英伟达 H100 机器基础设施报告 | AI生成和翻译
🖥️ Machine Report: 162.243.29.96
| Component |
Details |
| OS |
Ubuntu 22.04.4 LTS (Jammy) |
| Kernel |
5.15.0-113-generic |
| Location |
NYC2 (New York) |
| Type |
KVM Virtual Machine |
🧠 CPU
| |
|
| Model |
Intel Xeon Platinum 8468 |
| Cores |
20 cores, 1 thread per core (no HT) |
| Architecture |
x86_64, Sapphire Rapids |
| Cache |
L1d: 640 KiB / L1i: 640 KiB / L2: 80 MiB |
| ISA |
AVX-512 (full suite), AVX2, VNNI, BF16, FP16 |
🎮 GPU
| |
|
| Model |
NVIDIA H100 80GB HBM3 |
| VRAM |
81,559 MiB (~80 GB) |
| Driver |
575.57.08 |
| CUDA |
12.9 |
| Status |
Idle (0% util, 37°C, 73W / 700W TDP) |
| MIG |
Disabled |
💾 Memory
| |
|
| RAM |
235 GiB total (233 GiB available) |
| Swap |
None |
💿 Disk
| |
|
| Root |
/dev/vda1 — 698 GB (20 GB used, 679 GB free, 3%) |
🌐 Network
| Interface |
IP |
| eth0 (public) |
162.243.29.96/24 |
| eth0 (private) |
10.13.0.5/16 |
| eth1 |
10.100.0.2/20 |
| docker0 |
172.17.0.1/16 |
Summary
顶级配置。 H100 80GB + 235GB RAM + 20 Xeon cores + 700GB disk。非常适合 LLM inference —— 可以轻松运行大型模型 (70B+)。需要我在这里也安装 Ollama 吗？