GPU TDP / Power Limit

GPU TDP and power limit configuration affects performance, thermals, and energy efficiency of training workloads.

GPU TDP / Power Limit is a Infrastructure failure seen during ML training runs. GPU TDP and power limit configuration affects performance, thermals, and energy efficiency of training workloads. Common tags: Tdp, Power Limit, Gpu Clocks, Power.

Power limit set too low for workload. Default power limit too conservative. GPU throttling due to power not thermal. Inefficient GPU 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.

• GPU performance is throttled
• Training is slower than expected
• Power consumption is higher than expected

• Long-running training jobs
• Power-constrained GPU servers
• Datacenter with power caps

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 power limit reached
• ✓Verified signal present: GPU clocks lower than base
• ✓Verified signal present: Power consumption at 100% of limit
• ✓A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.

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.

• Power limit set too low for workload
• Default power limit too conservative
• GPU throttling due to power not thermal
• Inefficient GPU utilization