Skip to content

faq.md

Latest commit

 

History

History
177 lines (140 loc) · 7.05 KB

faq.md

File metadata and controls

177 lines (140 loc) · 7.05 KB

Frequently Asked Questions (FAQ)

Quick answers to common questions about Intel® Edge Developer Kits.

Table of Contents

Getting Started

Q: I'm new to AI development. Where should I start?

A: Follow this learning path:

  1. Complete our Getting Started Guide
  2. Try the OpenWebUI + Ollama use case (10 minutes)
  3. Read AI Development Basics for concepts
  4. Join our community discussions to ask questions

Q: Do I need to buy specific hardware?

A: The scripts work with many Intel® platforms, but we recommend:

  • Students/Beginners: Any Intel® Core™ Ultra system with integrated graphics
  • Professionals: Intel® Core™ Ultra + Intel® Arc™ GPU (B580 or better)
  • Researchers: High-end CPU + Intel® Arc™ B60 Pro for maximum performance

Check our Hardware Selection Guide for detailed recommendations.

Q: How long does installation take?

A:

  • Quick install: 15-30 minutes (automatic)
  • With reboot: Add 5-10 minutes for system restart
  • Full validation: 45-60 minutes including use case testing

Hardware Compatibility

Q: Is my Intel® hardware supported?

A: Check the compatibility matrix. Supported platforms include:

  • Intel® Core™ Ultra (Series 1 & 2)
  • Intel® Arc™ Graphics (A-Series, B-Series)
  • Intel® 14th Gen Core™ processors
  • Intel® Core™ N-series

Q: What about older Intel® hardware?

A: Older hardware may work with reduced functionality:

  • 10th-12th Gen Intel® CPUs: Basic functionality, no NPU features
  • Older Intel® GPUs: Limited AI acceleration
  • Very old systems: May require manual driver installation

For best results, use hardware from our validated list.

Installation Issues

Q: The installer says "unsupported kernel version"

A: You need the HWE (Hardware Enablement) kernel:

sudo apt install linux-generic-hwe-24.04
sudo reboot
# Then rerun the installer

Q: Installation fails with permission errors

A: Make sure you're running with sudo:

sudo ./main_installer.sh

Not just ./main_installer.sh

Q: GPU not detected after installation

A: Try these steps:

  1. Reboot your system
  2. Enable Resizable BAR in BIOS (for Arc GPUs)
  3. Run the GPU installer separately: sudo ./gpu_installer.sh
  4. Check detection: lspci | grep -i vga

Q: Docker permission denied errors

A: Add your user to the docker group:

sudo usermod -aG docker $USER
# Log out and log back in

Performance & Optimization

Q: AI models are running slowly. How can I speed them up?

A: Check these optimization steps:

  1. GPU Memory: Ensure you have enough VRAM for your model
  2. Resizable BAR: Enable in BIOS for Intel Arc GPUs
  3. Model Size: Try smaller model variants for faster inference
  4. Device Selection: Verify workloads are using GPU, not CPU

Q: How do I check if NPU is being used?

A: NPU usage depends on the application:

  • Check detection: ls /dev/intel-npu*
  • Monitor usage: Some apps show device utilization in logs
  • Configure manually: Set environment variables like STT_DEVICE=NPU in docker-compose files

Q: Can I run multiple AI models simultaneously?

A: Yes, but consider:

  • Memory limits: Each model needs GPU/system memory
  • Performance impact: Multiple models compete for resources
  • Multi-GPU setups: Distribute models across different GPUs

Use Cases & Applications

Q: Which use case should I try first?

A: Based on your interests:

Q: Can I modify the use cases for my needs?

A: Absolutely! The use cases are starting points:

  • Code: All source code is available and modifiable
  • Models: Swap AI models for different capabilities
  • Configuration: Adjust settings via environment variables
  • Integration: Use as components in larger applications

Q: How do I deploy to production?

A: Consider these steps:

  1. Security: Review and harden configurations
  2. Monitoring: Add logging and health checks
  3. Scaling: Use Kubernetes or similar orchestration
  4. Updates: Establish model and software update processes

Troubleshooting

Q: Where can I find detailed error logs?

A: Check these locations:

  • Installer logs: Terminal output during installation
  • System logs: /var/log/ directory
  • Docker logs: docker logs [container_name]
  • Application logs: Usually in the use case directory

Q: Use case containers won't start

A: Common fixes:

  1. Check ports: Ensure required ports aren't in use
  2. Docker status: Verify Docker is running: systemctl status docker
  3. Permissions: Check file permissions in the use case directory
  4. Resources: Ensure enough disk space and memory

Q: Models won't download or load

A: Try these solutions:

  • Internet connection: Verify network connectivity
  • Disk space: Ensure enough space for model files (often several GB)
  • Permissions: Check that Docker can write to volume mounts
  • Alternative sources: Some models have mirror download locations

Q: Performance is worse than expected

A: Optimization checklist:

  • ✅ Latest drivers installed
  • ✅ Resizable BAR enabled (Intel Arc GPUs)
  • ✅ Adequate cooling (check for thermal throttling)
  • ✅ Power settings optimized for performance
  • ✅ Models running on GPU, not CPU fallback

Still Need Help?

Found a Bug?

  • Report it: GitHub Issues
  • Include: System info, error messages, steps to reproduce

Have a Question?

Want to Contribute?

  • Improvements: Submit pull requests with fixes or enhancements
  • Documentation: Help improve guides and examples
  • Use Cases: Share your custom implementations

Didn't find your answer? Ask in our community discussions - we're here to help!