htop, but for RDMA traffic — a real-time TUI monitor for RDMA network interfaces.
Monitors per-device throughput (Gbps, packets/s, drops), RDMA read/write counters, retransmits, health events, and shows which processes are using each RDMA device — all via RDMA netlink, the same interface used by rdma statistic.
- Linux (netlink-based — macOS/Windows are not supported)
- RDMA-capable NICs (e.g., Mellanox/NVIDIA ConnectX, AWS EFA)
On Ubuntu 22.04 (jammy), 24.04 (noble), or 26.04 (resolute) — amd64 and arm64:
sudo add-apt-repository ppa:crazyguitar/rdmatop
sudo apt update
sudo apt install rdmatopcargo install rdmatopmake # cargo build
make install # cargo installrdmatopPress r in the TUI to start recording and r again to stop. rdmatop captures
every device's tx/rx Gbps and packets/s per interval and writes a Chrome-JSON
trace (rdmatop-<n>.json) you can drag into ui.perfetto.dev
— each device/port becomes its own set of counter tracks. Timestamps are relative
to when you pressed r, so the trace spans exactly your record window.
Use rdmatop to monitor RDMA traffic while running GPU
communication benchmarks:
- PyTorch — intranode NVLink/XGMI traffic
- IB Perftest — two-node
ib_write_bwbenchmark - UCX Perftest — two-node
ucx_perftestbandwidth / latency - NCCL — collective communication
- NIXL — point-to-point KV cache transfer
- NVSHMEM — one-sided GPU communication
- PPLX Kernels — MoE all-to-all dispatch/combine
- UCCL — DeepEP-compatible expert-parallel dispatch/combine
- RDMA Statistics — shell-based RDMA stats
- Kubernetes — DaemonSet deployment for Kubernetes
- Device enumeration —
RDMA_NLDEV_CMD_GETvia netlink to discover all RDMA devices - HW counters —
RDMA_NLDEV_CMD_STAT_GETper device/port, same asrdma statistic show - Process detection —
RDMA_NLDEV_CMD_RES_QP_GETto map QPs → PIDs, enriched with/procdata - Throughput — Two snapshots per interval, delta / elapsed for rates
See CONTRIBUTING.md for build/test instructions, design ground rules, and how to submit changes.
Apache-2.0