NVIDIA GPUs for Machine Learning
用于机器学习的 NVIDIA GPU
常用于机器学习和 LLM 推理/训练的 NVIDIA GPU 综合列表
| GPU 名称 | 内存 | 内存类型 | 架构 | 功耗 | CUDA 核心 | Tensor 核心 | 年份 |
|---|---|---|---|---|---|---|---|
| NVIDIA Tesla H200 NVL | 141 GB | HBM3e | Hopper | 600 W | 16,896 | ✓ | 2024 |
| NVIDIA Tesla H100 SXM5 | 80 GB | HBM3 | Hopper | 700 W | 16,896 | ✓ | 2022 |
| NVIDIA Tesla H100 SXM5 | 80 GB | HBM3 | Hopper | 700 W | 16,896 | ✓ | 2022 |
| NVIDIA Tesla H100 PCIe | 80 GB | HBM2e | Hopper | 350 W | 14,592 | ✓ | 2022 |
| NVIDIA Tesla A100 (80GB) | 80 GB | HBM2 | Ampere | 400 W | 6,912 | ✓ | 2020 |
| NVIDIA Tesla A100 (40GB) | 40 GB | HBM2 | Ampere | 400 W | 6,912 | ✓ | 2020 |
| NVIDIA Tesla A40 | 48 GB | GDDR6 | Ampere | 300 W | 10,752 | ✓ | 2020 |
| NVIDIA Tesla A30 | 24 GB | HBM2 | Ampere | 165 W | 3,584 | ✓ | 2021 |
| NVIDIA Tesla A10 | 24 GB | GDDR6 | Ampere | 150 W | 9,216 | ✓ | 2021 |
| NVIDIA Tesla A2 | 16 GB | GDDR6 | Ampere | 60 W | 1,280 | ✓ | 2021 |
| NVIDIA Tesla V100 (32GB) | 32 GB | HBM2 | Volta | 300 W | 5,120 | ✓ | 2018 |
| NVIDIA Tesla V100 (16GB) | 16 GB | HBM2 | Volta | 300 W | 5,120 | ✓ | 2017 |
| NVIDIA Tesla T4 | 16 GB | GDDR6 | Turing | 70 W | 2,560 | ✓ | 2018 |
| NVIDIA Tesla P100 (16GB) | 16 GB | HBM2 | Pascal | 250 W | 3,584 | ✗ | 2016 |
| NVIDIA Tesla P40 | 24 GB | GDDR5 | Pascal | 250 W | 3,840 | ✗ | 2016 |
| NVIDIA RTX 6000 Ada | 48 GB | GDDR6 | Ada Lovelace | 300 W | 18,176 | ✓ | 2022 |
| NVIDIA RTX A6000 | 48 GB | GDDR6 | Ampere | 300 W | 10,752 | ✓ | 2020 |
| NVIDIA RTX A5000 | 24 GB | GDDR6 | Ampere | 230 W | 8,192 | ✓ | 2021 |
| NVIDIA RTX A4000 | 16 GB | GDDR6 | Ampere | 140 W | 6,144 | ✓ | 2021 |
| NVIDIA L40 | 48 GB | GDDR6 | Ada Lovelace | 300 W | 18,176 | ✓ | 2022 |
| NVIDIA L4 | 24 GB | GDDR6 | Ada Lovelace | 72 W | 7,424 | ✓ | 2023 |
| NVIDIA GeForce RTX 4090 | 24 GB | GDDR6X | Ada Lovelace | 450 W | 16,384 | ✓ | 2022 |
| NVIDIA GeForce RTX 3090 | 24 GB | GDDR6X | Ampere | 350 W | 10,496 | ✓ | 2020 |
注意：这些数据来自各种来源，可能不完整。有关最准确和最新的规格，请参阅 NVIDIA 的官方文档。