You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Software-in-the-loop behavioral model of a GPU cluster modelled on NVIDIA GB200 NVL72, scaled down to two compute trays, on one Linux machine. Real Slurm or Kubernetes, PyTorch and NCCL run against emulated GPUs (CUDA, NVML, nvidia-smi), NVLink/NVSwitch partitions, Redfish BMCs and InfiniBand.
Install NVIDIA driver + Fabric Manager on multi-GPU NVSwitch boxes (A100/H100 SXM). Auto-picks a kernel-compatible, version-matched branch; verifies the DKMS build. Packaged as a .deb.
8x8 NoC router in Verilog RTL with virtual channels, credit flow control, multicast and an FP8 in-network reduction ALU. Implemented on ASAP7 7nm with Yosys, OpenROAD and OpenSTA through global routing; verified with cocotb.