gpudirect

Build userspace NVMe drivers and storage applications with CUDA support

Extended and advanced applications to the Rivermax SDK (Networking SDK for Media and Data Streaming).

Open source version of DOCA GPUNetIO and DOCA Verbs libraries (limited features) to enable GDAKI technology on RDMA (IB and RoCE)

⚡ Guidance, samples, and tools for HPC workloads on AKS clusters with RDMA and InfiniBand support, including GPUDirect RDMA.

GPUDirect Async suite

GPUDirect Async implementation of HPGMG-FV CUDA

a test of GPU direct with CUDA/OpenACC.

can - a simple dense matrix-matrix multiplication benchmark with MPI/OpenMP/OpenACC. MPI version is based on Cannon's algorithm.

Comparing GPU-communication frameworks. World-first multi-GPU maximum cardinality bipartite matching.

GPUDirect Async implementation of CoMD-CUDA

A hybrid testbed for evaluating top open-source LLMs (like gpt-oss-20b and Llama 3.3) on local, cloud GPUs, and AWS Inferentia2/Trainium instances, focusing on vLLM optimization, capacity management, kernel bypass, hardware-software co-design, as well as supporting infrastructure such as NCCL, RDMA, NVMeoF.

NCCue - a lightweight RDMA + GPU inspection toolkit for analyzing NCCL-style communication patterns

Reproducible diagnostics for NCCL, CUDA, RDMA, Kubernetes, SLURM, and distributed GPU incidents

Manage Azure Kubernetes Service with interactive CLI tools for kubectl, logs, exec, console, ConfigMaps, and rollout restarts