# GPU基准测试工具
这是一个用于测试NVIDIA GPU性能的综合工具，可以评估GPU的各项性能指标。
## 功能特点
- 环境检查：自动检测NVIDIA驱动、CUDA和PyTorch环境
- 系统依赖安装：自动安装必要的系统依赖
- CUDA基础能力测试：包括设备信息查询和内存带宽测试
- GPU计算性能与稳定性测试：使用gpu-burn进行压力测试
- 大模型推理测试：测试Stable Diffusion和LLaMA等模型的推理性能
## 安装方法
```bash
# 从PyPI安装
pip install gpu-benchmark-linux
# 或者从源码安装
git clone https://github.com/yourusername/gpu-benchmark-linux.git
cd gpu-benchmark-linux
pip install -e .
```
## 使用方法
```bash
# 运行完整测试
gpu-benchmark-linux
# 或者使用Python模块方式运行
python -m gpu-benchmark-linux
# 运行特定测试
gpu-benchmark-linux --test cuda
gpu-benchmark-linux --test model
```
## 测试结果
测试结果将保存在`./gpu-benchmark-linux_results`目录下，包含详细的日志信息。
## 系统要求
- NVIDIA GPU
- NVIDIA驱动
- CUDA Toolkit（推荐）
- Python 3.7+
Raw data
{
"_id": null,
"home_page": null,
"name": "gpu-benchmark-linux",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.7",
"maintainer_email": null,
"keywords": "gpu, benchmark, cuda, nvidia, testing, performance",
"author": null,
"author_email": "GPU Benchmark Team <your.email@example.com>",
"download_url": "https://files.pythonhosted.org/packages/ac/98/aa954b08060d90f80bfd9c83de50345e000f2c52d343d2896c4a703d4741/gpu_benchmark_linux-0.1.0.tar.gz",
"platform": null,
"description": "# GPU\u57fa\u51c6\u6d4b\u8bd5\u5de5\u5177\n\n\u8fd9\u662f\u4e00\u4e2a\u7528\u4e8e\u6d4b\u8bd5NVIDIA GPU\u6027\u80fd\u7684\u7efc\u5408\u5de5\u5177\uff0c\u53ef\u4ee5\u8bc4\u4f30GPU\u7684\u5404\u9879\u6027\u80fd\u6307\u6807\u3002\n\n## \u529f\u80fd\u7279\u70b9\n\n- \u73af\u5883\u68c0\u67e5\uff1a\u81ea\u52a8\u68c0\u6d4bNVIDIA\u9a71\u52a8\u3001CUDA\u548cPyTorch\u73af\u5883\n- \u7cfb\u7edf\u4f9d\u8d56\u5b89\u88c5\uff1a\u81ea\u52a8\u5b89\u88c5\u5fc5\u8981\u7684\u7cfb\u7edf\u4f9d\u8d56\n- CUDA\u57fa\u7840\u80fd\u529b\u6d4b\u8bd5\uff1a\u5305\u62ec\u8bbe\u5907\u4fe1\u606f\u67e5\u8be2\u548c\u5185\u5b58\u5e26\u5bbd\u6d4b\u8bd5\n- GPU\u8ba1\u7b97\u6027\u80fd\u4e0e\u7a33\u5b9a\u6027\u6d4b\u8bd5\uff1a\u4f7f\u7528gpu-burn\u8fdb\u884c\u538b\u529b\u6d4b\u8bd5\n- \u5927\u6a21\u578b\u63a8\u7406\u6d4b\u8bd5\uff1a\u6d4b\u8bd5Stable Diffusion\u548cLLaMA\u7b49\u6a21\u578b\u7684\u63a8\u7406\u6027\u80fd\n\n## \u5b89\u88c5\u65b9\u6cd5\n\n```bash\n# \u4ecePyPI\u5b89\u88c5\npip install gpu-benchmark-linux\n\n# \u6216\u8005\u4ece\u6e90\u7801\u5b89\u88c5\ngit clone https://github.com/yourusername/gpu-benchmark-linux.git\ncd gpu-benchmark-linux\npip install -e .