Namespaces
	access

	codeplay

	experimental

	ext

	half_precision

	info

	native

	property

Classes
class	accelerator_selector
	Implementation of a device_selector that selects an accelerator device. More...

class	accessor
	Public facing interface class for allowing users access to buffer objects, image objects and local memory from within kernel functions and the host. More...

class	accessor< elemT, 0, kMode, access::target::local, access::placeholder::false_t >
	Specialization for a 0-dimensional local accessor. More...

class	accessor< elemT, kDims, kMode, access::target::host_buffer, access::placeholder::false_t >

class	accessor< elemT, kDims, kMode, access::target::host_image, access::placeholder::false_t >
	Specialization for a host image accessor. More...

class	accessor< elemT, kDims, kMode, access::target::host_task, IsPlaceholder >

class	accessor< elemT, kDims, kMode, access::target::image, access::placeholder::false_t >
	Specialization for an image accessor. More...

class	accessor< elemT, kDims, kMode, access::target::image_array, access::placeholder::false_t >

class	accessor< elemT, kDims, kMode, access::target::local, access::placeholder::false_t >
	Specialization for a local accessor. More...

class	accessor< elemT, kDims, kMode, access::target::subgroup_local, access::placeholder::false_t >
	Specialization for a subgroup_local accessor. More...

class	accessor_error
	Represents an error regarding cl::sycl::accessor objects defined. More...

class	amd_selector
	Implementation of a device_selector that selects an AMD platform. More...

struct	atomic
	Atomic class template. More...

class	backend_traits
	Defines backend-specific type traits. More...

class	backend_traits< backend::opencl >
	Specializes backend traits for the OpenCL backend. More...

class	buffer
	buffer is the public interface for the buffer object implementation. More...

class	buffer< const T, dims, AllocatorT >
	Specialization for const buffers, that allows the creation of buffers on the device from const data. More...

class	compile_program_error
	Represents an error that happened during compilation. More...

class	context
	Interface for abstracting and interacting with an underlying cl_context object. More...

class	cpu_selector
	Implementation of a device_selector that selects a CPU device. More...

class	default_selector
	Implementation of a device_selector that selects either a CPU or a GPU, and falls back to a host mode device if none can be found. More...

class	device
	Interface for abstracting and interacting with an underlying cl_device_id object. More...

class	device_error
	Base SYCL device error group. More...

class	device_event
	Event for asynchronous operations on the device side. More...

class	device_selector
	Abstract class that can be implemented to tell the runtime how to perform device selection. More...

struct	device_type

struct	elem
	Struct with values that help with accessing vector members. More...

class	event
	Abstraction of a cl_event object. More...

class	event_error
	Represents an error related to a cl::sycl::event. More...

class	exception
	SYCL exception class, defined Section 3.2 of the specification, for general SYCL error. More...

class	exception_list
	List of exceptions thrown asynchronously, contains objects of type exception_ptr_class. More...

class	feature_not_supported
	Represents an exception when an optional feature or extension is used in a kernel but its not available on the device the SYCL kernel is being enqueued on. More...

class	gpu_selector
	Implementation of a device_selector that selects a GPU device. More...

class	group
	The cl::sycl::group object is a container for all information about a work group. More...

class	h_item
	Identifies an instance of a parallel_for_work_item function object executing at each point in a local range passed to a parallel_for_work_item call. More...

class	handler
	A handler gives user access to command group scope functionality, such as API calls. More...

class	host_selector
	Implementation of a host_selector that selects the host device. More...

class	id
	dims-dimensional index More...

class	image

struct	info_convert< cl_context *, context >
	This function converts a cl context object to a sycl context object it is used to allow get_info<cl::sycl::typename::context> to return a sycl context object. More...

struct	info_convert< cl_device_id *, device >
	This function converts a cl device object to a sycl device object it is used to allow get_info<cl::sycl::device::parent_device> to return a sycl device object. More...

struct	info_convert< cl_platform_id *, platform >
	This function converts a cl platform object to a sycl platform object it is used to allow get_info<cl::sycl::device::platform> to return a sycl platform object. More...

struct	info_convert< size_t *, range< 3 > >
	Specialization of info_convert for converting a pointer size_t type to a range<3> type. More...

class	intel_selector
	Implementation of a device_selector that selects an Intel platform. More...

class	invalid_object_error
	Represents an error regarding any memory object being used by a kernel. More...

class	invalid_parameter_error
	Represents an error related to SYCL kernel parameters. More...

class	item
	The cl::sycl::item object is a container for all information about a work item. More...

class	kernel
	SYCL Kernel interface. More...

class	kernel_error
	Represents an error that occurred before or while enqueuing a SYCL kernel. More...

class	link_program_error
	Represents an error that happened during linking. More...

class	memory_allocation_error
	Represents a memory allocation error. More...

class	multi_ptr
	multi_ptr, generic pointer class. More...

class	multi_ptr< const void, asp >
	Generic pointer class, specialization for const void. More...

class	multi_ptr< void, asp >
	Generic pointer class, specialization for void. More...

class	nd_item
	The cl::sycl::nd_item object is a container for all information about a work-item. More...

class	nd_range
	Implements the nd_range class of the SYCL specification. More...

class	nd_range_error
	Represents an error related to a provided nd_range. More...

class	opencl_selector
	Implementation of an opencl_selector that selects either a CPU or a GPU. More...

class	platform
	Interface for abstracting and interacting with an underlying cl_platform_id object. More...

class	platform_error
	Represents a platform related error. More...

class	precision_manipulator
	Class that encapsulates a precision stream manipulator. More...

class	private_memory
	This class allows private memory allocation inside a handler::parallel_for_work_group. More...

class	profiling_error
	Represents an issue related to profiling (can only be raised if profiling is enabled). More...

class	program
	Public facing program class that provides an interface for abstracting the construction and building of a cl_program object, See section 3.5.5 of the SYCL 1.2 specification. More...

class	property_list
	Storage class for different properties. More...

class	queue
	The cl::sycl::queue object is the SYCL abstraction of the OpenCL object cl_command_queue. More...

class	range
	dims-dimensional range More...

class	runtime_error
	Base SYCL runtime error group. More...

class	sampler
	Public sampler class. More...

class	stream
	Class that encapsulates a buffer that can be passed to the device and streamed to. More...

struct	stream_vec
	Template struct that contains a function call operator that streams a vec object to a stream object. More...

struct	stream_vec< 16, true, elementT, kDimensions >
	Specialization of stream_vec for kVecPart = 16 and kEnabled = true. More...

struct	stream_vec< 2, true, elementT, kDimensions >
	Specialization of stream_vec for kVecPart = 2 and kEnabled = true. More...

struct	stream_vec< 3, true, elementT, kDimensions >
	Specialization of stream_vec for kVecPart = 3 and kEnabled = true. More...

struct	stream_vec< 4, true, elementT, kDimensions >
	Specialization of stream_vec for kVecPart = 4 and kEnabled = true. More...

struct	stream_vec< 8, true, elementT, kDimensions >
	Specialization of stream_vec for kVecPart = 8 and kEnabled = true. More...

class	swizzled_vec

class	vec

Typedefs
using	base_t = detail::accessor_buffer_interface< elemT, kDims, kMode, kTarget, access::placeholder::true_t >

using	atomic_int = atomic< cl_int >
	Atomic int object with default global address space. More...

using	atomic_uint = atomic< cl_uint >
	Atomic unsigned int object with default global address space. More...

using	atomic_float = atomic< cl_float >
	Atomic float object with default global address space. More...

template<backend Backend, typename SyclType >
using	backend_input_t = typename backend_traits< Backend >::template input_type< SyclType >
	Shorthand for retrieving the input_type for a specific backend. More...

template<backend Backend, typename SyclType >
using	backend_return_t = typename backend_traits< Backend >::template return_type< SyclType >
	Shorthand for retrieving the return_type for a specific backend. More...

using	cl_bool = bool
	Boolean interop type. More...

using	cl_char = signed char
	char (1 byte) interop type More...

using	cl_uchar = unsigned char
	unsigned char (1 byte) interop type More...

using	cl_short = int16_t
	short (2 bytes) interop type More...

using	cl_ushort = uint16_t
	unsigned short (2 bytes) interop type More...

using	cl_int = int
	int (4 bytes) interop type More...

using	cl_uint = unsigned int
	unsigned int (4 bytes) interop type More...

using	cl_float = float
	float interop type More...

using	cl_double = double
	double interop type More...

using	cl_long = int64_t
	long (8 bytes) interop type More...

using	cl_ulong = uint64_t
	unsigned long (8 bytes) interop type More...

using	cl_half = half
	half scalar type More...

using	cl_char2 = vec< cl_char, 2 >
	char vector interop type More...

using	cl_char3 = vec< cl_char, 3 >
	char vector interop type More...

using	cl_char4 = vec< cl_char, 4 >
	char vector interop type More...

using	cl_char8 = vec< cl_char, 8 >
	char vector interop type More...

using	cl_char16 = vec< cl_char, 16 >
	char vector interop type More...

using	cl_uchar2 = vec< cl_uchar, 2 >
	unsigned char vector interop type More...

using	cl_uchar3 = vec< cl_uchar, 3 >
	unsigned char vector interop type More...

using	cl_uchar4 = vec< cl_uchar, 4 >
	unsigned char vector interop type More...

using	cl_uchar8 = vec< cl_uchar, 8 >
	unsigned char vector interop type More...

using	cl_uchar16 = vec< cl_uchar, 16 >
	unsigned char vector interop type More...

using	cl_short2 = vec< cl_short, 2 >
	short vector interop type More...

using	cl_short3 = vec< cl_short, 3 >
	short vector interop type More...

using	cl_short4 = vec< cl_short, 4 >
	short vector interop type More...

using	cl_short8 = vec< cl_short, 8 >
	short vector interop type More...

using	cl_short16 = vec< cl_short, 16 >
	short vector interop type More...

using	cl_ushort2 = vec< cl_ushort, 2 >
	unsigned short vector interop type More...

using	cl_ushort3 = vec< cl_ushort, 3 >
	unsigned short vector interop type More...

using	cl_ushort4 = vec< cl_ushort, 4 >
	unsigned short vector interop type More...

using	cl_ushort8 = vec< cl_ushort, 8 >
	unsigned short vector interop type More...

using	cl_ushort16 = vec< cl_ushort, 16 >
	unsigned short vector interop type More...

using	cl_int2 = vec< cl_int, 2 >
	int vector interop type More...

using	cl_int3 = vec< cl_int, 3 >
	int vector interop type More...

using	cl_int4 = vec< cl_int, 4 >
	int vector interop type More...

using	cl_int8 = vec< cl_int, 8 >
	int vector interop type More...

using	cl_int16 = vec< cl_int, 16 >
	int vector interop type More...

using	cl_uint2 = vec< cl_uint, 2 >
	unsigned int vector interop type More...

using	cl_uint3 = vec< cl_uint, 3 >
	unsigned int vector interop type More...

using	cl_uint4 = vec< cl_uint, 4 >
	unsigned int vector interop type More...

using	cl_uint8 = vec< cl_uint, 8 >
	unsigned int vector interop type More...

using	cl_uint16 = vec< cl_uint, 16 >
	unsigned int vector interop type More...

using	cl_long2 = vec< cl_long, 2 >
	long vector interop type More...

using	cl_long3 = vec< cl_long, 3 >
	long vector interop type More...

using	cl_long4 = vec< cl_long, 4 >
	long vector interop type More...

using	cl_long8 = vec< cl_long, 8 >
	long vector interop type More...

using	cl_long16 = vec< cl_long, 16 >
	long vector interop type More...

using	cl_ulong2 = vec< cl_ulong, 2 >
	unsigned long vector interop type More...

using	cl_ulong3 = vec< cl_ulong, 3 >
	unsigned long vector interop type More...

using	cl_ulong4 = vec< cl_ulong, 4 >
	unsigned long vector interop type More...

using	cl_ulong8 = vec< cl_ulong, 8 >
	unsigned long vector interop type More...

using	cl_ulong16 = vec< cl_ulong, 16 >
	unsigned long vector interop type More...

using	cl_float2 = vec< cl_float, 2 >
	float vector interop type More...

using	cl_float3 = vec< cl_float, 3 >
	float vector interop type More...

using	cl_float4 = vec< cl_float, 4 >
	float vector interop type More...

using	cl_float8 = vec< cl_float, 8 >
	float vector interop type More...

using	cl_float16 = vec< cl_float, 16 >
	float vector interop type More...

using	cl_double2 = vec< cl_double, 2 >
	double vector interop type More...

using	cl_double3 = vec< cl_double, 3 >
	double vector interop type More...

using	cl_double4 = vec< cl_double, 4 >
	double vector interop type More...

using	cl_double8 = vec< cl_double, 8 >
	double vector interop type More...

using	cl_double16 = vec< cl_double, 16 >
	double vector interop type More...

using	cl_half2 = vec< cl_half, 2 >
	2 dimensional half vector interop type More...

using	cl_half3 = vec< cl_half, 3 >
	3 dimensional half vector interop type More...

using	cl_half4 = vec< cl_half, 4 >
	4 dimensional half vector interop type More...

using	cl_half8 = vec< cl_half, 8 >
	8 dimensional half vector interop type More...

using	cl_half16 = vec< cl_half, 16 >
	16 dimensional half vector interop type More...

template<typename T , typename Alloc = std::allocator<T>>
using	vector_class = std::vector< T, Alloc >

using	string_class = std::string

template<typename T >
using	function_class = std::function< T >

using	mutex_class = std::mutex

template<typename T , class D = std::default_delete<T>>
using	unique_ptr_class = std::unique_ptr< T, D >

template<typename T >
using	shared_ptr_class = std::shared_ptr< T >

template<typename T >
using	weak_ptr_class = std::weak_ptr< T >

template<typename T >
using	hash_class = std::hash< T >

template<size_t Size>
using	bitset_class = std::bitset< Size >

using	byte = unsigned char

using	exception_ptr_class = std::exception_ptr
	Class used to store exception objects and transfer them across thread, equivalent to std::exception_ptr. More...

using	async_handler = cl::sycl::function_class< void(exception_list)>
	async_handler type definition. More...

template<typename dataType >
using	global_ptr = multi_ptr< dataType, access::address_space::global_space >
	global_ptr pointer class definition for data pointing to the OpenCL global address space. More...

template<typename dataType >
using	local_ptr = multi_ptr< dataType, access::address_space::local_space >
	local_ptr pointer class definition for data pointing to the OpenCL local address space More...

template<typename dataType >
using	private_ptr = multi_ptr< dataType, access::address_space::private_space >
	private_ptr, pointer class definition for data pointing to the OpenCL private address space More...

