Microsoft Build 2023, Qualcomm with Microsoft illustrates the innovations in the field of on-device generative AI, which make these solutions even more accessible, reliable and private. These include generative AI running on Snapdragon compute platforms, as well as new ones routes for developers building applications for Snapdragon-powered Windows 11 PCs.
New user experiences
Ziad Asghar, Senior Vice President of Product Management di Qualcomm Technologies
For generative AI to truly go mainstream, much of the inference will need to run on edge devices. Our best-in-class AI hardware and software allows developers to take full advantage of our powerful AI capabilities. Thus delivering amazing new user experiences on Snapdragon-powered laptops, phones, and other devices.
On-device AI even in smaller devices
The future of AI must include both on-device and cloud-based AI. Running AI applications on the device improves cost effectiveness, privacy, personalization and latency. Snapdragon platforms include a dedicated Qualcomm AI Engine that processes AI workloads more efficient compared to processing purely on the CPU or GPU. This makes on-device AI possible even in small, thin and light devices.
More accessible, reliable and private
Qualcomm Technologies demonstrated tools to help develop generative AI on the next generation of Windows 11 PCs at the event. Stable Diffusion, a text-to-image generative AI model with over 1 billion parameters, now runs entirely on the device . The firm also claimed to expect models to run with up to 10 billion parameters on devices in the coming months, including large language models (LLM). To enhance the Qualcomm AI Stack and help developers build the next generation of on-device AI experiences, the Qualcomm AI Engine Direct SDK first made available during the Microsoft Build. With ONNX Runtime and Qualcomm AI Stack, you can run AI workloads on Snapdragon 8cx Gen 3 AI Engine on both Surface Pro 9 5G and Lenovo ThinkPad x13s.
Accelerate opportunities
Pavan Davuluri, Corporate VP, Windows Silicon and System Integration di Microsoft
Extending AI to the vast universe of devices and applications requires both cloud and device computing. By bringing together Microsoft’s leadership in cloud AI and Windows platform capabilities with Qualcomm Technologies’ expertise in on-device AI, we will accelerate opportunities for generative AI experiences.
More accessible, reliable and private
During Microsoft Build, Qualcomm Technologies also demonstrated the performance improvements available when popular applications are built natively for Snapdragon. Has debuted Also a new developer portal that brings together the tools, resources and support for developing Windows on Snapdragon in one easy-to-use place.