PaddleX/zhch/utils/check_cuda_env.py

import subprocess
import sys

def check_nvidia_environment():
    print("=== NVIDIA环境检查 ===")
    
    # 检查nvidia-smi
    try:
        result = subprocess.run(['nvidia-smi'], capture_output=True, text=True, timeout=10)
        if result.returncode == 0:
            print("✓ nvidia-smi 正常运行")
            # 提取驱动版本
            lines = result.stdout.split('\n')
            for line in lines:
                if 'Driver Version:' in line:
                    print(f"  {line.strip()}")
                    break
        else:
            print("✗ nvidia-smi 失败:")
            print(result.stderr)
    except Exception as e:
        print(f"✗ nvidia-smi 错误: {e}")
    
    # 检查NVML
    try:
        import pynvml
        pynvml.nvmlInit()
        driver_version = pynvml.nvmlSystemGetDriverVersion()
        print(f"✓ NVML初始化成功，驱动版本: {driver_version}")
        
        device_count = pynvml.nvmlDeviceGetCount()
        print(f"✓ 检测到 {device_count} 个GPU设备")
        
        for i in range(device_count):
            handle = pynvml.nvmlDeviceGetHandleByIndex(i)
            name = pynvml.nvmlDeviceGetName(handle)
            print(f"  GPU {i}: {name}")
            
    except Exception as e:
        print(f"✗ NVML错误: {e}")
    
    # 检查PaddlePaddle
    try:
        import paddle
        print(f"✓ PaddlePaddle版本: {paddle.__version__}")
        print(f"✓ CUDA编译支持: {paddle.device.is_compiled_with_cuda()}")
        if paddle.device.is_compiled_with_cuda():
            gpu_count = paddle.device.cuda.device_count()
            print(f"✓ PaddlePaddle检测到 {gpu_count} 个GPU")
    except Exception as e:
        print(f"✗ PaddlePaddle错误: {e}")

if __name__ == "__main__":
    check_nvidia_environment()