template<typename dataType >
using	constant_ptr = multi_ptr< dataType, access::address_space::constant_space >
	constant_ptr, pointer class definition for data living in the OpenCL constant address space More...

using	float2 = vec< float, 2 >

using	float3 = vec< float, 3 >

using	float4 = vec< float, 4 >

using	float8 = vec< float, 8 >

using	float16 = vec< float, 16 >

using	double2 = vec< double, 2 >

using	double3 = vec< double, 3 >

using	double4 = vec< double, 4 >

using	double8 = vec< double, 8 >

using	double16 = vec< double, 16 >

using	half2 = vec< half, 2 >

using	half3 = vec< half, 3 >

using	half4 = vec< half, 4 >

using	half8 = vec< half, 8 >

using	half16 = vec< half, 16 >

using	char2 = vec< char, 2 >

using	char3 = vec< char, 3 >

using	char4 = vec< char, 4 >

using	char8 = vec< char, 8 >

using	char16 = vec< char, 16 >

using	schar2 = vec< signed char, 2 >

using	schar3 = vec< signed char, 3 >

using	schar4 = vec< signed char, 4 >

using	schar8 = vec< signed char, 8 >

using	schar16 = vec< signed char, 16 >

using	uchar2 = vec< unsigned char, 2 >

using	uchar3 = vec< unsigned char, 3 >

using	uchar4 = vec< unsigned char, 4 >

using	uchar8 = vec< unsigned char, 8 >

using	uchar16 = vec< unsigned char, 16 >

using	short2 = vec< short, 2 >

using	short3 = vec< short, 3 >

using	short4 = vec< short, 4 >

using	short8 = vec< short, 8 >

using	short16 = vec< short, 16 >

using	ushort2 = vec< unsigned short, 2 >

using	ushort3 = vec< unsigned short, 3 >

using	ushort4 = vec< unsigned short, 4 >

using	ushort8 = vec< unsigned short, 8 >

using	ushort16 = vec< unsigned short, 16 >

using	int2 = vec< int, 2 >

using	int3 = vec< int, 3 >

using	int4 = vec< int, 4 >

using	int8 = vec< int, 8 >

using	int16 = vec< int, 16 >

using	uint2 = vec< unsigned int, 2 >

using	uint3 = vec< unsigned int, 3 >

using	uint4 = vec< unsigned int, 4 >

using	uint8 = vec< unsigned int, 8 >

using	uint16 = vec< unsigned int, 16 >

using	long2 = vec< long, 2 >

using	long3 = vec< long, 3 >

using	long4 = vec< long, 4 >

using	long8 = vec< long, 8 >

using	long16 = vec< long, 16 >

using	ulong2 = vec< unsigned long, 2 >

using	ulong3 = vec< unsigned long, 3 >

using	ulong4 = vec< unsigned long, 4 >

using	ulong8 = vec< unsigned long, 8 >

using	ulong16 = vec< unsigned long, 16 >

using	longlong2 = vec< long long, 2 >

using	longlong3 = vec< long long, 3 >

using	longlong4 = vec< long long, 4 >

using	longlong8 = vec< long long, 8 >

using	longlong16 = vec< long long, 16 >

using	ulonglong2 = vec< unsigned long long, 2 >

using	ulonglong3 = vec< unsigned long long, 3 >

using	ulonglong4 = vec< unsigned long long, 4 >

using	ulonglong8 = vec< unsigned long long, 8 >

using	ulonglong16 = vec< unsigned long long, 16 >

Enumerations
enum	aspect_impl { aspect_impl::COMPUTECPP_DEPRECATED_BY_SYCL_202001 =("Use host_debuggable instead"), aspect_impl::cpu, aspect_impl::gpu, aspect_impl::accelerator, aspect_impl::custom, aspect_impl::fp16, aspect_impl::fp64, aspect_impl::COMPUTECPP_DEPRECATED_BY_SYCL_202001, aspect_impl::COMPUTECPP_DEPRECATED_BY_SYCL_202001, aspect_impl::image, aspect_impl::online_compiler, aspect_impl::online_linker, aspect_impl::queue_profiling, aspect_impl::usm_device_allocations, aspect_impl::usm_host_allocations, aspect_impl::usm_shared_allocations, aspect_impl::usm_restricted_shared_allocations, aspect_impl::COMPUTECPP_DEPRECATED_BY_SYCL_202001, aspect_impl::emulated, aspect_impl::host_debuggable, aspect_impl::atomic64, aspect_impl::usm_atomic_host_allocations, aspect_impl::usm_atomic_shared_allocations, aspect_impl::usm_system_allocations }
	Enumerates the aspects which can be queried on a platform or device. More...

enum	memory_order : int { memory_order::relaxed = static_cast<int>(std::memory_order_relaxed) }
	Only relaxed memory order is supported in SYCL 1.2.1. More...

enum	backend { backend::host, backend::opencl }
	Defines backends available in ComputeCpp. More...

enum	bundle_state { bundle_state::input, bundle_state::object, bundle_state::executable }
	The compilation state of a kernel bundle. More...

enum	image_channel_order : unsigned int { image_channel_order::r = CL_R, image_channel_order::a = CL_A, image_channel_order::rg = CL_RG, image_channel_order::ra = CL_RA, image_channel_order::rgb = CL_RGB, image_channel_order::rgba = CL_RGBA, image_channel_order::bgra = CL_BGRA, image_channel_order::argb = CL_ARGB, image_channel_order::intensity = CL_INTENSITY, image_channel_order::luminance = CL_LUMINANCE, image_channel_order::rx = CL_Rx, image_channel_order::rgx = CL_RGx, image_channel_order::rgbx = CL_RGBx, image_channel_order::abgr = 0x10C3 }
	Specify the number of channels and the channel layout in which channels are stored in the image. More...

enum	image_channel_type : unsigned int { image_channel_type::snorm_int8 = CL_SNORM_INT8, image_channel_type::snorm_int16 = CL_SNORM_INT16, image_channel_type::unorm_int8 = CL_UNORM_INT8, image_channel_type::unorm_int16 = CL_UNORM_INT16, image_channel_type::unorm_short_565 = CL_UNORM_SHORT_565, image_channel_type::unorm_short_555 = CL_UNORM_SHORT_555, image_channel_type::unorm_int_101010 = CL_UNORM_INT_101010, image_channel_type::signed_int8 = CL_SIGNED_INT8, image_channel_type::signed_int16 = CL_SIGNED_INT16, image_channel_type::signed_int32 = CL_SIGNED_INT32, image_channel_type::unsigned_int8 = CL_UNSIGNED_INT8, image_channel_type::unsigned_int16 = CL_UNSIGNED_INT16, image_channel_type::unsigned_int32 = CL_UNSIGNED_INT32, image_channel_type::fp16 = CL_HALF_FLOAT, image_channel_type::fp32 = CL_FLOAT }
	Specify the size of the channels data type. More...

enum	program_state { program_state::none, program_state::compiled, program_state::linked }
	Enum describing the build state of the program. More...

enum	addressing_mode : unsigned int { addressing_mode::none = 4400, addressing_mode::clamp_to_edge = 4401, addressing_mode::clamp = 4402, addressing_mode::repeat = 4403, addressing_mode::mirrored_repeat = 4404 }
	Enum class for specifying the addressing mode of a sampler. More...

enum	filtering_mode : unsigned int { filtering_mode::nearest = 4416, filtering_mode::linear = 4417 }
	Enum class for specifying the filter mode of a sampler. More...

enum	coordinate_normalization_mode : unsigned int { coordinate_normalization_mode::normalized, coordinate_normalization_mode::unnormalized }
	Filtering mode description. More...

enum	write_back : bool { write_back::disable_write_back = false, write_back::enable_write_back = true }

enum	stream_manipulator { stream_manipulator::endl = 0, stream_manipulator::scientific = 1, stream_manipulator::hex = 2, stream_manipulator::oct = 3, stream_manipulator::showbase = 4, stream_manipulator::showpos = 5, stream_manipulator::dec = 6, stream_manipulator::noshowbase = 7, stream_manipulator::noshowpos = 8, stream_manipulator::fixed = 9, stream_manipulator::hexfloat = 10, stream_manipulator::defaultfloat = 11, stream_manipulator::flush = 12 }
	Enum class that enumerates the different stream manipulators. More...

enum	rounding_mode { rounding_mode::automatic, rounding_mode::rte, rounding_mode::rtz, rounding_mode::rtp, rounding_mode::rtn }
	Available vector rounding modes. More...

Functions
template<typename elemT , int kDims, access::mode kMode, access::target kTarget, access::placeholder IsPlaceholder>
COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT	COMPUTECPP_VALID_KERNEL_ARG_IF (kTarget !=access::target::host_buffer &&kTarget !=access::target::host_task, "Cannot pass host accessor to SYCL kernel") accessor
	Non-specialized accessor template that covers only global and constant buffers. More...

template<typename elemT , int kDims, access::mode kMode, access::target kTarget>
COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT	COMPUTECPP_VALID_KERNEL_ARG_IF (kTarget !=access::target::host_buffer &&kTarget !=access::target::host_task, "Cannot pass host accessor to SYCL kernel") accessor< elemT
	A public facing accessor that can be constructed outside of a command group. More...

	accessor (storage_mem &store, detail::access_range accessRange, const property_list &propList={})
	Constructs a ranged placeholder accessor. More...

accessor< elemT, 1, kMode, kTarget, access::placeholder::true_t >	get_accessor_with_added_offset (int addedOffset) const
	returns an accessor similar to *this, with the offset changed to this->get_offset()[0] + addedOffset. More...

	accessor (const property_list &propList={})
	Constructs a default placeholder accessor without associated storage. More...

template<typename AllocatorT = default_allocator>
	accessor (buffer< elemT, detail::acc_interface_dims< kDims >::value, AllocatorT > &bufferRef, const property_list &propList={})
	Constructs a placeholder accessor. More...

template<typename AllocatorT >
	COMPUTECPP_DEPRECATED_BY_SYCL_VER (201703, "Use overload where the range comes before the offset.") accessor(buffer< elemT
	Constructs a ranged placeholder accessor. More...

template<typename AllocatorT >
	accessor (buffer< elemT, kDims, AllocatorT > &bufferRef, range< kDims > accessRange, const property_list &propList={})
	Constructs a ranged placeholder accessor. More...

template<typename AllocatorT >
	accessor (buffer< elemT, kDims, AllocatorT > &bufferRef, range< kDims > accessRange, id< kDims > accessOffset, const property_list &propList={})
	Constructs a ranged placeholder accessor. More...

	accessor (buffer< elemT, kDims > &bufferRef, range< kDims > accessRange, const property_list &propList={})
	Constructs a ranged placeholder accessor. More...

	accessor (buffer< elemT, kDims > &bufferRef, range< kDims > accessRange, id< kDims > accessOffset, const property_list &propList={})
	Constructs a ranged placeholder accessor. More...

template<typename AllocatorT >
	accessor (buffer< elemT, kDims, AllocatorT > &bufferRef, handler &commandHandler, range< kDims > accessRange, const property_list &propList={})
	Constructs a ranged placeholder accessor. More...

template<typename AllocatorT >
	accessor (buffer< elemT, kDims, AllocatorT > &bufferRef, handler &commandHandler, range< kDims > accessRange, id< kDims > accessOffset, const property_list &propList={})
	Constructs a ranged placeholder accessor. More...

	accessor (buffer< elemT, kDims > &bufferRef, handler &commandHandler, range< kDims > accessRange, const property_list &propList={})
	Constructs a ranged placeholder accessor. More...

	accessor (buffer< elemT, kDims > &bufferRef, handler &commandHandler, range< kDims > accessRange, id< kDims > accessOffset, const property_list &propList={})
	Constructs a ranged placeholder accessor. More...

template<typename AllocatorT >
	accessor (buffer< elemT, kDims, AllocatorT > &bufferRef, handler &commandHandler, const property_list &propList={})
	Constructs a placeholder buffer accessor. More...

	accessor (buffer< elemT, kDims > &bufferRef, handler &commandHandler, const property_list &propList={})
	Constructs a placeholder buffer accessor. More...

template<typename AllocatorT , COMPUTECPP_ENABLE_IF(elemT,(kDims==0)) >
	accessor (buffer< elemT, 1, AllocatorT > &bufferRef, handler &commandHandler, const property_list &propList={})
	Constructs a placeholder 0-dimensional buffer accessor. More...

accessor< elemT, kDims, kMode, kTarget, access::placeholder::false_t >	get_access (handler &commandHandler) const
	Obtains a normal accessor from the placeholder accessor. More...

template<COMPUTECPP_ENABLE_IF(elemT,(kDims==1)) >
accessor< elemT, 1, kMode, kTarget, access::placeholder::true_t >	operator+ (int rhs) const
	Creates and returns a new accessor with its offset changed by rhs. More...

template<COMPUTECPP_ENABLE_IF(elemT,(kDims==1)) >
accessor< elemT, 1, kMode, kTarget, access::placeholder::true_t >	operator- (int rhs) const
	Creates and returns a new accessor with its offset changed by rhs. More...

template<COMPUTECPP_ENABLE_IF(elemT,(kDims==1)) >
accessor< elemT, 1, kMode, kTarget, access::placeholder::true_t > &	operator+= (int rhs)
	Changes this->offset by rhs as this->get_offset()[0] + rhs. More...

template<COMPUTECPP_ENABLE_IF(elemT,(kDims==1)) >
accessor< elemT, 1, kMode, kTarget, access::placeholder::true_t > &	operator-= (int rhs)
	Changes this->offset by rhs as this->get_offset()[0] - rhs. More...

template<typename functorT , int dimensions>
	COMPUTECPP_DEPRECATED_BY_SYCL_VER (201703, "Use group::parallel_for_work_item instead.") void parallel_for_work_item(group< dimensions > groupID
	inner loop of parallel_for_work_group, the Hierarchical API. More...

template<typename T , access::address_space addressSpace>
T	atomic_load (atomic< T, addressSpace > object, memory_order mem_order=memory_order::relaxed)
	Global function atomic_load. More...

template<typename T , access::address_space addressSpace>
void	atomic_store (atomic< T, addressSpace > object, T operand, memory_order mem_order=memory_order::relaxed)
	Global function atomic_store. More...

template<typename T , access::address_space addressSpace>
T	atomic_exchange (atomic< T, addressSpace > object, T operand, memory_order mem_order=memory_order::relaxed)
	Global function atomic_exchange. More...

