← Back to Directory
✨
Denpex
Developer Tools
Overview
Denpex delivers automated cluster diagnostics and root-cause analysis for GPU training infrastructure, isolating hardware anomalies, silent data corruption and network bottlenecks across NVIDIA H100 and B200 clusters, supporting PyTorch, vLLM, SGLang and TensorRT-LLM with SLURM/Kubernetes integration.
Key Features
- Automated root-cause analysis from uploaded training/inference logs
- 16,400+ failure signatures with deterministic matching plus AI fallback
- Distributed rank analysis, NCCL timeout and cascade failure detection
- Integrates with Weights & Biases, MLflow, TensorBoard, SLURM and Kubernetes
Best For
ML infrastructure teams running large GPU training fleets who need fast root-cause analysis of job failures.
Pros & Cons
Cons
- Free tier is capped at 3 diagnoses per day.
- Diagnostic accuracy on real production fleets was not independently tested by SEODataPulse.
Advertisement
Pulse Verdict
Editorial review by SEODataPulse · Editorial rating: 4/5 · Updated
This is an editorial assessment, not an average of customer ratings.
“An automated GPU cluster diagnostics platform that identifies root causes of training failures (hardware anomalies, silent data corruption, network bottlenecks) across NVIDIA H100/B200 clusters, free for light use and paid for production fleets.”
Pricing
Free: 3 diagnoses/day, no signup required. Paid: Team $415/month (25,000 GPU-hours), with Scale, Growth and Data Center tiers above that.
Pricing changes often — confirm current plans on the official site.