RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a

RCCL could not map a peer IPC handle across AMD Infinity Fabric, so the ring could not be formed. The AMD equivalent of an NCCL P2P transport failure, peer access between GPUs is unavailable or blocked. This entry explains how to confirm the cause, apply the fix, and separate it from adjacent non-nvidia failures.

Symptom

RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a

Root cause

RCCL could not map a peer IPC handle across AMD Infinity Fabric, so the ring could not be formed. The AMD equivalent of an NCCL P2P transport failure, peer access between GPUs is unavailable or blocked. The decisive evidence is the first log line that precedes "RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a" and differs from a healthy run.

Recommended fix

NCCL_P2P_DISABLE=1

How Denpex helps

Denpex investigates RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a using the evidence you provide or your connected workload collects. Earlier rank, host or application evidence is needed to distinguish an initiating failure from a downstream report.

The literal signature is "RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a". It is a hardware failure associated with AMD ROCm, RCCL, Intel Gaudi, and HCCL. The line identifies the failing operation or subsystem, while the surrounding evidence decides whether it is the initiating fault or a downstream symptom.

RCCL could not map a peer IPC handle across AMD Infinity Fabric, so the ring could not be formed. The AMD equivalent of an NCCL P2P transport failure, peer access between GPUs is unavailable or blocked. The failure becomes visible at this call site because the operation first requires the missing resource, valid state, healthy peer, or correct result. Earlier log lines and a known-good control carry more causal value than the final wrapper exception.

• The workload stops or loses forward progress after emitting "RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a".
• A retry on the same configuration reproduces the failure because the causal state has not changed.
• The outer framework exception can hide the rank, node, allocation, or dependency that failed first.
• Increasing timeouts or reducing workload size can suppress the symptom without correcting the cause.

• AMD ROCm, RCCL, Intel Gaudi, and HCCL
• production-shaped multi-accelerator workloads
• containerized and bare-metal deployments of the same stack

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.

• ✓Find the first occurrence of "RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a" and preserve at least 100 lines before it.
• ✓Identify which rank, node, device, or process emitted the first related warning.
• ✓IPC across containers or PID namespaces is the usual blocker; a process cannot map another's handle without shared IPC namespace. Test with a two-GPU rccl-tests all-reduce before blaming the framework.
• ✓Repeat the same input after the targeted change and require the signature to disappear.
• ✓Resume from job start only after the control passes.

• RCCL could not map a peer IPC handle across AMD Infinity Fabric, so the ring could not be formed. The AMD equivalent of an NCCL P2P transport failure, peer access between GPUs is unavailable or blocked.
• The decisive evidence is the first log line that precedes "RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a" and differs from a healthy run.
• A nearby failure remains a competing hypothesis until a control separates configuration, capacity, transport, and hardware causes.

RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a

Use this text as a lookup key in logs and upstream issue trackers. It is not presented as a captured customer log. Confirm the cause from your own preceding events, versions, configuration and the cited references.

verify peer access is actually available, rocm-smi --showtopo shows the link type between each GPU pair. Inside a container, confirm /dev/kfd and /dev/dri are both mapped and the process has render-group permission. This changes the condition in the causal diagnosis instead of hiding the outer exception. The repeated control proves ownership before recovery from job start.

# Preserve evidence before restarting
rg -n -i 'error|exception|timeout|failed' <log-file>
nvidia-smi
python -m torch.utils.collect_env

# Find the exact signature in the complete log
rg -n -F -- "RCCL WARN Ring failed IPC handle mapping failed across AMD Infinity Fabric gfx90a" <log-file>

Adapt the snippet to your framework. The same pattern holds for PyTorch Lightning, Hugging Face Trainer, DeepSpeed, Megatron-LM, and vLLM training wrappers. Where the wrapper exposes a config flag (for examplelr_scheduler_type in Trainer), prefer the flag over the imperative API to keep the schedule declarative and reproducible.

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.

NCCL is where a distributed job reports failure, which is not the same as where it failed. The hub lists every common NCCL error next to what it actually indicates, and the environment variables that tell them apart.