OMP: Error #15: Initializing libomp.so but found libgomp.so already initialized

Two OpenMP runtimes were loaded into one process (Intel/LLVM libomp alongside GNU libgomp). Behaviour is undefined, in practice thread pools collide, causing deadlocks, oversubscription, or crashes in multi-threaded inference. This entry explains how to confirm the cause, apply the fix, and separate it from adjacent quant failures.

The literal signature is "OMP: Error #15: Initializing libomp.so but found libgomp.so already initialized". It is a environment failure associated with latency-sensitive CUDA, Triton, and OpenMP runtimes. The line identifies the failing operation or subsystem, while the surrounding evidence decides whether it is the initiating fault or a downstream symptom.

Two OpenMP runtimes were loaded into one process (Intel/LLVM libomp alongside GNU libgomp). Behaviour is undefined, in practice thread pools collide, causing deadlocks, oversubscription, or crashes in multi-threaded inference. 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 "OMP: Error #15: Initializing libomp.so but found libgomp.so already initialized".
• 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.

• latency-sensitive CUDA, Triton, and OpenMP runtimes
• production-shaped multi-accelerator workloads
• containerized and bare-metal deployments of the same stack

Apply the following checklist to a small reproduction: each box below is a positive signal that you are looking at this exact failure rather than a sibling in the same taxonomy.

• ✓Find the first occurrence of "OMP: Error #15: Initializing libomp.so but found libgomp.so already initialized" and preserve at least 100 lines before it.
• ✓Identify which rank, node, device, or process emitted the first related warning.
• ✓the common pairing is a conda numpy/MKL alongside a pip-installed torch or a compiled extension. Standardise on one channel for everything that links OpenMP.
• ✓Repeat the same input after the targeted change and require the signature to disappear.
• ✓Resume from process restart with a consistent runtime only after the control passes.

OMP: Error #15: Initializing libomp.so but found libgomp.so already initialized

Timestamps and exact values vary across runs, but the pattern. An info-level start, an early WARN, an ERROR carrying the symptom. Is the actual fingerprint you should alert on. The Denpex platform flags this combination automatically.

find which libraries pull in each runtime, ldd on the offending .so files, or set LD_DEBUG=libs to trace load order. MKL/Intel builds pull libomp; many PyPI wheels built with GCC pull libgomp. This changes the condition in the causal diagnosis instead of hiding the outer exception. The repeated control proves ownership before recovery from process restart with a consistent runtime.

# 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 -- "OMP: Error #15: Initializing libomp.so but found libgomp.so already initialized" <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.

• Two OpenMP runtimes were loaded into one process (Intel/LLVM libomp alongside GNU libgomp). Behaviour is undefined, in practice thread pools collide, causing deadlocks, oversubscription, or crashes in multi-threaded inference.
• The decisive evidence is the first log line that precedes "OMP: Error #15: Initializing libomp.so but found libgomp.so already initialized" and differs from a healthy run.
• A nearby failure remains a competing hypothesis until a control separates configuration, capacity, transport, and hardware causes.