For singapore and apac distributed training and inference engineers
Isolate InfiniBand NDR rail faults and multi-node NCCL stalls in Singapore.
Correlate mlx5 HCA port counters, kernel Xid events, and rank sequence logs before draining healthy H100/H200 nodes on immersion-cooled fabrics.
Free 30-day Team evaluation. No card or automatic subscription. Work email verification required.
Replay a Singapore H100 InfiniBand training stall
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.
Upload your log file. Large files are processed in your browser. Only selected evidence is submitted when you run diagnosis.
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
Preserve clean node and port evidence for Singapore datacenter escalation instead of restarting entire multi-node runs.
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.
Follow the failure through the cascade

Prepare the hardware escalation

Ask for the evidence that is missing

Keep the investigation in the editor

SMC (Sustainable Metal Cloud Singapore) - Observed Runtime Evidence
Incident pattern summary
An 8-node H100 SXM5 training run in Singapore stalls on WorkNCCL AllReduce after an optical transceiver drop on one InfiniBand leaf port. Denpex separates the source HCA port from surviving ranks.
smc-sg-h100-04:1980:2044 [2] misc/ibvwrap.cc:284 NCCL WARN Call to ibv_poll_cq returned error 12 (Transport retry counter exceeded)
smc-sg-h100-04:1980:2044 [2] NCCL WARN NET/IB : mlx5_2 port 1 link down event
[E ProcessGroupNCCL.cpp:915] [Rank 18] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=412, OpType=ALLREDUCE)Recommended verification flags
Set these environment flags in your training container or launch script to capture high-resolution collective logs:
NCCL_DEBUG=INFONCCL_IB_HCA=mlx5TORCH_NCCL_ASYNC_ERROR_HANDLING=1CUDA_DEVICE_ORDER=PCI_BUS_IDDiagnostic Scope Boundary
Denpex analyzes exported logs, kernel outputs, and topology captures. It does not touch your private Singapore VPC or InfiniBand subnet manager.
Other GPU Cloud Failure Workflows
Compare evidence patterns across multi-node fabrics, Kubernetes pods, and bare-metal environments.