C Specification

To record a CUDA kernel launch, call:

// Provided by VK_NV_cuda_kernel_launch
void vkCmdCudaLaunchKernelNV(
    VkCommandBuffer                             commandBuffer,
    const VkCudaLaunchInfoNV*                   pLaunchInfo);


  • commandBuffer is the command buffer into which the command will be recorded.

  • pLaunchInfo is a pointer to a VkCudaLaunchInfoNV structure in which the grid (similar to workgroup) dimension, function handle and related arguments are defined.


When the command is executed, a global workgroup consisting of gridDimX × gridDimY × gridDimZ local workgroups is assembled.

Valid Usage (Implicit)
  • VUID-vkCmdCudaLaunchKernelNV-commandBuffer-parameter
    commandBuffer must be a valid VkCommandBuffer handle

  • VUID-vkCmdCudaLaunchKernelNV-pLaunchInfo-parameter
    pLaunchInfo must be a valid pointer to a valid VkCudaLaunchInfoNV structure

  • VUID-vkCmdCudaLaunchKernelNV-commandBuffer-recording
    commandBuffer must be in the recording state

  • VUID-vkCmdCudaLaunchKernelNV-commandBuffer-cmdpool
    The VkCommandPool that commandBuffer was allocated from must support graphics, or compute operations

  • VUID-vkCmdCudaLaunchKernelNV-videocoding
    This command must only be called outside of a video coding scope

Host Synchronization
  • Host access to the VkCommandPool that commandBuffer was allocated from must be externally synchronized

Command Properties
Command Buffer Levels Render Pass Scope Video Coding Scope Supported Queue Types Command Type






See Also

Document Notes

For more information, see the Vulkan Specification

This page is extracted from the Vulkan Specification. Fixes and changes should be made to the Specification, not directly.

Copyright 2014-2024 The Khronos Group Inc.

SPDX-License-Identifier: CC-BY-4.0