template<typename T , access::address_space addressSpace>
cl_bool	atomic_compare_exchange_strong (atomic< T, addressSpace > object, T &expected, T desired, memory_order success=memory_order::relaxed, memory_order fail=memory_order::relaxed)
	Global function atomic_compare_exchange. More...

template<typename T , access::address_space addressSpace>
T	atomic_fetch_add (atomic< T, addressSpace > object, T operand, memory_order mem_order=memory_order::relaxed)
	Global function atomic_add. More...

template<typename T , access::address_space addressSpace>
T	atomic_fetch_sub (atomic< T, addressSpace > object, T operand, memory_order mem_order=memory_order::relaxed)
	Global function atomic_sub. More...

template<typename T , access::address_space addressSpace>
T	atomic_fetch_and (atomic< T, addressSpace > object, T operand, memory_order mem_order=memory_order::relaxed)
	Global function atomic_and. More...

template<typename T , access::address_space addressSpace>
T	atomic_fetch_or (atomic< T, addressSpace > object, T operand, memory_order mem_order=memory_order::relaxed)
	Global function atomic_or. More...

template<typename T , access::address_space addressSpace>
T	atomic_fetch_xor (atomic< T, addressSpace > object, T operand, memory_order mem_order=memory_order::relaxed)
	Global function atomic_xor. More...

template<typename T , access::address_space addressSpace>
T	atomic_fetch_min (atomic< T, addressSpace > object, T operand, memory_order mem_order=memory_order::relaxed)
	Global function atomic_min. More...

template<typename T , access::address_space addressSpace>
T	atomic_fetch_max (atomic< T, addressSpace > object, T operand, memory_order mem_order=memory_order::relaxed)
	Global function atomic_max. More...

template<typename F1 , typename F2 , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && std::is_same< F1, F2 >::value)\|\|(detail::builtin::is_genfloath< F1 >::value && std::is_same< F2, half >::value)\|\|(detail::builtin::is_genfloatf< F1 >::value && std::is_same< F2, float >::value)\|\|(detail::builtin::is_genfloatd< F1 >::value && std::is_same< F2, double >::value)) >
F1	clamp (F1 x, F2 minval, F2 maxval) noexcept
	Returns `fmin(fmax(x, minval), maxval)`. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	degrees (const F x) noexcept
	Converts radians to degrees, i.e. More...

template<typename F1 , typename F2 , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && std::is_same< F1, F2 >::value)\|\|(detail::builtin::is_genfloath< F1 >::value && std::is_same< F2, half >::value)\|\|(detail::builtin::is_genfloatf< F1 >::value && std::is_same< F2, float >::value)\|\|(detail::builtin::is_genfloatd< F1 >::value && std::is_same< F2, double >::value)) >
F1	abs (F1 x, F2 y) noexcept
	Compute absolute value of a floating-point number. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)\|\|(detail::builtin::is_genfloath< F1 >::value && detail::builtin::is_sgenfloat< F2, half >::value)\|\|(detail::builtin::is_genfloatf< F1 >::value && detail::builtin::is_sgenfloat< F2, float >::value)\|\|(detail::builtin::is_genfloatd< F1 >::value && detail::builtin::is_sgenfloat< F2, double >::value)) >
F	max (F1 x, F2 y) noexcept
	Returns y if `x < y`, otherwise it returns x. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)\|\|(detail::builtin::is_genfloath< F1 >::value && detail::builtin::is_sgenfloat< F2, half >::value)\|\|(detail::builtin::is_genfloatf< F1 >::value && detail::builtin::is_sgenfloat< F2, float >::value)\|\|(detail::builtin::is_genfloatd< F1 >::value && detail::builtin::is_sgenfloat< F2, double >::value)) >
F	min (F1 x, F2 y) noexcept
	Returns y if `x > y`, otherwise it returns x. More...

template<typename F1 , typename F2 , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && std::is_same< F1, F2 >::value)\|\|(detail::builtin::is_genfloath< F1 >::value && std::is_same< F2, half >::value)\|\|(detail::builtin::is_genfloatf< F1 >::value && std::is_same< F2, float >::value)\|\|(detail::builtin::is_genfloatd< F1 >::value && std::is_same< F2, double >::value)) >
F1	mix (F1 x, F1 y, F2 a) noexcept
	Returns the linear blend of x&y implemented as: x + (y - x) * a. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	radians (const F x) noexcept
	Converts degrees to radians, i.e. More...

template<typename F1 , typename F2 , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && std::is_same< F1, F2 >::value)\|\|(std::is_same< F1, half >::value && detail::builtin::is_genfloath< F2 >::value)\|\|(std::is_same< F1, float >::value && detail::builtin::is_genfloatf< F2 >::value)\|\|(std::is_same< F1, double >::value && detail::builtin::is_genfloatd< F2 >::value)) >
F2	step (F1 edge, F2 x) noexcept
	Returns 0.0 if `x < edge`, otherwise it returns 1.0. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	acos (F x) noexcept
	Inverse cosine function. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	acosh (F x) noexcept
	Inverse hyperbolic cosine. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	acospi (F x) noexcept
	Compute `acos(x)/π` More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	asin (F x) noexcept
	Inverse sine function. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	asinh (F x) noexcept
	Inverse hyperbolic sine. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	asinpi (F x) noexcept
	Compute `asin(x)/π` More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	atan (F y_over_x) noexcept
	Inverse tangent function. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
F	atan2 (F y, F x) noexcept
	Compute `atan(y/x)`. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	atanh (F x) noexcept
	Inverse hyperbolic tangent. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	atanpi (F x) noexcept
	Compute `atan(x)/π` More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
F	atan2pi (F y, F x) noexcept
	Compute `atan(y/x)/π`. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	cbrt (F x) noexcept
	Compute cube-root. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	ceil (F x) noexcept
	Round to integral value using the round to positive infinity rounding mode. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	copysign (F1 x, F2 y) noexcept
	Returns x with its sign changed to match the sign of y. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	cos (F x) noexcept
	Compute cosine. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	cosh (F x) noexcept
	Compute hyperbolic cosine. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	cospi (F x) noexcept
	Compute cos(πx) More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	erfc (F x) noexcept
	Complementary error function. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	erf (F x) noexcept
	Error function encountered in integrating the normal distribution. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	exp (F x) noexcept
	Compute the base-e exponential of x. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	exp2 (F x) noexcept
	Exponential base 2 function. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	exp10 (F x) noexcept
	Exponential base 10 function. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	expm1 (F x) noexcept
	Compute `exp(x) - 1.0`. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	fabs (F x) noexcept
	Compute absolute value of a floating-point number. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	fdim (F1 x, F2 y) noexcept
	`x - y` if `x > y`, +0 if x is less than or equal to y. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	floor (F x) noexcept
	Round to integral value using the round to negative infinity rounding mode. More...

template<typename F1 , typename F2 , typename F3 , typename F = detail::common_return_t<F1, F2, F3>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value && detail::builtin::is_genfloat< F3 >::value)) >
F	fma (F1 a, F2 b, F3 c) noexcept
	Returns the correctly rounded floating-point representation of the sum of c with the infinitely precise product of a and b. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES( detail::builtin::is_genfloat< F1 >::value &&(detail::builtin::is_genfloat< F2 >::value\|\| detail::builtin::is_sgenfloat< F2, detail::scalar_t< F >>::value)) >
F	fmax (F1 x, F2 y) noexcept
	Returns y if `x < y`, otherwise it returns x. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES( detail::builtin::is_genfloat< F1 >::value &&(detail::builtin::is_genfloat< F2 >::value\|\| detail::builtin::is_sgenfloat< F2, detail::scalar_t< F >>::value)) >
F	fmin (F1 x, F2 y) noexcept
	Returns y if y < x, otherwise it returns x. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	fmod (F1 x, F2 y) noexcept
	Modulus. More...

template<typename F , access::address_space AddressSpace, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
F	fract (F x, detail::builtin_ptr< F, AddressSpace > iptr) noexcept
	Returns `fmin(x - floor (x), 0x1.fffffep-1f)`. More...

template<typename F , typename Integral , access::address_space AddressSpace, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >
F	frexp (F x, detail::builtin_ptr< Integral, AddressSpace > exp) noexcept
	Extract mantissa and exponent from x. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	hypot (F1 x, F2 y) noexcept
	Compute the value of the square root of `x2 + y2` without undue overflow or underflow. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::correct_int_t< F >	ilogb (F x) noexcept
	Return the exponent as an integer value. More...

template<typename F , typename Integral , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >
F	ldexp (F x, Integral k) noexcept
	Multiply x by 2 to the power k. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
F	ldexp (F x, const int k) noexcept
	Multiply x by 2 to the power k. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	lgamma (F x) noexcept
	Log gamma function. More...

template<typename F , typename Integral , access::address_space AddressSpace, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >
F	lgamma_r (F x, detail::builtin_ptr< Integral, AddressSpace > signp) noexcept
	Log gamma function. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	log (F x) noexcept
	Compute natural logarithm. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	log2 (F x) noexcept
	Compute a base 2 logarithm. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	log10 (F x) noexcept
	Compute a base 10 logarithm. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	log1p (F x) noexcept
	Compute `loge(1.0 + x)`. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	logb (F x) noexcept
	Compute the exponent of x, which is the integral part of logr (\|x\|). More...

template<typename F1 , typename F2 , typename F3 , typename F = detail::common_return_t<F1, F2, F3>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value && detail::builtin::is_genfloat< F3 >::value)) >
F	mad (F1 a, F2 b, F3 c) noexcept
	mad approximates `a * b + c`. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	maxmag (F1 x, F2 y) noexcept
	Returns x if \|x\| > \|y\|, y if \|y\| > \|x\|, otherwise `fmax(x, y)`. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	minmag (F1 x, F2 y) noexcept
	Returns x if \|x\| < \|y\|, y if \|y\| < \|x\|, otherwise `fmin(x, y)`. More...

template<typename F , access::address_space AddressSpace, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
F	modf (F x, detail::builtin_ptr< F, AddressSpace > iptr) noexcept
	Decompose a floating-point number. More...

template<typename Integral , COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_ugenint< Integral >::value, detail::builtin::is_ugenlonginteger< Integral >::value >::value) >
auto	nan (Integral nancode) noexcept -> detail::matching_float_t< Integral >
	Returns x if \|x\| < \|y\|, y if \|y\| < \|x\|, otherwise `fmin(x, y)`. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_ushortn< Integral >::value) >
detail::matching_float_t< Integral >	nan (Integral nancode) noexcept

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	nextafter (F1 x, F2 y) noexcept
	Computes the next representable single-precision floating-point value following x in the direction of y. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	pow (F1 x, F2 y) noexcept
	Compute x to the power y. More...

template<typename F , typename Integral , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >
F	pown (F x, Integral y) noexcept
	Compute x to the power y, where y is an integer. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	powr (F1 x, F2 y) noexcept
	Compute x to the power y, where x >= 0. More...

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	remainder (F1 x, F2 y) noexcept
	Compute the value r such that `r = x - n * y`, where n is the integer nearest the exact value of x / y. More...

template<typename F , typename Integral , access::address_space AddressSpace, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >
F	remquo (F x, F y, detail::builtin_ptr< Integral, AddressSpace > quo) noexcept
	The remquo function computes the value r such that `r = x - k * y`, where k is the integer nearest the exact value of x/y. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	rint (F x) noexcept
	Round to integral value (using round to nearest even rounding mode) in floating-point format. More...

template<typename F , typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value, detail::builtin::is_genint< Integral >::value) >
F	rootn (F x, Integral y) noexcept
	Compute x to the power 1/y. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	round (F x) noexcept
	Return the integral value nearest to x rounding halfway cases away from zero, regardless of the current rounding direction. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	rsqrt (F x) noexcept
	Compute inverse square root. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	sin (F x) noexcept
	Compute sine. More...

template<typename F , access::address_space AddressSpace, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
F	sincos (F x, detail::builtin_ptr< F, AddressSpace > cosval) noexcept
	Compute sine and cosine of x. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	sinh (F x) noexcept
	Compute hyperbolic sine. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	sinpi (F x) noexcept
	Compute `sin(π * x)`. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	sqrt (F x) noexcept
	Compute square root. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	tan (F x) noexcept
	Compute tangent. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	tanh (F x) noexcept
	Compute hyperbolic tangent. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	tanpi (F x) noexcept
	Compute `tan(π * x)`. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	tgamma (F x) noexcept
	Compute the gamma function. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	trunc (F x) noexcept
	Round to integral value using the round to zero rounding mode. More...

template<typename F , int N, COMPUTECPP_REQUIRES((detail::is_custom_half_type< F >::value\|\| std::is_same< F, float >::value\|\| std::is_same< F, double >::value) &&(N==3\|\|N==4)) >
vec< F, N >	cross (const vec< F, N > p0, const vec< F, N > p1) noexcept
	Returns the cross product of p0.xyz and p1.xyz. More...

template<typename F1 , typename F2 , typename F = detail::select_float_t<detail::common_return_t<F1, F2>>, COMPUTECPP_REQUIRES((detail::builtin::is_gen_geo_anyfloat< F1 >::value && detail::builtin::is_gen_geo_anyfloat< F2 >::value)) >
F	dot (F1 p0, F2 p1) noexcept
	Compute dot product. More...

template<typename F1 , typename F2 , typename F = detail::select_float_t<detail::common_return_t<F1, F2>>, COMPUTECPP_REQUIRES((detail::builtin::is_gen_geo_anyfloat< F1 >::value && detail::builtin::is_gen_geo_anyfloat< F2 >::value)) >
F	distance (F1 p0, F2 p1) noexcept
	Returns the distance between p0 and p1. More...

template<typename F , typename return_t = detail::select_float_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES(detail::builtin::is_gen_geo_anyfloat< F >::value) >
return_t	length (F p) noexcept
	Return the length of vector p, i.e. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_gen_geo_anyfloat< F >::value) >
return_t	normalize (F p) noexcept
	Returns a vector in the same direction as p but with a length of 1. More...

template<typename F1 , typename F2 , typename F = detail::scalar_t<detail::common_return_t<F1, F2>>, COMPUTECPP_REQUIRES((computecpp::gsl::or_< detail::builtin::is_gengeohalf< F1 >::value, detail::builtin::is_gengeofloat< F1 >::value >::value) &&(computecpp::gsl::or_< detail::builtin::is_gengeohalf< F2 >::value, detail::builtin::is_gengeofloat< F2 >::value >::value)) >
F	fast_distance (F1 p0, F2 p1) noexcept
	Returns f ast `length(p0 - p1)`. More...

template<typename F , typename return_t = detail::scalar_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_gengeohalf< F >::value, detail::builtin::is_gengeofloat< F >::value >::value) >
return_t	fast_length (F p) noexcept
	Returns the length of vector p computed as: `sqrt((half)(pow(p.x,2) + pow(p.y,2)+ ...))`. More...

