Learn about new features in Vulkan 1.1, including subgroup functionality, the shader toolchain for HLSL in Vulkan, memory management, and more. The day will include breakout sessions to facilitate discussions on specific use cases, and lessons learnt by developers porting to Vulkan. Speakers confirmed from AMD, Google, LunarG, NVIDIA, Qualcomm and Samsung, with more to come! There will be plenty of opportunities to provide your feedback, and a Q&A panel comprised of the speakers of the day – so bring your toughest questions! After the sessions, there will be a social (with beer).
Graphics developers using Vulkan or other graphics APIs
|10:00||Introduction, Overview and Vulkan 1.1 Recap
Alon Or-bach, Samsung Electronics (@alonorbach)
A brief introduction to what to expect from the Vulkan Developer Day, what’s new in the world of Vulkan and how the Vulkan working group want to hear from developers on the direction of the API.
|10:20||Render Passes in Vulkan
Bill Licea-Kane, Qualcomm
What is to be gained by Render Pass/Subpass? Are there caveats? Why are there subpass dependencies and barriers? What happens before and what happens after? Are there known limitations?
|11:00||Vulkan Memory Management
Jordan Logan, AMD
Learn the advantages and challenges of managing memory in Vulkan. You will learn good practices and common patterns that can be applied to your code to use GPU memory more efficiently, as well as some more advanced tips and tricks related to this topic. You will also learn about a library that is available to help you with this task.
|11:45||Vulkan Assistant Layer and Layer Factory
Introduction to the Vulkan Assistant Layer and the Vulkan Layer Factory, including short demos highlighting ways these tools can aid Vulkan developers, and where we go from here.
|12:15||Shader Toolchain: HLSL in Vulkan
Lei Zhang, Google
Support for HLSL in Vulkan has gained a lot of ground over the past year with expanded language support, legalization, and reflection. Come hear the latest developments in DXC and spirv-opt backing this effort. Learn about recently added compiler features that make HLSL source sharing easier than ever.
|13:45||Vulkan Subgroup Functionality
Daniel Koch, NVIDIA (@booner_k)
Vulkan 1.1 subgroups explained, , partitioned subgroups, NVIDIA implementation details, comparison with HLSL SM 6.0 Wave Ops.
|14:30||ANGLE: OpenGL on Vulkan
Jamie Madill, Google
The ANGLE team is developing an open source OpenGL ES implementation on Vulkan. This short presentation will cover its potential to reduce fragmentation on Android.
|14:45||SwiftShader: Reference Implementation and Fallback
Nicolas Capens, Google (@c0d1f1ed)
The SwiftShader project is aiming to become Google's reference implementation for Vulkan, to provide continuous testing and to offer a reliable fallback. Come hear about our architecture and design decisions, and provide your input!
|15:20||Using Vulkan-Hpp, The C++ Header for Vulkan
Markus Tavenrath, NVIDIA
Vulkan-Hpp is a C++ binding of the Vulkan-API with the goal of bringing features of the C++ language to the Vulkan API while staying as close as possible to the original API and with no added runtime cost.Come hear about the latest enhancements and provide feedback for further improvements of the bindings.
Hai Nguyen, Google
Descriptor indexing is one of the most exciting extensions to land on Vulkan. This extension brings large descriptor sets and dynamic resource indexing enables advanced rendering and compute techniques. Come learn how to employ this in Vulkan and the caveats that go along with it.
Your chance to ask our speakers questions on all things Vulkan
What lessons have we learnt in the two years since Vulkan was released? What issues should developers watch out for? What difficulties have you faced porting to Vulkan? What could the working group do better? Get your questions ready for a lively discussion!
More Vulkan discussion ... but with beer.
Conference Code of Conduct: The Khronos Group is dedicated to providing a harassment-free conference experience for everyone. Visit our Code of Conduct page to learn more.