For distributed training and cloud infrastructure engineers

Separate provider infrastructure faults from distributed training bugs.

Each GPU cloud exposes different networking fabrics, container overlays, and virtualization boundaries. Preserve the exact log evidence needed before retrying a failed run or escalating a support ticket.

Replay one resolved incident Try 30 days free

Free 30-day Team evaluation. No card or automatic subscription. Work email verification required.

Live diagnosis, no signup

Replay a resolved multi-node cluster incident

Upload your log file. Review the selected evidence and masking, then run diagnosis. Keep the original log for any follow-up questions.

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

Try Denpex for 30 days. FREE.

No credit card. No automatic subscription. See whether Denpex helps with an incident you already understand.

Start your free 30-day Team evaluation

Up to 50 diagnoses a day for 30 days. Verify your work email to activate an eligible company evaluation, then replay a resolved incident and compare the evidence and next steps with your team's confirmed cause.

What a useful evaluation should establish

Isolate the single failing node or interface without tearing down healthy cluster resources or blaming collateral timeout victims.

Use Denpex alongside your existing telemetry tools to investigate submitted incident evidence. With monitoring configured on a supported plan, detected failures can trigger alerts and an automatic diagnosis from the captured evidence.

Inside Denpex

See the evidence, not just the claim

Follow the causal chain, inspect the quoted evidence and prepare the next handoff. These actual product captures use founder-account examples and the VS Code extension, with the source identified under each image. Open a capture to inspect it at full size.

Denpex founder-account example: a causal graph connects an Xid 48 ECC event to a rank exit and downstream collective failures, with quoted evidence for each stage. — Live founder-account capture from the bundled NCCL/ECC example. Follow the hardware event, rank exit and downstream symptoms with a supporting log quote at each stage.Inspect original image

Denpex hardware handoff summary asks the operator to capture diagnostic evidence and verify RMA eligibility with the vendor. — Actual hardware handoff summary from the same founder-account example. Capture diagnostic evidence before an approved reset, then confirm RMA eligibility with the vendor.Inspect original image

Denpex historical example says a collective timeout does not establish the initiating rank, cause or failure domain and requests distinguishing evidence before a remedy. — Founder-account example restored from history. The original timeout remains unattributed, and the UI keeps the missing-evidence assessment visible.Inspect original image

Denpex VS Code extension displaying a bundled OOM-before-NCCL example, matched log evidence, verification guidance and an investigation handoff. — Actual VS Code extension capture of the bundled OOM-before-NCCL example. Review matched log evidence, verification guidance and the investigation handoff without leaving the editor.Inspect original image

Available Cloud Provider Failure Triage Workflows

Select your GPU cloud provider to review failure signatures, terminal tracebacks, and verification environment flags.

RunPoden

Separate container socket misrouting from a true worker stall on RunPod.

Keep container network interface bindings, Pod or Instant Cluster lifecycle logs, and rank-0 rendezvous traces beside the first ProcessGroupNCCL watchdog exception.

NCCL_DEBUG=INFONCCL_SOCKET_IFNAME=eth0

Lambdaen

Distinguish an InfiniBand mlx5 fabric flap from an application rank stall on Lambda.

Correlate Mellanox RDMA transport retry counters, NVSwitch health telemetry, and rank-level Python tracebacks to isolate which 8x H100 node triggered the collective abort.

NCCL_DEBUG=INFONCCL_DEBUG_SUBSYS=INIT,NET

Vast.aien

Isolate motherboard PCIe P2P restrictions and /dev/shm limits on Vast.ai hosts.

Verify IOMMU peer-to-peer memory access, Docker shared memory allocation, and spot instance preemption markers before blaming model code on heterogeneous Vast.ai rigs.

NCCL_P2P_DISABLE=1NCCL_SHM_DISABLE=0

Scalewayen

Separate RoCE v2 GID routing mismatches from DeepSpeed ZeRO-3 stage stalls on Scaleway.

Preserve NCCL RoCE Global Identifier (GID) table selections, Priority Flow Control (PFC) indicators, and DeepSpeed parameter partitioning logs when multi-node jobs freeze.

NCCL_DEBUG=INFONCCL_IB_GID_INDEX=3

CoreWeaveen

Link Kubernetes pod restarts directly to underlying Xid 79 or NVLink hardware faults.

Align Kubelet eviction events, DCGM health exporter alerts, and kernel Xid codes so your team knows whether to reschedule onto a healthy node or debug container OOMs.

NCCL_DEBUG=INFOTORCH_NCCL_ASYNC_ERROR_HANDLING=1

Crusoe Clouden

Pinpoint which straggler rank or RoCE link slowed down your Crusoe Cloud cluster.

Compare per-rank step completion drift, shared storage checkpoint durations, and RDMA pause frames before a slow worker trips the global NCCL watchdog.

NCCL_DEBUG=INFOTORCH_NCCL_ENABLE_MONITORING=1

Hetzneren