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_gengeohalf< F >::value, detail::builtin::is_gengeofloat< F >::value >::value) >
return_t	fast_normalize (F p) noexcept
	Returns a vector in the same direction as p but with a length of 1. More...

template<typename Integral , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value)) >
auto	abs (Integral x) noexcept -> detail::make_genuint_t< Integral >
	Returns `\|x\|` More...

template<typename Integral , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value)) >
auto	abs_diff (Integral x, Integral y) noexcept -> detail::make_genuint_t< Integral >
	Returns `\|x - y\|` without modulo overflow. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	add_sat (Integral x, Integral y) noexcept
	Returns `x + y` and saturates the result. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	hadd (Integral x, Integral y) noexcept
	Returns `(x + y) >> 1`. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	rhadd (Integral x, Integral y) noexcept
	Returns `(x + y + 1) >> 1`. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	clamp (Integral x, Integral minval, Integral maxval) noexcept
	Returns `min(max(x, minval), maxval)`. More...

template<typename Integral , typename S , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value && detail::builtin::is_sgeninteger< S >::value)) >
Integral	clamp (Integral x, S minval, S maxval) noexcept
	Returns `min(max(x, minval), maxval)`. More...

template<typename Integral , typename return_t = detail::collapse_swizzled_vec_t<Integral>, COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
return_t	clz (Integral x) noexcept
	Returns the number of leading 0-bits in `x`, starting at the most significant bit position. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	mad_hi (Integral a, Integral b, Integral c) noexcept
	Returns `mul_hi(a, b) + c`. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	mad_sat (Integral a, Integral b, Integral c) noexcept
	Returns `a * b + c` and saturates the result. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	max (Integral x, Integral y) noexcept
	Returns y if `x < y`, otherwise it returns x. More...

template<typename Integral , typename S , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value && detail::builtin::is_sgeninteger< S >::value)) >
Integral	max (Integral x, S y) noexcept
	Returns y if `x < y`, otherwise it returns x. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	min (Integral x, Integral y) noexcept
	Returns y if `y < x`, otherwise it returns x. More...

template<typename Integral , typename S , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value && detail::builtin::is_sgeninteger< S >::value)) >
Integral	min (Integral x, S y) noexcept
	Returns y if `y < x`, otherwise it returns x. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	mul_hi (Integral x, Integral y) noexcept
	Computes `x * y` and returns the high half of the product of `x` and `y`. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	rotate (Integral v, Integral i) noexcept
	For each element in v, the bits are shifted left by the number of bits given by the corresponding element in i (subject to usual shift modulo rules described in section 6.3). More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
Integral	sub_sat (Integral x, Integral y) noexcept
	Returns `x - y` and saturates the result. More...

template<typename I1 , typename I2 , COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_geninteger8bit< I1 >::value && detail::builtin::is_ugeninteger8bit< I2 >::value, detail::builtin::is_geninteger16bit< I1 >::value && detail::builtin::is_ugeninteger16bit< I2 >::value, detail::builtin::is_geninteger32bit< I1 >::value && detail::builtin::is_ugeninteger32bit< I2 >::value >::value) >
auto	upsample (I1 hi, I2 lo) noexcept -> decltype(::cl::sycl::detail::double_width_cast(std::declval< I1 >()))
	Returns `result[i] = (hi[i] << 8) \| lo[i]` More...

template<typename Integral , typename return_t = detail::collapse_swizzled_vec_t<Integral>, COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >
return_t	popcount (Integral x) noexcept
	Returns the number of non-zero bits in `x`. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger32bit< Integral >::value) >
Integral	mad24 (Integral x, Integral y, Integral z) noexcept
	Multipy two 24-bit integer values `x` and `y` and add the 32-bit integer result to the 32-bit integer `z`. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger32bit< Integral >::value) >
Integral	mul24 (Integral x, Integral y) noexcept
	Multiply two 24-bit integer values `x` and `y`. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	isequal (F x, F y) noexcept
	Returns the component-wise compare of `x == y`. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	isnotequal (F x, F y) noexcept
	Returns the component-wise compare of `x != y`. More...

template<typename F1 , typename F2 , typename F = detail::matching_integral_t<detail::common_return_t<F1, F2>>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >
F	isgreater (F1 x, F2 y) noexcept
	Returns the component-wise compare of `x > y`. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	isgreaterequal (F x, F y) noexcept
	Returns the component-wise compare of `x >= y`. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	isless (F x, F y) noexcept
	Returns the component-wise compare of `x < y`. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	islessequal (F x, F y) noexcept
	Returns the component-wise compare of `x <= y`. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	islessgreater (F x, F y) noexcept
	Returns the component-wise compare of `(x < y) \|\| (x > y)`. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	isfinite (F x) noexcept
	Test for finite value. More...

template<typename F , typename return_t = detail::matching_integral_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	isinf (F x) noexcept
	Test for infinity value (positive or negative). More...

template<typename F , typename return_t = detail::matching_integral_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	isnan (F x) noexcept
	Test for a NaN. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	isnormal (F x) noexcept
	Test for a normal value. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	isordered (F x, F y) noexcept
	Test if arguments are ordered. More...

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
detail::matching_integral_t< F >	isunordered (F x, F y) noexcept
	Test if arguments are unordered. More...

template<typename F , typename return_t = detail::matching_integral_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >
return_t	signbit (F x) noexcept
	Test for sign bit. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_igeninteger< Integral >::value) >
int	any (Integral x) noexcept
	Returns 1 if the most significant bit in any component of x is set; otherwise returns 0. More...

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_igeninteger< Integral >::value) >
int	all (Integral x) noexcept
	Returns 1 if the most significant bit in all components of x is set; otherwise returns 0. More...

template<typename T , COMPUTECPP_REQUIRES(detail::builtin::is_gentype< T >::value) >
T	bitselect (T a, T b, T c) noexcept
	Each bit of the result is the corresponding bit of a if the corresponding bit of c is 0. More...

template<typename T1 , typename T2 , COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_geninteger< T1 >::value && detail::builtin::is_igeninteger< T2 >::value, detail::builtin::is_geninteger< T1 >::value && detail::builtin::is_ugeninteger< T2 >::value, detail::builtin::is_genfloatf< T1 >::value && detail::builtin::is_genint< T2 >::value, detail::builtin::is_genfloatf< T1 >::value && detail::builtin::is_ugenint< T2 >::value, detail::builtin::is_genfloatd< T1 >::value && detail::builtin::is_igeninteger64bit< T2 >::value, detail::builtin::is_genfloatd< T1 >::value && detail::builtin::is_ugeninteger64bit< T2 >::value >::value) >
T1	select (T1 a, T1 b, T2 c) noexcept
	For each component of a vector type: `result[i] = (MSB of c[i] is set)? b[i] : a[i]` For a scalar type: `result = c ? b : a`. More...

	COMPUTECPP_GET_INFO_SPECIALIZATION_DECL (device, usm_restricted_shared_allocations) COMPUTECPP_GET_INFO_SPECIALIZATION_DECL(device

template<>
sub_group_independent_forward_progress info::param_traits< info::device, info::device::name >::return_type	device::get_info< info::device::name > () const
	Gets OpenCL information for the underlying cl_device_id. More...

template<>
info::param_traits< info::device, info::device::vendor >::return_type	device::get_info< info::device::vendor > () const
	Gets OpenCL information for the underlying cl_device_id. More...

template<>
info::param_traits< info::device, info::device::version >::return_type	device::get_info< info::device::version > () const
	Gets OpenCL information for the underlying cl_device_id. More...

template<>
COMPUTECPP_EXPORT vector_class< device >	device::create_sub_devices< info::partition_property::partition_equally > (size_t nbSubDev) const
	Template specialisation of create_sub_devices() for info::partition_property::partition_equally. More...

template<>
COMPUTECPP_EXPORT vector_class< device >	device::create_sub_devices< info::partition_property::partition_by_counts > (const vector_class< size_t > &counts) const
	Template specialisation of create_sub_devices() for info::partition_property::partition_by_counts. More...

template<>
COMPUTECPP_EXPORT vector_class< device >	device::create_sub_devices< info::partition_property::partition_by_affinity_domain > (info::partition_affinity_domain affinityDomain) const
	Template specialisation of create_sub_devices() for info::partition_property::partition_by_affinity_domain. More...

template<typename dataType , cl::sycl::access::address_space Space>
std::enable_if< !std::is_same< typename multi_ptr< dataType, Space >::ptr_t, dataType * >::value, multi_ptr< dataType, Space > >::type	make_ptr (dataType *ptr)
	Create a multi_ptr object from a raw pointer. More...

template<typename dataType , cl::sycl::access::address_space Space>
multi_ptr< dataType, Space >	make_ptr (typename multi_ptr< dataType, Space >::ptr_t ptr)

template<>
info::param_traits< info::platform, info::platform::name >::return_type	platform::get_info< info::platform::name > () const
	Get OpenCL information for the underlying cl_platform_id. More...

template<>
info::param_traits< info::platform, info::platform::vendor >::return_type	platform::get_info< info::platform::vendor > () const
	Get OpenCL information for the underlying cl_platform_id. More...

precision_manipulator	setprecision (int precision)
	Inline function that constructs a precision_manipulator object, used in conjunction with stream operators. More...

detail::width_manipulator	setw (const int width)
	Creates a stream manipulator for setting the stream width. More...

COMPUTECPP_EXPORT std::ostream &	operator<< (std::ostream &out, const stream &rhs)
	Stream operator that streams a stream object to an ostream. More...

template<typename valueT >
void	append_hexidecimal (const stream &os, const valueT &value)
	Template function that appends a hexadecimal value to a stream object. More...

template<typename valueT >
int	append_integral (const stream &os, const valueT &value, bool useManipulators=true)
	Template function that appends a integral value to a stream object. More...

template<typename valueT >
void	append_floating_point (const stream &os, const valueT &value)
	Template function that appends a floating point value to a stream object. More...

void	append_string (const stream &os, const char *value)
	Inline function that appends a string value to a stream object. More...

void	append_type_string (const stream &os, char *typeStr, int numChars, int dimensions=0)
	Appends a string representation of a type to the stream. More...

void	append_char (const stream &os, const char &value)
	Inline function that appends a char value to the a object. More...

void	append_bool (const stream &os, const bool &value)
	Inline function that appends a bool value to the a object. More...

void	append_index_array (const stream &os, const detail::index_array &value, int numElements)
	Appends an index_array object to the stream. More...

template<typename pointerT >
const stream &	stream_value (const stream &os, pointerT *rhs)
	Template function that streams a pointer value to a stream object. More...

template<typename dataType , cl::sycl::access::address_space asp>
const stream &	stream_value (const stream &os, const multi_ptr< dataType, asp > &rhs)
	Streams a multi_ptr object into a stream. More...

template<class dataType , COMPUTECPP_ENABLE_IF(dataType,(std::is_integral< dataType >::value)) >
const stream &	stream_value (const stream &os, dataType rhs)
	Streams an integral type into a stream. More...

const stream &	stream_value (const stream &os, const half &rhs)
	Streams a half value to a stream object. More...

const stream &	stream_value (const stream &os, const float &rhs)
	Inline function that streams a float value to a stream object. More...

const stream &	stream_value (const stream &os, const double &rhs)
	Inline function that streams a double value to a stream object. More...

const stream &	stream_value (const stream &os, const bool &rhs)
	Inline function that streams a bool value to a stream object. More...

const stream &	stream_value (const stream &os, const char &rhs)
	Inline function that streams a const char & value to a stream object. More...

const stream &	stream_value (const stream &os, const char *rhs)
	Inline function that streams a const char * value to a stream object. More...

const stream &	stream_value (const stream &os, const stream_manipulator &rhs)
	Inline function that streams a manipulator to a stream object, applying the appropriate action. More...

const stream &	stream_value (const stream &os, const precision_manipulator &rhs)
	Inline function that streams a precision manipulator to a stream object, applying the precision value. More...

const stream &	stream_value (const stream &os, const detail::width_manipulator &rhs)
	Sets the width used for streaming data to the value specified by rhs. More...

template<int dimensions>
const stream &	stream_value (const stream &os, const id< dimensions > &rhs, char *typeStr=nullptr)
	Streams an id object into a stream object. More...

template<int dimensions>
const stream &	stream_value (const stream &os, const range< dimensions > &rhs, char *typeStr=nullptr)
	Streams a range object into a stream object. More...

template<int dimensions, bool with_offset>
const stream &	stream_value (const stream &os, const item< dimensions, with_offset > &rhs, char *typeStr=nullptr)
	Streams an item object into a stream object. More...

template<typename valueT >
const stream &	operator<< (const stream &os, valueT rhs)
	Template stream operator that streams a value to a stream object. More...

template<typename elementT , int kDimensions>
const stream &	operator<< (const stream &os, const vec< elementT, kDimensions > rhs)
	Template stream operator that streams a vec object to a stream object. More...

template<typename elementT , int kDimensions, int... swizzleIndexes>
const stream &	operator<< (const stream &os, const swizzled_vec< elementT, kDimensions, swizzleIndexes... > rhs)
	Streams a swizzled_vec object to a stream object. More...

