Skip to content

debug

Get detailed information from all detectors for troubleshooting.

Usage

env-doctor debug

What It Shows

All Detector Results

  • Raw output from every registered detector
  • Detection status (success, warning, error, not_found)
  • Version information

Detection Metadata

  • Internal detection methods used
  • File paths examined
  • Detailed status information

Registry Information

  • List of all available detectors
  • Detector execution order

Error Details

  • Full exception traces
  • Diagnostic information for failures

Example Output

🔍 DEBUG MODE - Detailed Detector Information
============================================================
Registered Detectors: cuda_toolkit, nvidia_driver, python_library, wsl2, cudnn

--- WSL2 ---
Status: Status.SUCCESS
Component: wsl2
Version: wsl2
Metadata: {
  'environment': 'WSL2',
  'gpu_forwarding': 'enabled',
  'wsl_lib_path': '/usr/lib/wsl/lib',
  'libcuda_found': True
}

--- NVIDIA DRIVER ---
Status: Status.SUCCESS
Component: nvidia_driver
Version: 535.146.02
Path: /usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1
Metadata: {
  'max_cuda_version': '12.2',
  'detection_method': 'nvml',
  'gpu_name': 'NVIDIA GeForce RTX 3090',
  'gpu_memory': 24576
}

--- CUDA TOOLKIT ---
Status: Status.SUCCESS
Component: cuda_toolkit
Version: 12.1.1
Path: /usr/local/cuda-12.1
Metadata: {
  'nvcc_path': '/usr/local/cuda-12.1/bin/nvcc',
  'cuda_home': '/usr/local/cuda-12.1',
  'multiple_installations': ['/usr/local/cuda-12.1', '/usr/local/cuda-11.8']
}

--- CUDNN ---
Status: Status.SUCCESS
Component: cudnn
Version: 8.9.7
Path: /usr/local/cuda-12.1/lib64/libcudnn.so.8
Metadata: {
  'cuda_version': '12.x',
  'header_path': '/usr/local/cuda-12.1/include/cudnn.h'
}

--- PYTHON LIBRARIES ---
Status: Status.SUCCESS
Component: python_library
Libraries: {
  'torch': {
    'version': '2.1.0+cu121',
    'cuda_version': '12.1',
    'detection_method': 'import'
  },
  'tensorflow': {
    'status': 'not_installed'
  }
}

When to Use Debug Mode

Unexpected Results

When env-doctor check shows results that don't match your expectations:

# Regular check shows CUDA not found, but you know it's installed
env-doctor check
# Shows: ⚠️ No system CUDA toolkit found

# Debug to see what happened
env-doctor debug
# Shows: Detection attempted at /usr/local/cuda (symlink broken)

Contributing or Reporting Issues

Include debug output when:

  • Reporting bugs
  • Contributing new detectors
  • Validating behavior changes

Understanding Detection Methods

Debug shows how each component was detected:

  • NVML for driver detection
  • File system search for CUDA
  • Python imports for libraries

See Also