C Specification

The VkPhysicalDeviceShaderFloat16Int8Features structure is defined as:

// Provided by VK_VERSION_1_2
typedef struct VkPhysicalDeviceShaderFloat16Int8Features {
    VkStructureType    sType;
    void*              pNext;
    VkBool32           shaderFloat16;
    VkBool32           shaderInt8;
} VkPhysicalDeviceShaderFloat16Int8Features;

or the equivalent

// Provided by VK_KHR_shader_float16_int8
typedef VkPhysicalDeviceShaderFloat16Int8Features VkPhysicalDeviceShaderFloat16Int8FeaturesKHR;
// Provided by VK_KHR_shader_float16_int8
typedef VkPhysicalDeviceShaderFloat16Int8Features VkPhysicalDeviceFloat16Int8FeaturesKHR;

Members

This structure describes the following features:

  • sType is a VkStructureType value identifying this structure.

  • pNext is NULL or a pointer to a structure extending this structure.

Description

  • shaderFloat16 indicates whether 16-bit floats (halfs) are supported in shader code. This also indicates whether shader modules can declare the Float16 capability. However, this only enables a subset of the storage classes that SPIR-V allows for the Float16 SPIR-V capability: Declaring and using 16-bit floats in the Private, Workgroup (for non-Block variables), and Function storage classes is enabled, while declaring them in the interface storage classes (e.g., UniformConstant, Uniform, StorageBuffer, Input, Output, and PushConstant) is not enabled.

  • shaderInt8 indicates whether 8-bit integers (signed and unsigned) are supported in shader code. This also indicates whether shader modules can declare the Int8 capability. However, this only enables a subset of the storage classes that SPIR-V allows for the Int8 SPIR-V capability: Declaring and using 8-bit integers in the Private, Workgroup (for non-Block variables), and Function storage classes is enabled, while declaring them in the interface storage classes (e.g., UniformConstant, Uniform, StorageBuffer, Input, Output, and PushConstant) is not enabled.

If the VkPhysicalDeviceShaderFloat16Int8Features structure is included in the pNext chain of the VkPhysicalDeviceFeatures2 structure passed to vkGetPhysicalDeviceFeatures2, it is filled in to indicate whether each corresponding feature is supported. VkPhysicalDeviceShaderFloat16Int8Features can also be used in the pNext chain of VkDeviceCreateInfo to selectively enable these features.

Valid Usage (Implicit)

  • VUID-VkPhysicalDeviceShaderFloat16Int8Features-sType-sType
    sType must be VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SHADER_FLOAT16_INT8_FEATURES

See Also

VK_KHR_shader_float16_int8, VK_VERSION_1_2, VkBool32, VkStructureType

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

Copyright 2014-2024 The Khronos Group Inc.

SPDX-License-Identifier: CC-BY-4.0