Variables
COMPUTECPP_ACCESSOR_LINUX_ALIGNMENT	COMPUTECPP_CONVERT_ATTR

COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT	kDims

COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT	kMode

COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT	kTarget

AllocatorT &	bufferRef

AllocatorT id< kDims >	accessOffset

AllocatorT id< kDims > range< kDims >	accessRange

COMPUTECPP_ACCESSOR_LINUX_ALIGNMENT	COMPUTECPP_CONVERT_ATTR_PLACEHOLDER

const functorT &	functor

range< dimensions >	localRange

class cl::sycl::device_event	COMPUTECPP_PRIVATE_MEMORY_ATTR

class COMPUTECPP_EXPORT COMPUTECPP_SAMPLER_WINDOWS_ALIGNMENT cl::sycl::sampler	COMPUTECPP_CONVERT_ATTR_SAMPLER

const stream_manipulator	scientific = stream_manipulator::scientific
	Const value defined to the scientific stream manipulator, used in conjunction with stream operators. More...

const stream_manipulator	dec = stream_manipulator::dec
	Stream manipulator for displaying numbers in base 10. More...

const stream_manipulator	hex = stream_manipulator::hex
	Const value defined to the hex stream manipulator, used in conjunction with stream operators. More...

const stream_manipulator	oct = stream_manipulator::oct
	Const value defined to the oct stream manipulator, used in conjunction with stream operators. More...

const stream_manipulator	showbase = stream_manipulator::showbase
	Const value defined to the showbase stream manipulator, used in conjunction with stream operators. More...

const stream_manipulator	noshowbase = stream_manipulator::noshowbase
	Stream manipulator for reverting showbase. More...

const stream_manipulator	showpos = stream_manipulator::showpos
	Const value defined to the showpos stream manipulator, used in conjunction with stream operators. More...

const stream_manipulator	noshowpos = stream_manipulator::noshowpos
	Stream manipulator for reverting showpos. More...

const stream_manipulator	endl = stream_manipulator::endl
	Const value defined to the endl stream manipulator, used in conjunction with stream operators. More...

const stream_manipulator	fixed = stream_manipulator::fixed
	Stream manipulator that forces floating point numbers to be displayed using fixed precision. More...

const stream_manipulator	hexfloat = stream_manipulator::hexfloat
	Stream manipulator that displays floating point numbers in base 16. More...

const stream_manipulator	defaultfloat = stream_manipulator::defaultfloat
	Stream manipulator that reverts the display of floating point numbers back to the default representation. More...

const stream_manipulator	flush = stream_manipulator::flush
	Stream manipulator that flushes the current stream buffer. More...

Typedef Documentation

◆ async_handler

using cl::sycl::async_handler = typedef cl::sycl::function_class<void(exception_list)>

async_handler type definition.

This is the type expected by a device to report asynchronous errors.

Definition at line 93 of file exception_list.h.

◆ atomic_float

using cl::sycl::atomic_float = typedef atomic<cl_float>

Atomic float object with default global address space.

Definition at line 128 of file atomic.h.

◆ atomic_int

using cl::sycl::atomic_int = typedef atomic<cl_int>

Atomic int object with default global address space.

Definition at line 122 of file atomic.h.

◆ atomic_uint

using cl::sycl::atomic_uint = typedef atomic<cl_uint>

Atomic unsigned int object with default global address space.

Definition at line 125 of file atomic.h.

◆ backend_input_t

template<backend Backend, typename SyclType >

using cl::sycl::backend_input_t = typedef typename backend_traits<Backend>::template input_type<SyclType>

Shorthand for retrieving the input_type for a specific backend.

Template Parameters

Backend	The backend to inspect
SyclType	Type of the SYCL object

Note: Added in SYCL 2020

Definition at line 100 of file backend.h.

◆ backend_return_t

template<backend Backend, typename SyclType >

using cl::sycl::backend_return_t = typedef typename backend_traits<Backend>::template return_type<SyclType>

Shorthand for retrieving the return_type for a specific backend.

Template Parameters

Backend	The backend to inspect
SyclType	Type of the SYCL object

Note: Added in SYCL 2020

Definition at line 109 of file backend.h.

◆ base_t

using cl::sycl::base_t = typedef detail::accessor_buffer_interface<elemT, kDims, kMode, kTarget, access::placeholder::true_t>

Definition at line 923 of file buffer_accessor.h.

◆ bitset_class

template<size_t Size>

using cl::sycl::bitset_class = typedef std::bitset<Size>

Definition at line 322 of file common.h.

◆ byte

using cl::sycl::byte = typedef unsigned char

Definition at line 328 of file common.h.

◆ char16

using cl::sycl::char16 = typedef vec<char, 16>

Definition at line 56 of file vec_types_defines.h.

◆ char2

using cl::sycl::char2 = typedef vec<char, 2>

Definition at line 52 of file vec_types_defines.h.

◆ char3

using cl::sycl::char3 = typedef vec<char, 3>

Definition at line 53 of file vec_types_defines.h.

◆ char4

using cl::sycl::char4 = typedef vec<char, 4>

Definition at line 54 of file vec_types_defines.h.

◆ char8

using cl::sycl::char8 = typedef vec<char, 8>

Definition at line 55 of file vec_types_defines.h.

◆ cl_bool

using cl::sycl::cl_bool = typedef bool

Boolean interop type.

Definition at line 30 of file cl_types.h.

◆ cl_char

using cl::sycl::cl_char = typedef signed char

char (1 byte) interop type

Definition at line 32 of file cl_types.h.

◆ cl_char16

using cl::sycl::cl_char16 = typedef vec<cl_char, 16>

char vector interop type

Definition at line 37 of file cl_vec_types.h.

◆ cl_char2

using cl::sycl::cl_char2 = typedef vec<cl_char, 2>

char vector interop type

Definition at line 29 of file cl_vec_types.h.

◆ cl_char3

using cl::sycl::cl_char3 = typedef vec<cl_char, 3>

char vector interop type

Definition at line 31 of file cl_vec_types.h.

◆ cl_char4

using cl::sycl::cl_char4 = typedef vec<cl_char, 4>

char vector interop type

Definition at line 33 of file cl_vec_types.h.

◆ cl_char8

using cl::sycl::cl_char8 = typedef vec<cl_char, 8>

char vector interop type

Definition at line 35 of file cl_vec_types.h.

◆ cl_double

using cl::sycl::cl_double = typedef double

double interop type

Definition at line 46 of file cl_types.h.

◆ cl_double16

using cl::sycl::cl_double16 = typedef vec<cl_double, 16>

double vector interop type

Definition at line 136 of file cl_vec_types.h.

◆ cl_double2

using cl::sycl::cl_double2 = typedef vec<cl_double, 2>

double vector interop type

Definition at line 128 of file cl_vec_types.h.

◆ cl_double3

using cl::sycl::cl_double3 = typedef vec<cl_double, 3>

double vector interop type

Definition at line 130 of file cl_vec_types.h.

◆ cl_double4

using cl::sycl::cl_double4 = typedef vec<cl_double, 4>

double vector interop type

Definition at line 132 of file cl_vec_types.h.

◆ cl_double8

using cl::sycl::cl_double8 = typedef vec<cl_double, 8>

double vector interop type

Definition at line 134 of file cl_vec_types.h.

◆ cl_float

using cl::sycl::cl_float = typedef float

float interop type

Definition at line 44 of file cl_types.h.

◆ cl_float16

using cl::sycl::cl_float16 = typedef vec<cl_float, 16>

float vector interop type

Definition at line 125 of file cl_vec_types.h.

◆ cl_float2

using cl::sycl::cl_float2 = typedef vec<cl_float, 2>

float vector interop type

Definition at line 117 of file cl_vec_types.h.

◆ cl_float3

using cl::sycl::cl_float3 = typedef vec<cl_float, 3>

float vector interop type

Definition at line 119 of file cl_vec_types.h.

◆ cl_float4

using cl::sycl::cl_float4 = typedef vec<cl_float, 4>

float vector interop type

Definition at line 121 of file cl_vec_types.h.

◆ cl_float8

using cl::sycl::cl_float8 = typedef vec<cl_float, 8>

float vector interop type

Definition at line 123 of file cl_vec_types.h.

◆ cl_half

using cl::sycl::cl_half = typedef half

half scalar type

Definition at line 54 of file cl_types.h.

◆ cl_half16

using cl::sycl::cl_half16 = typedef vec<cl_half, 16>

16 dimensional half vector interop type

Definition at line 147 of file cl_vec_types.h.

◆ cl_half2

using cl::sycl::cl_half2 = typedef vec<cl_half, 2>

2 dimensional half vector interop type

Definition at line 139 of file cl_vec_types.h.

◆ cl_half3

using cl::sycl::cl_half3 = typedef vec<cl_half, 3>

3 dimensional half vector interop type

Definition at line 141 of file cl_vec_types.h.

◆ cl_half4

using cl::sycl::cl_half4 = typedef vec<cl_half, 4>

4 dimensional half vector interop type

Definition at line 143 of file cl_vec_types.h.

◆ cl_half8

using cl::sycl::cl_half8 = typedef vec<cl_half, 8>

8 dimensional half vector interop type

Definition at line 145 of file cl_vec_types.h.

◆ cl_int

using cl::sycl::cl_int = typedef int

int (4 bytes) interop type

Definition at line 40 of file cl_types.h.

◆ cl_int16

using cl::sycl::cl_int16 = typedef vec<cl_int, 16>

int vector interop type

Definition at line 81 of file cl_vec_types.h.

◆ cl_int2

using cl::sycl::cl_int2 = typedef vec<cl_int, 2>

int vector interop type

Definition at line 73 of file cl_vec_types.h.

◆ cl_int3

using cl::sycl::cl_int3 = typedef vec<cl_int, 3>

int vector interop type

Definition at line 75 of file cl_vec_types.h.

◆ cl_int4

using cl::sycl::cl_int4 = typedef vec<cl_int, 4>

int vector interop type

Definition at line 77 of file cl_vec_types.h.

◆ cl_int8

using cl::sycl::cl_int8 = typedef vec<cl_int, 8>

int vector interop type

Definition at line 79 of file cl_vec_types.h.

◆ cl_long

using cl::sycl::cl_long = typedef int64_t

long (8 bytes) interop type

Definition at line 49 of file cl_types.h.

◆ cl_long16

using cl::sycl::cl_long16 = typedef vec<cl_long, 16>

long vector interop type

Definition at line 103 of file cl_vec_types.h.

◆ cl_long2

using cl::sycl::cl_long2 = typedef vec<cl_long, 2>

long vector interop type

Definition at line 95 of file cl_vec_types.h.

◆ cl_long3

using cl::sycl::cl_long3 = typedef vec<cl_long, 3>

long vector interop type

Definition at line 97 of file cl_vec_types.h.

◆ cl_long4

using cl::sycl::cl_long4 = typedef vec<cl_long, 4>

long vector interop type

Definition at line 99 of file cl_vec_types.h.

◆ cl_long8

using cl::sycl::cl_long8 = typedef vec<cl_long, 8>

long vector interop type

Definition at line 101 of file cl_vec_types.h.

◆ cl_short

using cl::sycl::cl_short = typedef int16_t

short (2 bytes) interop type

Definition at line 36 of file cl_types.h.

◆ cl_short16

using cl::sycl::cl_short16 = typedef vec<cl_short, 16>

short vector interop type

Definition at line 59 of file cl_vec_types.h.

◆ cl_short2

using cl::sycl::cl_short2 = typedef vec<cl_short, 2>

short vector interop type

Definition at line 51 of file cl_vec_types.h.

◆ cl_short3

using cl::sycl::cl_short3 = typedef vec<cl_short, 3>

short vector interop type

Definition at line 53 of file cl_vec_types.h.

◆ cl_short4

using cl::sycl::cl_short4 = typedef vec<cl_short, 4>

short vector interop type

Definition at line 55 of file cl_vec_types.h.

◆ cl_short8

using cl::sycl::cl_short8 = typedef vec<cl_short, 8>

short vector interop type

Definition at line 57 of file cl_vec_types.h.

◆ cl_uchar

using cl::sycl::cl_uchar = typedef unsigned char

unsigned char (1 byte) interop type

Definition at line 34 of file cl_types.h.

◆ cl_uchar16

using cl::sycl::cl_uchar16 = typedef vec<cl_uchar, 16>

unsigned char vector interop type

Definition at line 48 of file cl_vec_types.h.

◆ cl_uchar2

using cl::sycl::cl_uchar2 = typedef vec<cl_uchar, 2>

unsigned char vector interop type

Definition at line 40 of file cl_vec_types.h.

◆ cl_uchar3

using cl::sycl::cl_uchar3 = typedef vec<cl_uchar, 3>

unsigned char vector interop type

Definition at line 42 of file cl_vec_types.h.

◆ cl_uchar4

using cl::sycl::cl_uchar4 = typedef vec<cl_uchar, 4>

unsigned char vector interop type

Definition at line 44 of file cl_vec_types.h.

◆ cl_uchar8

using cl::sycl::cl_uchar8 = typedef vec<cl_uchar, 8>

unsigned char vector interop type

Definition at line 46 of file cl_vec_types.h.

◆ cl_uint

using cl::sycl::cl_uint = typedef unsigned int

unsigned int (4 bytes) interop type

Definition at line 42 of file cl_types.h.

◆ cl_uint16

using cl::sycl::cl_uint16 = typedef vec<cl_uint, 16>

unsigned int vector interop type

Definition at line 92 of file cl_vec_types.h.

◆ cl_uint2

using cl::sycl::cl_uint2 = typedef vec<cl_uint, 2>

unsigned int vector interop type

Definition at line 84 of file cl_vec_types.h.

◆ cl_uint3

using cl::sycl::cl_uint3 = typedef vec<cl_uint, 3>

unsigned int vector interop type

Definition at line 86 of file cl_vec_types.h.

◆ cl_uint4

using cl::sycl::cl_uint4 = typedef vec<cl_uint, 4>

unsigned int vector interop type

Definition at line 88 of file cl_vec_types.h.

◆ cl_uint8

using cl::sycl::cl_uint8 = typedef vec<cl_uint, 8>

unsigned int vector interop type

Definition at line 90 of file cl_vec_types.h.

◆ cl_ulong

using cl::sycl::cl_ulong = typedef uint64_t

unsigned long (8 bytes) interop type

Definition at line 51 of file cl_types.h.

◆ cl_ulong16

using cl::sycl::cl_ulong16 = typedef vec<cl_ulong, 16>

unsigned long vector interop type

Definition at line 114 of file cl_vec_types.h.

◆ cl_ulong2

using cl::sycl::cl_ulong2 = typedef vec<cl_ulong, 2>

unsigned long vector interop type

Definition at line 106 of file cl_vec_types.h.

◆ cl_ulong3

using cl::sycl::cl_ulong3 = typedef vec<cl_ulong, 3>

unsigned long vector interop type

Definition at line 108 of file cl_vec_types.h.

◆ cl_ulong4

using cl::sycl::cl_ulong4 = typedef vec<cl_ulong, 4>

unsigned long vector interop type

Definition at line 110 of file cl_vec_types.h.

◆ cl_ulong8

using cl::sycl::cl_ulong8 = typedef vec<cl_ulong, 8>

unsigned long vector interop type

Definition at line 112 of file cl_vec_types.h.

◆ cl_ushort

using cl::sycl::cl_ushort = typedef uint16_t

unsigned short (2 bytes) interop type

Definition at line 38 of file cl_types.h.

◆ cl_ushort16

using cl::sycl::cl_ushort16 = typedef vec<cl_ushort, 16>

unsigned short vector interop type

Definition at line 70 of file cl_vec_types.h.

◆ cl_ushort2

using cl::sycl::cl_ushort2 = typedef vec<cl_ushort, 2>

unsigned short vector interop type

Definition at line 62 of file cl_vec_types.h.

