import torch
print('PyTorch version:', torch.__version__)
print('CUDA available:', torch.cuda.is_available())
if torch.cuda.is_available():
    print('CUDA device:', torch.cuda.get_device_name(0))
    print('CUDA version:', torch.version.cuda)
    print('VRAM:', round(torch.cuda.get_device_properties(0).total_memory / 1024**3, 2), 'GB')
    # Test actual GPU computation
    try:
        x = torch.randn(100, 100, device='cuda')
        y = torch.randn(100, 100, device='cuda')
        z = torch.mm(x, y)
        print('GPU computation test: PASSED')
        print('Result shape:', z.shape, 'sum:', z.sum().item())
    except Exception as e:
        print('GPU computation test: FAILED -', e)
        print('Falling back to CPU mode')
else:
    print('WARNING: CUDA not available, will use CPU')