VkLatencySubmissionPresentIdNV

The VkLatencySubmissionPresentIdNV structure is defined as:

typedef struct VkLatencySubmissionPresentIdNV {
    VkStructureType sType;
    const void* pNext;
    uint64_t presentID;
} VkLatencySubmissionPresentIdNV;

pub struct LatencySubmissionPresentIdNV {
    s_type: vk::StructureType,
    p_next: *const c_void,
    present_id: u64,
}

• sType is a VkStructureType value identifying this structure.
• pNext is NULL or a pointer to a structure extending this structure.
• presentID is the application-defined identifier of the application-rendered frame that this submission belongs to. It is not related to presentIds passed via VkPresentIdKHR::pPresentIds or VkPresentId2KHR::pPresentIds.

If the implicit submission attribution rules defined in Frame Attribution are not sufficient, an application can associate submissions with application-rendered frames by chaining VkLatencySubmissionPresentIdNV to its vkQueueSubmit or vkQueueSubmit2 calls. A non-zero presentID passed via VkLatencySubmissionPresentIdNVmust be greater than or equal to any non-zero presentID previously passed via VkLatencySubmissionPresentIdNV on the same queue.

A logical device enters explicit submission attribution mode when any queue on that device receives a VkLatencySubmissionPresentIdNV with a non-zero presentID. While a device is in explicit submission attribution mode, each vkQueueSubmit or vkQueueSubmit2 call will be associated with the presentID value of the first of these rules to apply:

• The presentID in VkLatencySubmissionPresentIdNV chained to the submission, if non-zero.
• The most recent presentID passed via VkLatencySubmissionPresentIdNV on the same queue, if non-zero.
• The maximum presentID previously passed via VkLatencySubmissionPresentIdNV on the device.

A vkQueueSubmit or vkQueueSubmit2 call that chains VkLatencySubmissionPresentIdNV with a presentID of zero clears the tracked presentID for that queue. The device will exit explicit submission attribution mode when all queues' tracked presentID values have been cleared. It will then return to using the implicit attribution rules.

Explicit submission attribution is only respected by implementations advertising support for VK_NV_low_latency2 with a VkExtensionProperties::specVersion greater than or equal to 3. Implementations advertising a lower specVersion ignore VkLatencySubmissionPresentIdNV, so applications should verify the advertised specVersion before relying on explicit submission attribution.