◆ cl_ushort3

using cl::sycl::cl_ushort3 = typedef vec<cl_ushort, 3>

unsigned short vector interop type

Definition at line 64 of file cl_vec_types.h.

◆ cl_ushort4

using cl::sycl::cl_ushort4 = typedef vec<cl_ushort, 4>

unsigned short vector interop type

Definition at line 66 of file cl_vec_types.h.

◆ cl_ushort8

using cl::sycl::cl_ushort8 = typedef vec<cl_ushort, 8>

unsigned short vector interop type

Definition at line 68 of file cl_vec_types.h.

◆ constant_ptr

template<typename dataType >

using cl::sycl::constant_ptr = typedef multi_ptr<dataType, access::address_space::constant_space>

constant_ptr, pointer class definition for data living in the OpenCL constant address space

Template Parameters

dataType Data type the object manipulates.

Definition at line 1157 of file multi_pointer.h.

◆ double16

using cl::sycl::double16 = typedef vec<double, 16>

Definition at line 40 of file vec_types_defines.h.

◆ double2

using cl::sycl::double2 = typedef vec<double, 2>

Definition at line 36 of file vec_types_defines.h.

◆ double3

using cl::sycl::double3 = typedef vec<double, 3>

Definition at line 37 of file vec_types_defines.h.

◆ double4

using cl::sycl::double4 = typedef vec<double, 4>

Definition at line 38 of file vec_types_defines.h.

◆ double8

using cl::sycl::double8 = typedef vec<double, 8>

Definition at line 39 of file vec_types_defines.h.

◆ exception_ptr_class

using cl::sycl::exception_ptr_class = typedef std::exception_ptr

Class used to store exception objects and transfer them across thread, equivalent to std::exception_ptr.

Definition at line 27 of file exception_list.h.

◆ float16

using cl::sycl::float16 = typedef vec<float, 16>

Definition at line 34 of file vec_types_defines.h.

◆ float2

using cl::sycl::float2 = typedef vec<float, 2>

Definition at line 30 of file vec_types_defines.h.

◆ float3

using cl::sycl::float3 = typedef vec<float, 3>

Definition at line 31 of file vec_types_defines.h.

◆ float4

using cl::sycl::float4 = typedef vec<float, 4>

Definition at line 32 of file vec_types_defines.h.

◆ float8

using cl::sycl::float8 = typedef vec<float, 8>

Definition at line 33 of file vec_types_defines.h.

◆ function_class

template<typename T >

using cl::sycl::function_class = typedef std::function<T>

Definition at line 305 of file common.h.

◆ global_ptr

template<typename dataType >

using cl::sycl::global_ptr = typedef multi_ptr<dataType, access::address_space::global_space>

global_ptr pointer class definition for data pointing to the OpenCL global address space.

Template Parameters

dataType Data type the object manipulates.

Definition at line 1136 of file multi_pointer.h.

◆ half16

using cl::sycl::half16 = typedef vec<half, 16>

Definition at line 49 of file vec_types_defines.h.

◆ half2

using cl::sycl::half2 = typedef vec<half, 2>

Definition at line 45 of file vec_types_defines.h.

◆ half3

using cl::sycl::half3 = typedef vec<half, 3>

Definition at line 46 of file vec_types_defines.h.

◆ half4

using cl::sycl::half4 = typedef vec<half, 4>

Definition at line 47 of file vec_types_defines.h.

◆ half8

using cl::sycl::half8 = typedef vec<half, 8>

Definition at line 48 of file vec_types_defines.h.

◆ hash_class

template<typename T >

using cl::sycl::hash_class = typedef std::hash<T>

Definition at line 319 of file common.h.

◆ int16

using cl::sycl::int16 = typedef vec<int, 16>

Definition at line 86 of file vec_types_defines.h.

◆ int2

using cl::sycl::int2 = typedef vec<int, 2>

Definition at line 82 of file vec_types_defines.h.

◆ int3

using cl::sycl::int3 = typedef vec<int, 3>

Definition at line 83 of file vec_types_defines.h.

◆ int4

using cl::sycl::int4 = typedef vec<int, 4>

Definition at line 84 of file vec_types_defines.h.

◆ int8

using cl::sycl::int8 = typedef vec<int, 8>

Definition at line 85 of file vec_types_defines.h.

◆ local_ptr

template<typename dataType >

using cl::sycl::local_ptr = typedef multi_ptr<dataType, access::address_space::local_space>

local_ptr pointer class definition for data pointing to the OpenCL local address space

Template Parameters

dataType Data type the object manipulates.

Definition at line 1143 of file multi_pointer.h.

◆ long16

using cl::sycl::long16 = typedef vec<long, 16>

Definition at line 98 of file vec_types_defines.h.

◆ long2

using cl::sycl::long2 = typedef vec<long, 2>

Definition at line 94 of file vec_types_defines.h.

◆ long3

using cl::sycl::long3 = typedef vec<long, 3>

Definition at line 95 of file vec_types_defines.h.

◆ long4

using cl::sycl::long4 = typedef vec<long, 4>

Definition at line 96 of file vec_types_defines.h.

◆ long8

using cl::sycl::long8 = typedef vec<long, 8>

Definition at line 97 of file vec_types_defines.h.

◆ longlong16

using cl::sycl::longlong16 = typedef vec<long long, 16>

Definition at line 110 of file vec_types_defines.h.

◆ longlong2

using cl::sycl::longlong2 = typedef vec<long long, 2>

Definition at line 106 of file vec_types_defines.h.

◆ longlong3

using cl::sycl::longlong3 = typedef vec<long long, 3>

Definition at line 107 of file vec_types_defines.h.

◆ longlong4

using cl::sycl::longlong4 = typedef vec<long long, 4>

Definition at line 108 of file vec_types_defines.h.

◆ longlong8

using cl::sycl::longlong8 = typedef vec<long long, 8>

Definition at line 109 of file vec_types_defines.h.

◆ mutex_class

using cl::sycl::mutex_class = typedef std::mutex

Definition at line 307 of file common.h.

◆ private_ptr

template<typename dataType >

using cl::sycl::private_ptr = typedef multi_ptr<dataType, access::address_space::private_space>

private_ptr, pointer class definition for data pointing to the OpenCL private address space

Template Parameters

dataType Data type the object manipulates.

Definition at line 1150 of file multi_pointer.h.

◆ schar16

using cl::sycl::schar16 = typedef vec<signed char, 16>

Definition at line 62 of file vec_types_defines.h.

◆ schar2

using cl::sycl::schar2 = typedef vec<signed char, 2>

Definition at line 58 of file vec_types_defines.h.

◆ schar3

using cl::sycl::schar3 = typedef vec<signed char, 3>

Definition at line 59 of file vec_types_defines.h.

◆ schar4

using cl::sycl::schar4 = typedef vec<signed char, 4>

Definition at line 60 of file vec_types_defines.h.

◆ schar8

using cl::sycl::schar8 = typedef vec<signed char, 8>

Definition at line 61 of file vec_types_defines.h.

◆ shared_ptr_class

template<typename T >

using cl::sycl::shared_ptr_class = typedef std::shared_ptr<T>

Definition at line 313 of file common.h.

◆ short16

using cl::sycl::short16 = typedef vec<short, 16>

Definition at line 74 of file vec_types_defines.h.

◆ short2

using cl::sycl::short2 = typedef vec<short, 2>

Definition at line 70 of file vec_types_defines.h.

◆ short3

using cl::sycl::short3 = typedef vec<short, 3>

Definition at line 71 of file vec_types_defines.h.

◆ short4

using cl::sycl::short4 = typedef vec<short, 4>

Definition at line 72 of file vec_types_defines.h.

◆ short8

using cl::sycl::short8 = typedef vec<short, 8>

Definition at line 73 of file vec_types_defines.h.

◆ string_class

using cl::sycl::string_class = typedef std::string

Definition at line 302 of file common.h.

◆ uchar16

using cl::sycl::uchar16 = typedef vec<unsigned char, 16>

Definition at line 68 of file vec_types_defines.h.

◆ uchar2

using cl::sycl::uchar2 = typedef vec<unsigned char, 2>

Definition at line 64 of file vec_types_defines.h.

◆ uchar3

using cl::sycl::uchar3 = typedef vec<unsigned char, 3>

Definition at line 65 of file vec_types_defines.h.

◆ uchar4

using cl::sycl::uchar4 = typedef vec<unsigned char, 4>

Definition at line 66 of file vec_types_defines.h.

◆ uchar8

using cl::sycl::uchar8 = typedef vec<unsigned char, 8>

Definition at line 67 of file vec_types_defines.h.

◆ uint16

using cl::sycl::uint16 = typedef vec<unsigned int, 16>

Definition at line 92 of file vec_types_defines.h.

◆ uint2

using cl::sycl::uint2 = typedef vec<unsigned int, 2>

Definition at line 88 of file vec_types_defines.h.

◆ uint3

using cl::sycl::uint3 = typedef vec<unsigned int, 3>

Definition at line 89 of file vec_types_defines.h.

◆ uint4

using cl::sycl::uint4 = typedef vec<unsigned int, 4>

Definition at line 90 of file vec_types_defines.h.

◆ uint8

using cl::sycl::uint8 = typedef vec<unsigned int, 8>

Definition at line 91 of file vec_types_defines.h.

◆ ulong16

using cl::sycl::ulong16 = typedef vec<unsigned long, 16>

Definition at line 104 of file vec_types_defines.h.

◆ ulong2

using cl::sycl::ulong2 = typedef vec<unsigned long, 2>

Definition at line 100 of file vec_types_defines.h.

◆ ulong3

using cl::sycl::ulong3 = typedef vec<unsigned long, 3>

Definition at line 101 of file vec_types_defines.h.

◆ ulong4

using cl::sycl::ulong4 = typedef vec<unsigned long, 4>

Definition at line 102 of file vec_types_defines.h.

◆ ulong8

using cl::sycl::ulong8 = typedef vec<unsigned long, 8>

Definition at line 103 of file vec_types_defines.h.

◆ ulonglong16

using cl::sycl::ulonglong16 = typedef vec<unsigned long long, 16>

Definition at line 116 of file vec_types_defines.h.

◆ ulonglong2

using cl::sycl::ulonglong2 = typedef vec<unsigned long long, 2>

Definition at line 112 of file vec_types_defines.h.

◆ ulonglong3

using cl::sycl::ulonglong3 = typedef vec<unsigned long long, 3>

Definition at line 113 of file vec_types_defines.h.

◆ ulonglong4

using cl::sycl::ulonglong4 = typedef vec<unsigned long long, 4>

Definition at line 114 of file vec_types_defines.h.

◆ ulonglong8

using cl::sycl::ulonglong8 = typedef vec<unsigned long long, 8>

Definition at line 115 of file vec_types_defines.h.

◆ unique_ptr_class

template<typename T , class D = std::default_delete<T>>

using cl::sycl::unique_ptr_class = typedef std::unique_ptr<T, D>

Definition at line 310 of file common.h.

◆ ushort16

using cl::sycl::ushort16 = typedef vec<unsigned short, 16>

Definition at line 80 of file vec_types_defines.h.

◆ ushort2

using cl::sycl::ushort2 = typedef vec<unsigned short, 2>

Definition at line 76 of file vec_types_defines.h.

◆ ushort3

using cl::sycl::ushort3 = typedef vec<unsigned short, 3>

Definition at line 77 of file vec_types_defines.h.

◆ ushort4

using cl::sycl::ushort4 = typedef vec<unsigned short, 4>

Definition at line 78 of file vec_types_defines.h.

◆ ushort8

using cl::sycl::ushort8 = typedef vec<unsigned short, 8>

Definition at line 79 of file vec_types_defines.h.

◆ vector_class

template<typename T , typename Alloc = std::allocator<T>>

using cl::sycl::vector_class = typedef std::vector<T, Alloc>

Definition at line 300 of file common.h.

◆ weak_ptr_class

template<typename T >

using cl::sycl::weak_ptr_class = typedef std::weak_ptr<T>

Definition at line 316 of file common.h.

Enumeration Type Documentation

◆ addressing_mode

enum cl::sycl::addressing_mode : unsigned int

strong

Enum class for specifying the addressing mode of a sampler.

Values are hard coded to match those of OpenCL in order to allow a simple cast when converting.

Enumerator
none
clamp_to_edge
clamp
repeat
mirrored_repeat

Definition at line 67 of file sampler.h.

◆ aspect_impl

enum cl::sycl::aspect_impl

strong

Enumerates the aspects which can be queried on a platform or device.

Enumerator
COMPUTECPP_DEPRECATED_BY_SYCL_202001
cpu
gpu
accelerator
custom
fp16
fp64
COMPUTECPP_DEPRECATED_BY_SYCL_202001
COMPUTECPP_DEPRECATED_BY_SYCL_202001
image
online_compiler
online_linker
queue_profiling
usm_device_allocations
usm_host_allocations
usm_shared_allocations
usm_restricted_shared_allocations
COMPUTECPP_DEPRECATED_BY_SYCL_202001
emulated
host_debuggable
atomic64
usm_atomic_host_allocations
usm_atomic_shared_allocations
usm_system_allocations

Definition at line 26 of file aspect.h.

◆ backend

enum cl::sycl::backend

strong

Defines backends available in ComputeCpp.

Enumerator
host	Native C++ device.
opencl	OpenCL device.

Definition at line 34 of file backend.h.

◆ bundle_state

enum cl::sycl::bundle_state

strong

The compilation state of a kernel bundle.

There are three possible states: input, object, and executable. These represent the online compilation and linking status of a kernel bundle. An input bundle is typically some intermediate representation such as SPIR-V. At this stage, specialization variables may be set. This can be compiled to obtain a kernel bundle in the object state. The object state can then be linked to obtain a kernel in the executable state, suitable to be invoked on device.

Any given kernel bundle is not guaranteed to be available in any particular state.

Enumerator
input	Input: device images must still be compiled and linked before they can be executed.
object	A compiled but not linked kernel bundle.
executable	A compiled and linked kernel bundle. Can be executed.

Definition at line 32 of file bundle_state.h.

◆ coordinate_normalization_mode

enum cl::sycl::coordinate_normalization_mode : unsigned int

strong

Filtering mode description.

Enumerator
normalized
unnormalized

Definition at line 85 of file sampler.h.

◆ filtering_mode

enum cl::sycl::filtering_mode : unsigned int

strong

Enum class for specifying the filter mode of a sampler.

Values are hard coded to match those of OpenCL in order to allow a simple cast when converting.

Enumerator
nearest
linear

Definition at line 80 of file sampler.h.

