU into multiple instances, which can cause issues with PyTorch and NCCL if not configured correctly. Common tags: Mig, Multi Instance, Gpu, Nvidia.
Is this what broke your run? Paste your log.
You're reading about GPU MIG (Multi-Instance GPU) Mode. 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)
MIG mode splits GPU into multiple instances. PyTorch doesn't natively support MIG. NCCL may not work across MIG instances. MIG instances have limited memory and compute. 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
- PyTorch doesn't see all GPU memory
- NCCL fails to use MIG instances
- GPU shows multiple smaller devices
Common symptoms and what they mean
| Symptom | Why it happens |
|---|---|
| nvidia-smi shows MIG instances | MIG mode splits GPU into multiple instances |
| torch.cuda.device_count() returns wrong number | PyTorch doesn't natively support MIG |
| NCCL initialization fails with MIG error | NCCL may not work across MIG instances |
Which systems are affected
- A100/H100 GPUs with MIG enabled
- Multi-tenant GPU sharing
- GPU partitioning for isolation
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: nvidia-smi shows MIG instances
- ✓Verified signal present: torch.cuda.device_count() returns wrong number
- ✓Verified signal present: NCCL initialization fails with MIG 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
- MIG mode splits GPU into multiple instances
- PyTorch doesn't natively support MIG
- NCCL may not work across MIG instances
- MIG instances have limited memory and compute
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.