Pinned Loading

1. uccl-project/uccl uccl-project/uccl Public

   UCCL is an efficient communication library for GPUs, covering collectives, P2P (e.g., KV cache transfer, RL weight transfer), and EP (e.g., GPU-driven)

   C++ 1.4k 157

2. uccl-project/mKernel uccl-project/mKernel Public

   mKernel: fast multi-node, multi-GPU fused kernels

   Cuda 231 22

3. NEO-MLSys25/NEO NEO-MLSys25/NEO Public

   NEO is a LLM inference engine built to save the GPU memory crisis by CPU offloading

   Python 97 24

4. eTran-NSDI25/eTran eTran-NSDI25/eTran Public

   eTran: Extensible Kernel Transport with eBPF

   C 51 5

5. DINT-NSDI24/DINT DINT-NSDI24/DINT Public

   DINT: Fast In-Kernel Distributed Transactions with eBPF

   C 52 2

6. Electrode-NSDI23/Electrode Electrode-NSDI23/Electrode Public

   Electrode: Accelerating Distributed Protocols with eBPF

   C++ 53 9