◆ image_channel_order

enum cl::sycl::image_channel_order : unsigned int

strong

Specify the number of channels and the channel layout in which channels are stored in the image.

Enumerator
r
a
rg
ra
rgb
rgba
bgra
argb
intensity
luminance
rx
rgx
rgbx
abgr

Definition at line 74 of file image.h.

◆ image_channel_type

enum cl::sycl::image_channel_type : unsigned int

strong

Specify the size of the channels data type.

Enumerator
snorm_int8
snorm_int16
unorm_int8
unorm_int16
unorm_short_565
unorm_short_555
unorm_int_101010
signed_int8
signed_int16
signed_int32
unsigned_int8
unsigned_int16
unsigned_int32
fp16
fp32

Definition at line 93 of file image.h.

◆ memory_order

enum cl::sycl::memory_order : int

strong

Only relaxed memory order is supported in SYCL 1.2.1.

Enumerator
relaxed

Definition at line 59 of file atomic.h.

◆ program_state

enum cl::sycl::program_state

strong

Enum describing the build state of the program.

Enumerator
none
compiled
linked

Definition at line 100 of file program.h.

◆ rounding_mode

enum cl::sycl::rounding_mode

strong

Available vector rounding modes.

Enumerator
automatic
rte
rtz
rtp
rtn

Definition at line 66 of file vec_common.h.

◆ stream_manipulator

enum cl::sycl::stream_manipulator

strong

Enum class that enumerates the different stream manipulators.

Enumerator
endl
scientific
hex
oct
showbase
showpos
dec
noshowbase
noshowpos
fixed
hexfloat
defaultfloat
flush

Definition at line 109 of file stream.h.

◆ write_back

enum cl::sycl::write_back : bool

strong

Enumerator
disable_write_back
enable_write_back

Definition at line 40 of file storage_mem.h.

Function Documentation

◆ abs() [1/2]

template<typename F1 , typename F2 , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && std::is_same< F1, F2 >::value)||(detail::builtin::is_genfloath< F1 >::value && std::is_same< F2, half >::value)||(detail::builtin::is_genfloatf< F1 >::value && std::is_same< F2, float >::value)||(detail::builtin::is_genfloatd< F1 >::value && std::is_same< F2, double >::value)) >

F1 cl::sycl::abs	(	F1	x,
		F2	y
	)

noexcept

Compute absolute value of a floating-point number.

Redirects to fabs(x, y).

Template Parameters

F1	must model genfloat.
F2	must model either genfloat or be the scalar equivalent of F1.

Definition at line 74 of file math_common.h.

◆ abs() [2/2]

template<typename Integral , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value)) >

auto cl::sycl::abs ( Integral x ) -> detail::make_genuint_t<Integral>

noexcept

Returns |x|

Template Parameters

Integral must model geninteger.

Parameters

x

Returns: x if 0 <= x, -x otherwise.

Definition at line 96 of file math_integral.h.

◆ abs_diff()

template<typename Integral , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value)) >

auto cl::sycl::abs_diff	(	Integral	x,
		Integral	y
	)		-> detail::make_genuint_t<Integral>

noexcept

Returns |x - y| without modulo overflow.

Template Parameters

Integral must model geninteger.

Definition at line 106 of file math_integral.h.

◆ accessor() [1/14]

cl::sycl::accessor	(	storage_mem &	store,
		detail::access_range	accessRange,
		const property_list &	propList = `{}`
	)

protected

Constructs a ranged placeholder accessor.

Parameters

store	Storage object representing the buffer/image
accessRange	the offset and range that this accessor can access.
propList	Additional properties

Definition at line 931 of file buffer_accessor.h.

◆ accessor() [2/14]

cl::sycl::accessor ( const property_list & propList = {} )

Constructs a default placeholder accessor without associated storage.

Parameters

propList Additional properties

Definition at line 972 of file buffer_accessor.h.

◆ accessor() [3/14]

template<typename AllocatorT = default_allocator>

cl::sycl::accessor	(	buffer< elemT, detail::acc_interface_dims< kDims >::value, AllocatorT > &	bufferRef,
		const property_list &	propList = `{}`
	)

explicit

Constructs a placeholder accessor.

Template Parameters

AllocatorT Specifies the type of the buffer objects allocator.

Parameters

bufferRef	Reference to the buffer object the accessor is to access data from.
propList	Additional properties

Definition at line 981 of file buffer_accessor.h.

◆ accessor() [4/14]

template<typename AllocatorT >

cl::sycl::accessor	(	buffer< elemT, kDims, AllocatorT > &	bufferRef,
		range< kDims >	accessRange,
		const property_list &	propList = `{}`
	)

Constructs a ranged placeholder accessor.

Template Parameters

AllocatorT Type of the buffer allocator

Parameters

bufferRef	Buffer object where access is being requested
accessRange	Range of data to access
propList	Additional properties

Definition at line 1010 of file buffer_accessor.h.

◆ accessor() [5/14]

template<typename AllocatorT >

cl::sycl::accessor	(	buffer< elemT, kDims, AllocatorT > &	bufferRef,
		range< kDims >	accessRange,
		id< kDims >	accessOffset,
		const property_list &	propList = `{}`
	)

Constructs a ranged placeholder accessor.

Template Parameters

AllocatorT Specifies the type of the buffer objects allocator.

Parameters

bufferRef	Buffer object where access is being requested
accessRange	Range of data to access
accessOffset	Point where data access starts
propList	Additional properties

Definition at line 1022 of file buffer_accessor.h.

◆ accessor() [6/14]

cl::sycl::accessor	(	buffer< elemT, kDims > &	bufferRef,
		range< kDims >	accessRange,
		const property_list &	propList = `{}`
	)

Constructs a ranged placeholder accessor.

Parameters

bufferRef	Buffer object where access is being requested
accessRange	Range of data to access
propList	Additional properties

Definition at line 1033 of file buffer_accessor.h.

◆ accessor() [7/14]

cl::sycl::accessor	(	buffer< elemT, kDims > &	bufferRef,
		range< kDims >	accessRange,
		id< kDims >	accessOffset,
		const property_list &	propList = `{}`
	)

Constructs a ranged placeholder accessor.

Parameters

bufferRef	Buffer object where access is being requested
accessRange	Range of data to access
accessOffset	Point where data access starts
propList	Additional properties

Definition at line 1043 of file buffer_accessor.h.

◆ accessor() [8/14]

template<typename AllocatorT >

cl::sycl::accessor	(	buffer< elemT, kDims, AllocatorT > &	bufferRef,
		handler &	commandHandler,
		range< kDims >	accessRange,
		const property_list &	propList = `{}`
	)

Constructs a ranged placeholder accessor.

Template Parameters

AllocatorT Type of the buffer allocator

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Command group handler
accessRange	Range of data to access
propList	Additional properties

Definition at line 1056 of file buffer_accessor.h.

◆ accessor() [9/14]

template<typename AllocatorT >

cl::sycl::accessor	(	buffer< elemT, kDims, AllocatorT > &	bufferRef,
		handler &	commandHandler,
		range< kDims >	accessRange,
		id< kDims >	accessOffset,
		const property_list &	propList = `{}`
	)

Constructs a ranged placeholder accessor.

Template Parameters

AllocatorT Specifies the type of the buffer objects allocator.

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Reference to the handler object for the command group scope that the accessor is being constructed within.
accessRange	Range of data to access
accessOffset	Point where data access starts
propList	Additional properties

Definition at line 1071 of file buffer_accessor.h.

◆ accessor() [10/14]

cl::sycl::accessor	(	buffer< elemT, kDims > &	bufferRef,
		handler &	commandHandler,
		range< kDims >	accessRange,
		const property_list &	propList = `{}`
	)

Constructs a ranged placeholder accessor.

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Command group handler
accessRange	Range of data to access
propList	Additional properties

Definition at line 1083 of file buffer_accessor.h.

◆ accessor() [11/14]

cl::sycl::accessor	(	buffer< elemT, kDims > &	bufferRef,
		handler &	commandHandler,
		range< kDims >	accessRange,
		id< kDims >	accessOffset,
		const property_list &	propList = `{}`
	)

Constructs a ranged placeholder accessor.

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Reference to the handler object for the command group scope that the accessor is being constructed within.
accessRange	Range of data to access
accessOffset	Point where data access starts
propList	Additional properties

Definition at line 1096 of file buffer_accessor.h.

◆ accessor() [12/14]

template<typename AllocatorT >

cl::sycl::accessor	(	buffer< elemT, kDims, AllocatorT > &	bufferRef,
		handler &	commandHandler,
		const property_list &	propList = `{}`
	)

Constructs a placeholder buffer accessor.

Template Parameters

AllocatorT Type of the buffer allocator

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Command group handler
propList	Additional properties

Definition at line 1109 of file buffer_accessor.h.

◆ accessor() [13/14]

cl::sycl::accessor	(	buffer< elemT, kDims > &	bufferRef,
		handler &	commandHandler,
		const property_list &	propList = `{}`
	)

Constructs a placeholder buffer accessor.

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Command group handler
propList	Additional properties

Definition at line 1118 of file buffer_accessor.h.

◆ accessor() [14/14]

template<typename AllocatorT , COMPUTECPP_ENABLE_IF(elemT,(kDims==0)) >

cl::sycl::accessor	(	buffer< elemT, 1, AllocatorT > &	bufferRef,
		handler &	commandHandler,
		const property_list &	propList = `{}`
	)

Constructs a placeholder 0-dimensional buffer accessor.

Template Parameters

AllocatorT	Type of the buffer allocator
COMPUTECPP_ENABLE_IF	Only enabled for 0 dimensions

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Command group handler
propList	Additional properties

Definition at line 1130 of file buffer_accessor.h.

◆ acos()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::acos ( F x )

noexcept

Inverse cosine function.

Template Parameters

F	must model genfloat.

Definition at line 40 of file math_floating_point.h.

◆ acosh()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::acosh ( F x )

noexcept

Inverse hyperbolic cosine.

Template Parameters

F	must model genfloat.

Definition at line 49 of file math_floating_point.h.

◆ acospi()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::acospi ( F x )

noexcept

Compute acos(x)/π

Template Parameters

F	must model genfloat.

Definition at line 58 of file math_floating_point.h.

◆ add_sat()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::add_sat	(	Integral	x,
		Integral	y
	)

noexcept

Returns x + y and saturates the result.

Template Parameters

Integral must model geninteger.

Definition at line 124 of file math_integral.h.

◆ all()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_igeninteger< Integral >::value) >

int cl::sycl::all ( Integral x )

noexcept

Returns 1 if the most significant bit in all components of x is set; otherwise returns 0.

Template Parameters

Integral must model igeninteger.

Definition at line 204 of file math_relational.h.

◆ any()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_igeninteger< Integral >::value) >

int cl::sycl::any ( Integral x )

noexcept

Returns 1 if the most significant bit in any component of x is set; otherwise returns 0.

Template Parameters

Integral must model igeninteger.

Definition at line 194 of file math_relational.h.

◆ append_bool()

void cl::sycl::append_bool	(	const stream &	os,
		const bool &	value
	)

inline

Inline function that appends a bool value to the a object.

Parameters

os	The stream object to be appended to.
value	A bool value to be appended to the stream object.

Definition at line 843 of file stream.h.

◆ append_char()

void cl::sycl::append_char	(	const stream &	os,
		const char &	value
	)

inline

Inline function that appends a char value to the a object.

Parameters

os	The stream object to be appended to.
value	A char value to be appended to the stream object.

Definition at line 834 of file stream.h.

◆ append_floating_point()

template<typename valueT >

void cl::sycl::append_floating_point	(	const stream &	os,
		const valueT &	value
	)

inline

Template function that appends a floating point value to a stream object.

Template Parameters

valueT The type of the floating point value to be appended.

Parameters

os	The stream object to be appended to.
value	A floating point value to be appended to the stream object.

Definition at line 672 of file stream.h.

◆ append_hexidecimal()

template<typename valueT >

void cl::sycl::append_hexidecimal	(	const stream &	os,
		const valueT &	value
	)

Template function that appends a hexadecimal value to a stream object.

Template Parameters

valueT The type of the hexadecimal value to be appended.

Parameters

os	The stream object to be appended to.
value	A hexadecimal value to be appended to the stream object.

Definition at line 568 of file stream.h.

◆ append_index_array()

void cl::sycl::append_index_array	(	const stream &	os,
		const detail::index_array &	value,
		int	numElements
	)

inline

Appends an index_array object to the stream.

Parameters

os	Stream object to append to
value	Array of indexes
numElements	Number of elements of the index array

Definition at line 856 of file stream.h.

◆ append_integral()

template<typename valueT >

int cl::sycl::append_integral	(	const stream &	os,
		const valueT &	value,
		bool	useManipulators = `true`
	)

inline

Template function that appends a integral value to a stream object.

Template Parameters

valueT The type of the integral value to be appended.

Parameters

os	The stream object to be appended to.
value	A integral value to be appended to the stream object.
useManipulators	Whether to apply stream manipulators when streaming

Definition at line 599 of file stream.h.

◆ append_string()

void cl::sycl::append_string	(	const stream &	os,
		const char *	value
	)

inline

Inline function that appends a string value to a stream object.

Parameters

os	The stream object to be appended to.
value	A string value to be appended to the stream object.

Definition at line 801 of file stream.h.

◆ append_type_string()

void cl::sycl::append_type_string	(	const stream &	os,
		char *	typeStr,
		int	numChars,
		int	dimensions = `0`
	)

inline

Appends a string representation of a type to the stream.

Parameters

os	Stream object to append to
typeStr	String representing the type name. Can contain a single percentage character that will get replaced by the value of dimensions
numChars	Number of characters in the provided type string
dimensions	Optional number of dimensions of the type

Definition at line 814 of file stream.h.

◆ asin()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::asin ( F x )

noexcept

Inverse sine function.

Template Parameters

F	must model genfloat.

Definition at line 67 of file math_floating_point.h.

◆ asinh()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::asinh ( F x )

noexcept

Inverse hyperbolic sine.

Template Parameters

F	must model genfloat.

Definition at line 76 of file math_floating_point.h.

◆ asinpi()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::asinpi ( F x )

noexcept

Compute asin(x)/π

Template Parameters

F	must model genfloat.

Definition at line 85 of file math_floating_point.h.

◆ atan()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::atan ( F y_over_x )

noexcept

Inverse tangent function.

Template Parameters

F	must model genfloat.

Definition at line 94 of file math_floating_point.h.

◆ atan2()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

F cl::sycl::atan2	(	F	y,
		F	x
	)

noexcept

Compute atan(y/x).

