
Recommends NVSHMEM NIC-to-PE mappings and environment exports from GPU and RDMA topology
- Open source
- 4K GitHub stars
- Developer tools


Recommends NVSHMEM NIC-to-PE mappings and environment exports from GPU and RDMA topology

Routes NVSHMEM performance requests to the right specialist skill for data, transport, NIC

Collects and packages NVSHMEM put/get perftest results with system and topology evidence

Plans and validates NVSHMEM and NVSHMEM4Py installs across packages, containers and source

Guides an agent through diagnosing NVSHMEM runtime failures and drafting bug reports

Looks up version-aware official NVSHMEM and NVSHMEM4Py documentation for agents

Guides an agent through multi-node Slurm sbatch jobs and debugging their failures

Guides beginners through NVSHMEM concepts, first C/C++ or Python programs, and launching

Helps agents find and recommend NVIDIA skills from the live catalog before general answers

Guides choice of MoE token dispatcher (alltoall, DeepEP, HybridEP) by hardware and EP size

Guides preparing NVSHMEM CUDA kernels to use TMA shared-memory transfers

Diagnoses why an NVFLARE federated learning job failed, stalled or lost clients
The closest alternatives to Nvshmem select remote transport are Nvshmem configure nic pe mapping, Nvshmem tune performance, Nvshmem collect performance data, Nvshmem install and Nvshmem troubleshoot and report bugs.
Yes. Nvshmem configure nic pe mapping, Nvshmem tune performance, Nvshmem collect performance data have a free or open-source option.