rdmatop
RDMA EFA InfiniBand RoCE NCCL NVSHMEM Monitoring 2026-06-15

rdmatop: Cross-Provider htop for RDMA Traffic
CommBench logo
LLMs Benchmark Code Generation GPU Communication NCCL RDMA CUDA MSCCLPP 2026-06-09

CommBench: Can LLMs Write Correct and Efficient GPU Communication Code?
OpenAI MRC
MRC SRv6 SRD Multi-plane UCCL-Tran 2026-05-26

Reading OpenAI's MRC Through a UCCL Lens
mKernel
Fused Kernels RDMA 2026-05-25

mKernel: Fast Multi-GPU, Multi-Node Fused Kernels
TrainMover Overview — Two-phase machine migration design
LLM Training Fault Tolerance Systems OSDI 2026-05-18

When Your AI Training Cluster Crashes at 3 AM: How TrainMover Cuts Recovery Time to 20 Seconds
UCCL-EP
RDMA EFA 2026-04-13

A Practitioner Guide to AWS EFA Programming
UCCL-EP
MoE DeepEP RDMA Expert Parallelism AMD EFA 2026-04-06

UCCL-EP: Portable Expert-Parallel Communication — Full Results
UCCL-EP
MoE DeepEP IBGDA RDMA 2025-10-27

Previewing UCCL-EP: Flexible and Efficient Expert Parallelism for Cloud and Beyond
KV transfer engine
NIXL NCCL RCCL Mooncake RDMA 2025-08-13

Everything You Want to Know about KV Cache Transfer Engine
About
NCCL RCCL RDMA 2025-06-30

How to Debug NCCL Performance Issues for ML Workloads?
About
Networking AI RDMA 2025-05-26

UCCL-Tran: An Extensible Software Transport Layer for GPU Networking
About
Networking AI Sky Computing 2025-05-26

About Us