dgx-spark-roce-crs804-fix
GB10 cluster stuck at ~10 Gb/s through a MikroTik CRS804? It was the switch — in three places How a 4× DGX Spark (GB10) cluster went from a hard ~10 Gb/s ceiling and a ~5 Gb/s single-flow RoCE collapse (NCCL all-reduce: 0.46 GB/s) to 109 Gb/s per rail and 23–26 GB/s NCCL busbw — by… Read More »dgx-spark-roce-crs804-fix