GDC 2025 Session Viewer
Runtime Hardware Acceleration of Neural Networks in Games (Presented by Qualcomm Technologies)
Carlos Castrillon (Senior Staff Engineer, Qualcomm Technologies, Inc.)
Location: Room 3022, West Hall
Date: Thursday, March 20
Time: 12:15 pm - 1:15 pm
Pass Type: All Access Pass, Core Pass, Summits Pass, Expo Pass, Audio Pass, Indie Games Summit Pass - Get your pass now!
Track: Programming
Format: Sponsored Session
Vault Recording: Not Recorded
Audience Level: All
During this session, we will cover real-time machine learning inference techniques on mobile platforms. The discussion will cover optimization techniques and will delve into the tools available as part of Qualcomm's AI stack of solutions for deploying large neural network models on mobile platforms. Additionally, we will go over integration of our solutions into Unreal Engine for enabling real-time ML inference in games.
Takeaway
Attendees will learn how to integrate Machine Learning models into game projects and take advantage of dedicated neural network hardware acceleration on mobile platforms. They will also learn how to execute optimization techniques for neural networks models using Qualcomm's SDK tools.
Intended Audience
This talk is intended for any game developers (particularly programmers and AI engineers) interested in running machine-learning models in game applications.