\n```\n\n## \u4f7f\u7528\u65b9\u6cd5\n\n```bash\n# \u8fd0\u884c\u5b8c\u6574\u6d4b\u8bd5\ngpu-benchmark-linux\n\n# \u6216\u8005\u4f7f\u7528Python\u6a21\u5757\u65b9\u5f0f\u8fd0\u884c\npython -m gpu-benchmark-linux\n\n# \u8fd0\u884c\u7279\u5b9a\u6d4b\u8bd5\ngpu-benchmark-linux --test cuda\ngpu-benchmark-linux --test model\n```\n\n## \u6d4b\u8bd5\u7ed3\u679c\n\n\u6d4b\u8bd5\u7ed3\u679c\u5c06\u4fdd\u5b58\u5728`./gpu-benchmark-linux_results`\u76ee\u5f55\u4e0b\uff0c\u5305\u542b\u8be6\u7ec6\u7684\u65e5\u5fd7\u4fe1\u606f\u3002\n\n## \u7cfb\u7edf\u8981\u6c42\n\n- NVIDIA GPU\n- NVIDIA\u9a71\u52a8\n- CUDA Toolkit\uff08\u63a8\u8350\uff09\n- Python 3.7+\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "GPU\u57fa\u51c6\u6d4b\u8bd5\u5de5\u5177 - \u7528\u4e8e\u8bc4\u4f30NVIDIA GPU\u6027\u80fd\u7684\u7efc\u5408\u5de5\u5177\u5305",
"version": "0.1.0",
"project_urls": {
"Bug Tracker": "https://github.com/yourusername/gpu-benchmark-linux/issues",
"Documentation": "https://github.com/yourusername/gpu-benchmark-linux#readme",
"Homepage": "https://github.com/yourusername/gpu-benchmark-linux"
},
"split_keywords": [
"gpu",
" benchmark",
" cuda",
" nvidia",
" testing",
" performance"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "a1677f5f6b7343cc0dec7f5a8c650bfc0225c1e50077878ee20da628a1e4e38b",
"md5": "d5dcfa1aa84bb3bcd26b40996b4742a7",
"sha256": "fdd10a2188c429d76464d36d787bb33d212eeee53785791bcb0c79eb8888aad7"
},
"downloads": -1,
"filename": "gpu_benchmark_linux-0.1.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "d5dcfa1aa84bb3bcd26b40996b4742a7",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.7",
"size": 13110,
"upload_time": "2025-08-15T09:46:34",
"upload_time_iso_8601": "2025-08-15T09:46:34.127079Z",
"url": "https://files.pythonhosted.org/packages/a1/67/7f5f6b7343cc0dec7f5a8c650bfc0225c1e50077878ee20da628a1e4e38b/gpu_benchmark_linux-0.1.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "ac98aa954b08060d90f80bfd9c83de50345e000f2c52d343d2896c4a703d4741",
"md5": "4b38dd1be43fe48845d4444374deca10",
"sha256": "f0fe6ebe3e2c94f55052b7dfd8f7de0342ddf0991d50e4b14f5d95e556b82679"
},
"downloads": -1,
"filename": "gpu_benchmark_linux-0.1.0.tar.gz",
"has_sig": false,
"md5_digest": "4b38dd1be43fe48845d4444374deca10",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.7",
"size": 11382,
"upload_time": "2025-08-15T09:46:36",
"upload_time_iso_8601": "2025-08-15T09:46:36.179057Z",
"url": "https://files.pythonhosted.org/packages/ac/98/aa954b08060d90f80bfd9c83de50345e000f2c52d343d2896c4a703d4741/gpu_benchmark_linux-0.1.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-08-15 09:46:36",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "yourusername",
"github_project": "gpu-benchmark-linux",
"github_not_found": true,
"lcname": "gpu-benchmark-linux"
}
JSON
