Announcing new Vulkan® ray tracing extension support in our latest driver

Following a brief spell in our beta driver, we’re delighted to announce that support for the following new Vulkan® ray tracing extensions is now available in our main Radeon™ Adrenalin driver 20.11.3 onwards:

  • VK_KHR_acceleration_structure
    • This extension provides acceleration structures for representing geometry that is spatially sorted. Acceleration structures are the most common way for quickly identifying potential ray intersections by ray tracing techniques.
  • VK_KHR_ray_tracing_pipeline
    • This extension introduces ray tracing pipelines along with new shader domains and an indirection table to link shader groups with acceleration structures.
  • VK_KHR_ray_query
    • This extension introduces ray queries, which can be used by any shader type outside of the optional dedicated ray tracing pipeline to return traversal results to the calling shader.
  • VK_KHR_deferred_host_operations
    • This extension defines the infrastructure and usage patterns for deferrable commands, which allows dependent extensions to defer their own operations. The VK_KHR_deferred_host_operations extension itself does not specify any commands as deferrable.
  • VK_KHR_pipeline_library
    • This extension introduces pipeline libraries. A pipeline library is a new special pipeline type that defines a set of pipeline states. It cannot be bound, rather it can be linked into other pipelines.

In addition to the ray tracing extensions above, there’s also now support for:

  • VK_KHR_shader_terminate_invocation
    • This extension indicates support for the SPV_KHR_terminate_invocation SPIR-V extension. The SPIR-V extension introduces a new instruction OpTerminateInvocation that allows a shader invocation to immediately terminate, which provides the behavior required by the GLSL discard statement.

Our Vulkan® Memory Allocator library is already compatible, as VkAccelerationStructures are created on top of VkBuffers.

More GPUOpen Vulkan® content

AMD GPUOpen Vulkan Memory Allocator

Vulkan® Memory Allocator

VMA is our single-header, MIT-licensed, C++ library for easily and efficiently managing memory allocation for your Vulkan® games and applications.


Vulkan® gives software developers control over the performance, efficiency, and capabilities of AMD Radeon™ GPUs and multi-core CPUs.

Developing Vulkan® applications

Discover our Vulkan blog posts, presentations, samples, and more. Find out how we can help you create and optimize your Vulkan applications!