OpenCL data types.
These are the data types available in the OpenCL C programming language. Click on a category name in the table below to see information about specific data types.
Data type category | Included data types |
---|---|
|
|
|
|
|
|
|
|
|
A data item declared to be a data type in memory is always aligned to the size of the data type in bytes.
For example, a float4
variable will be aligned to a 16-byte boundary, and a char2
variable will be aligned to a 2-byte boundary.
For 3-component vector data types, the size of the data type is 4 * sizeof(component)
.
This means that a 3-component vector data type will be aligned to a 4 * sizeof(component)
boundary.
The vload3
and vstore3
built-in functions can be used to read and write, respectively, 3-component vector data types from an array of packed scalar data type.
A built-in data type that is not a power of two bytes in size must be aligned to the next larger power of two.
This rule applies to built-in types only, not structs
or unions
.
The OpenCL compiler is responsible for aligning data items to the appropriate alignment as required by the data type.
For arguments to a __kernel
function declared to be a pointer to a data type, the OpenCL compiler can assume that the pointee is always appropriately aligned as required by the data type.
The behavior of an unaligned load or store is undefined, except for the vloadn
, vload_halfn
, vstoren
, and vstore_halfn
functions.
The vector load functions can read a vector from an address aligned to the element type of the vector.
The vector store functions can write a vector to an address aligned to the element type of the vector.
The user is responsible for ensuring that data passed into and out of OpenCL buffers are natively aligned relative to the start of the buffer as described above.
This implies that OpenCL buffers created with CL_MEM_USE_HOST_PTR
need to provide an appropriately aligned host memory pointer that is aligned to the data types used to access these buffers in a kernel(s).
As well, the user is responsible to ensure that data passed into and out of OpenCL images are properly aligned to the granularity of the data representing a single pixel (e.g.
image_num_channels
* sizeof
(image_channel_data_type
)) except for CL_RGB
and CL_RGBx
images where the data must be aligned to the granularity of a single channel in a pixel (i.e.
sizeof
(image_channel_data_type
)).
This implies that OpenCL images created with CL_MEM_USE_HOST_PTR
must align correctly.
The image alignment value can be queried using the CL_DEVICE_IMAGE_BASE_ADDRESS_ALIGNMENT
query.
In addition, source pointers for clEnqueueWriteBuffer
, clEnqueueWriteImage
and other operations that copy to the OpenCL runtime, as well as destination pointers for clEnqueueReadBuffer
, clEnqueueReadImage
and other operations that copy from the OpenCL runtime must follow the same alignment rules.
OpenCL makes no requirement about the alignment of OpenCL application defined data types outside of buffers and images, except that the underlying vector primitives (e.g.
__cl_float4
) where defined shall be directly accessible as such using appropriate named fields in the cl_type
union.
Nevertheless, it is recommended that the cl_platform.h
header should attempt to naturally align OpenCL defined application data types (e.g.
cl_float4
) according to their type.