From 888b59ab21bdcc8fb42b5ef91cd317c8a4c3877e Mon Sep 17 00:00:00 2001 From: amitkumarlambda Date: Tue, 18 Nov 2025 08:24:39 -0600 Subject: [PATCH] Update gb300_n18_ngc25.04_nemo.json Correct incorrect CPU and other values, that was carried over from other systems --- Lambda/systems/gb300_n18_ngc25.04_nemo.json | 38 ++++++++++----------- 1 file changed, 19 insertions(+), 19 deletions(-) diff --git a/Lambda/systems/gb300_n18_ngc25.04_nemo.json b/Lambda/systems/gb300_n18_ngc25.04_nemo.json index a3fff640..051b73f5 100644 --- a/Lambda/systems/gb300_n18_ngc25.04_nemo.json +++ b/Lambda/systems/gb300_n18_ngc25.04_nemo.json @@ -5,7 +5,7 @@ "system_name": "Lambda_GB300_n18", "number_of_nodes": "18", "host_processors_per_node": "2", - "host_processor_model_name": "Intel(R) Xeon(R) Platinum 8570", + "host_processor_model_name": "Neoverse-V2", "host_processor_core_count": "72", "host_processor_vcpu_count": "", "host_processor_frequency": "", @@ -14,34 +14,34 @@ "host_memory_capacity": "2.0 TB", "host_storage_type": "Local NVMe + Shared File System", "host_storage_capacity": "22 TB Local NVMe, Unlmited Shared File System", - "host_networking": "Infiniband; Data bandwidth for GPU-PCIe: 504GB/s; PCIe-NIC: 500GB/s", + "host_networking": "4x ConnectX-8 IB XDR 800 Gb/s", "host_networking_topology": "Ethernet/Infiniband on switching network", - "host_memory_configuration": "32x 64GB HMCG94AGBRA179N", + "host_memory_configuration": "", "accelerators_per_node": "4", "accelerator_model_name": "NVIDIA Blackwell Ultra GPU (GB300)", - "accelerator_host_interconnect": "PCIe Gen5 x16", + "accelerator_host_interconnect": "", "accelerator_frequency": "", "accelerator_on-chip_memories": "", "accelerator_memory_configuration": "HBM3e", "accelerator_memory_capacity": "280 GB", - "accelerator_interconnect": "18x 5th Gen NVLink, 14.4 TB/s aggregated bandwidth", + "accelerator_interconnect": "NVLINK Gen5 1800 GB/s + NVSWITCH Gen5", "accelerator_interconnect_topology": "", - "cooling": "Air-cooled", - "hw_notes": "GB300 1400W", - "framework": "PyTorch NVIDIA Release 25.04", - "framework_name": "ngc25.04_pytorch_nemo_2.0.0", + "cooling": "", + "hw_notes": "", + "framework": "PyTorch NVIDIA Release 25.09", + "framework_name": "", "other_software_stack": { - "cuda_version": "12.9.0.034", - "cuda_driver_version": "575.50", - "nccl_version": "2.26.3", - "cublas_version": "12.9.0.2", - "cudnn_version": "9.9.0.52", - "trt_version": "10.9.0.34+cuda12.8", - "dali_version": "1.48.0", - "mofed_version": "5.4-rdmacore50.0", + "cuda_version": "13.0.1.012", + "cuda_driver_version": "580.82.07", + "nccl_version": "v2.28.3-1", + "cublas_version": "13.0.2.14", + "cudnn_version": "9.13.1.26", + "trt_version": "10.13.3.9", + "dali_version": "1.51.2", + "mofed_version": "5.4-rdmacore56.0", "openmpi_version": "4.1.7", - "kernel_version": "Linux 6.8.0-52-generic", - "nvidia_kernel_driver": "570.124.06" + "kernel_version": "Linux 6.11.0-1013-nvidia-64k", + "nvidia_kernel_driver": "580.95.05" }, "operating_system": "Ubuntu 24.04.2 LTS", "sw_notes": ""