Template Parameters

F	must model genfloat.

Definition at line 103 of file math_floating_point.h.

◆ atan2pi()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

F cl::sycl::atan2pi	(	F	y,
		F	x
	)

noexcept

Compute atan(y/x)/π.

Template Parameters

F	must model genfloat.

Definition at line 130 of file math_floating_point.h.

◆ atanh()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::atanh ( F x )

noexcept

Inverse hyperbolic tangent.

Template Parameters

F	must model genfloat.

Definition at line 112 of file math_floating_point.h.

◆ atanpi()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::atanpi ( F x )

noexcept

Compute atan(x)/π

Template Parameters

F	must model genfloat.

Definition at line 121 of file math_floating_point.h.

◆ atomic_compare_exchange_strong()

template<typename T , access::address_space addressSpace>

cl_bool cl::sycl::atomic_compare_exchange_strong	(	atomic< T, addressSpace >	object,
		T &	expected,
		T	desired,
		memory_order	success = `memory_order::relaxed`,
		memory_order	fail = `memory_order::relaxed`
	)

inline

Global function atomic_compare_exchange.

Calls compare_exchange on SYCL atomic object.

Parameters

object	The atomic object to compare_exchange with
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: Whether comparison succeeds or fails

Definition at line 578 of file atomic.h.

◆ atomic_exchange()

template<typename T , access::address_space addressSpace>

T cl::sycl::atomic_exchange	(	atomic< T, addressSpace >	object,
		T	operand,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_exchange.

Calls exchange on SYCL atomic object.

Parameters

object	The atomic object to exchange with
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: the old value of *object

Definition at line 564 of file atomic.h.

◆ atomic_fetch_add()

template<typename T , access::address_space addressSpace>

T cl::sycl::atomic_fetch_add	(	atomic< T, addressSpace >	object,
		T	operand,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_add.

Calls add on SYCL atomic object.

Parameters

object	The atomic object to add to
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: The old value of *object

Definition at line 592 of file atomic.h.

◆ atomic_fetch_and()

template<typename T , access::address_space addressSpace>

T cl::sycl::atomic_fetch_and	(	atomic< T, addressSpace >	object,
		T	operand,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_and.

Calls and on SYCL atomic object.

Parameters

object	The atomic object to and with
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: The old value of *object

Definition at line 616 of file atomic.h.

◆ atomic_fetch_max()

template<typename T , access::address_space addressSpace>

T cl::sycl::atomic_fetch_max	(	atomic< T, addressSpace >	object,
		T	operand,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_max.

Calculates max(object, operand), storing the result in object.

Parameters

object	The atomic object to perform max on
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: The old value of *object

Definition at line 668 of file atomic.h.

◆ atomic_fetch_min()

template<typename T , access::address_space addressSpace>

T cl::sycl::atomic_fetch_min	(	atomic< T, addressSpace >	object,
		T	operand,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_min.

Calculates min(object, operand), storing the result in object.

Parameters

object	The atomic object to perform min on
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: The old value of *object

Definition at line 654 of file atomic.h.

◆ atomic_fetch_or()

template<typename T , access::address_space addressSpace>

T cl::sycl::atomic_fetch_or	(	atomic< T, addressSpace >	object,
		T	operand,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_or.

Calls or on SYCL atomic object.

Parameters

object	The atomic object to or with
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: The old value of *object

Definition at line 628 of file atomic.h.

◆ atomic_fetch_sub()

template<typename T , access::address_space addressSpace>

T cl::sycl::atomic_fetch_sub	(	atomic< T, addressSpace >	object,
		T	operand,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_sub.

Calls sub on SYCL atomic object.

Parameters

object	The atomic object to subtract from
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: The old value of *object

Definition at line 604 of file atomic.h.

◆ atomic_fetch_xor()

template<typename T , access::address_space addressSpace>

T cl::sycl::atomic_fetch_xor	(	atomic< T, addressSpace >	object,
		T	operand,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_xor.

Calls XOR on SYCL atomic object.

Parameters

object	The atomic object to XOR with
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: The old value of *object

Definition at line 640 of file atomic.h.

◆ atomic_load()

template<typename T , access::address_space addressSpace>

T cl::sycl::atomic_load	(	atomic< T, addressSpace >	object,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_load.

Calls load on SYCL atomic object.

Parameters

object	The atomic object to load from
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Returns: *object

Definition at line 540 of file atomic.h.

◆ atomic_store()

template<typename T , access::address_space addressSpace>

void cl::sycl::atomic_store	(	atomic< T, addressSpace >	object,
		T	operand,
		memory_order	mem_order = `memory_order::relaxed`
	)

inline

Global function atomic_store.

Calls store on SYCL atomic object.

Parameters

object	The atomic object to store to
mem_order	The memory ordering to use. Only memory_order_relaxed is supported.

Definition at line 551 of file atomic.h.

◆ bitselect()

template<typename T , COMPUTECPP_REQUIRES(detail::builtin::is_gentype< T >::value) >

T cl::sycl::bitselect	(	T	a,
		T	b,
		T	c
	)

noexcept

Each bit of the result is the corresponding bit of a if the corresponding bit of c is 0.

Otherwise it is the corresponding bit of b.

Template Parameters

T	must model gentype.

Definition at line 216 of file math_relational.h.

◆ cbrt()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::cbrt ( F x )

noexcept

Compute cube-root.

Template Parameters

F	must model genfloat.

Definition at line 139 of file math_floating_point.h.

◆ ceil()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::ceil ( F x )

noexcept

Round to integral value using the round to positive infinity rounding mode.

Template Parameters

F	must model genfloat.

Definition at line 149 of file math_floating_point.h.

◆ clamp() [1/3]

template<typename F1 , typename F2 , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && std::is_same< F1, F2 >::value)||(detail::builtin::is_genfloath< F1 >::value && std::is_same< F2, half >::value)||(detail::builtin::is_genfloatf< F1 >::value && std::is_same< F2, float >::value)||(detail::builtin::is_genfloatd< F1 >::value && std::is_same< F2, double >::value)) >

F1 cl::sycl::clamp	(	F1	x,
		F2	minval,
		F2	maxval
	)

noexcept

Returns fmin(fmax(x, minval), maxval).

Results are undefined if minval > maxval.

Template Parameters

F1	must model genfloat.
F2	must model either genfloat or be the scalar equivalent of F1.

Precondition: minval <= maxval

Definition at line 46 of file math_common.h.

◆ clamp() [2/3]

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::clamp	(	Integral	x,
		Integral	minval,
		Integral	maxval
	)

noexcept

Returns min(max(x, minval), maxval).

Results are undefined if minval > maxval.

Template Parameters

Integral must model geninteger.

Precondition: minval <= maxval.

Definition at line 155 of file math_integral.h.

◆ clamp() [3/3]

template<typename Integral , typename S , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value && detail::builtin::is_sgeninteger< S >::value)) >

Integral cl::sycl::clamp	(	Integral	x,
		S	minval,
		S	maxval
	)

noexcept

Returns min(max(x, minval), maxval).

Results are undefined if minval > maxval.

Template Parameters

Integral	must model geninteger
S	must model sgeninteger.

Precondition: minval <= maxval.

Definition at line 171 of file math_integral.h.

◆ clz()

template<typename Integral , typename return_t = detail::collapse_swizzled_vec_t<Integral>, COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

return_t cl::sycl::clz ( Integral x )

noexcept

Returns the number of leading 0-bits in x, starting at the most significant bit position.

Template Parameters

Integral must model geninteger.

Definition at line 184 of file math_integral.h.

◆ COMPUTECPP_DEPRECATED_BY_SYCL_VER() [1/2]

template<typename AllocatorT >

cl::sycl::COMPUTECPP_DEPRECATED_BY_SYCL_VER	(	201703	,
		"Use overload where the range comes before the offset."
	)

Constructs a ranged placeholder accessor.

Template Parameters

AllocatorT Specifies the type of the buffer objects allocator.

Parameters

bufferRef	Buffer object where access is being requested
accessOffset	Point where data access starts
accessRange	Range of data to access

Deprecated:: Need to reverse the order of the access offset and range, see 4.7.6.6 Buffer accessor interface in SYCL 1.2.1

◆ COMPUTECPP_DEPRECATED_BY_SYCL_VER() [2/2]

template<typename functorT , int dimensions>

cl::sycl::COMPUTECPP_DEPRECATED_BY_SYCL_VER	(	201703	,
		"Use group::parallel_for_work_item instead."
	)

inner loop of parallel_for_work_group, the Hierarchical API.

inner loop of parallel_for_work_group with a logical local range, the Hierarchical API.

Deprecated:: Use group::parallel_for_work_item instead

Deprecated:: Use group::parallel_for_work_item instead

◆ COMPUTECPP_GET_INFO_SPECIALIZATION_DECL()

cl::sycl::COMPUTECPP_GET_INFO_SPECIALIZATION_DECL	(	device	,
		usm_restricted_shared_allocations
	)

◆ COMPUTECPP_VALID_KERNEL_ARG_IF() [1/2]

template<typename elemT , int kDims, access::mode kMode, access::target kTarget, access::placeholder IsPlaceholder>

COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT cl::sycl::COMPUTECPP_VALID_KERNEL_ARG_IF	(	kTarget !	= `access::target::host_buffer && kTarget != access::target::host_task`,
		"Cannot pass host accessor to SYCL kernel"
	)

Non-specialized accessor template that covers only global and constant buffers.

Template Parameters

elemT	Underlying data type
kDims	Number of accessor dimensions
kMode	Access mode
kTarget	Access target
isPlaceholder	Whether the accessor is a placeholder. No effect in SYCL-2020.

Constructs a buffer accessor

Template Parameters

AllocatorT	Specifies the type of the buffer objects allocator.
COMPUTECPP_ENABLE_IF	Only enabled for global and constant buffer access, atomic mode not allowed

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Command group handler
propList	Additional properties

Constructs an accessor of access target access::target::global_buffer or access::target::constant_buffer by taking a buffer object, a handler, an offset and a range and initialises the base_accessor with the buffer, the handler, the access target, the access mode and the element size. This constructor is for constructing a sub accessor.

Template Parameters

AllocatorT	Specifies the type of the buffer objects allocator.
COMPUTECPP_ENABLE_IF	Only allows the constructor if the accessor is global, or is constant and not atomic.

Parameters

bufferRef	Reference to the buffer object the accessor is to access data from.
commandHandler	Reference to the handler object for the command group
accessOffset	The offset that the sub accessor should have access from.
accessRange	The range that the sub accessor should have access to. scope that the accessor is being constructed within.

Deprecated:: Need to reverse the order of the access offset and range, see 4.7.6.6 Buffer accessor interface in SYCL 1.2.1

Constructs a ranged buffer accessor

Template Parameters

AllocatorT	Type of the buffer allocator
COMPUTECPP_ENABLE_IF	Only allowed for global and constant buffer accessors which are not 0-dimensional

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Command group handler
accessRange	Range of data to access
propList	Additional properties

Constructs a ranged buffer accessor with an offset

Template Parameters

AllocatorT	Type of the buffer objects allocator
COMPUTECPP_ENABLE_IF	Only allows the constructor if the accessor is global, or is constant and not atomic, and is not 0-dimensional

Parameters

bufferRef	Buffer object where access is being requested
commandHandler	Command group handler
accessRange	Range of data to access
accessOffset	Offset from the beginning of the buffer
propList	Additional properties

Constructs an accessor from a storage_mem} object. Used to create normal accessors from placeholder ones.

Constructs a ranged placeholder accessor

Parameters

store	Storage object representing the buffer/image
accessRange	the offset and range that this accessor can access.
propList	Additional properties

Returns an accessor similar to *this, with the offset changed to this->get_offset()[0] + addedOffset. A call to this function will fail with ACCESSOR_ARGUMENTS_ERROR if the created accessor has an offset less than 0 or if the new offset plus the current access range exceed this->get_range()[0].

Definition at line 464 of file buffer_accessor.h.

◆ COMPUTECPP_VALID_KERNEL_ARG_IF() [2/2]

template<typename elemT , int kDims, access::mode kMode, access::target kTarget>

COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT cl::sycl::COMPUTECPP_VALID_KERNEL_ARG_IF	(	kTarget !	= `access::target::host_buffer && kTarget != access::target::host_task`,
		"Cannot pass host accessor to SYCL kernel"
	)

A public facing accessor that can be constructed outside of a command group.

Even though it can be constructed, it cannot be accessed outside of a command group. Before it can be accessed, it has to be registered in a command group handler.

This is an extension of the SYCL specification.

The reason this is a specialization instead of using enable_if in one class is that COMPUTECPP_CONVERT_ATTR has to be placed on the regular accessor class and COMPUTECPP_CONVERT_ATTR_PLACEHOLDER on the placeholder one.

In SYCL-2020, buffer accessors accept IsPlaceholder template parameters, but the value of the parameter has no bearing on whether this accessor is a placeholder. Instead, this is determined by the constructor. Consequently, this specialization is ignored in SYCL-2020 onwards.

Template Parameters

elemT	Underlying data type
kDims	Number of accessor dimensions
kMode	Access mode
kTarget	Access target

◆ copysign()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::copysign	(	F1	x,
		F2	y
	)

noexcept

Returns x with its sign changed to match the sign of y.

Template Parameters

F	must model genfloat.

Definition at line 161 of file math_floating_point.h.

◆ cos()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::cos ( F x )

noexcept

Compute cosine.

Template Parameters

F	must model genfloat.

Definition at line 170 of file math_floating_point.h.

◆ cosh()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::cosh ( F x )

noexcept

Compute hyperbolic cosine.

Template Parameters

F	must model genfloat.

Definition at line 179 of file math_floating_point.h.

◆ cospi()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::cospi ( F x )

noexcept

Compute cos(πx)

Template Parameters

F	must model genfloat.

Definition at line 188 of file math_floating_point.h.

◆ cross()

template<typename F , int N, COMPUTECPP_REQUIRES((detail::is_custom_half_type< F >::value|| std::is_same< F, float >::value|| std::is_same< F, double >::value) &&(N==3||N==4)) >

vec<F, N> cl::sycl::cross	(	const vec< F, N >	p0,
		const vec< F, N >	p1
	)

noexcept

Returns the cross product of p0.xyz and p1.xyz.

The w component of float4 result returned will be 0.0.

Template Parameters

F	must be half, float, or double.
N	N == 3 or N == 4

Definition at line 44 of file math_geometric.h.

◆ degrees()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::degrees ( const F x )

noexcept

Converts radians to degrees, i.e.

(180 / π) * x.

Template Parameters

F	must model genfloat.

Definition at line 56 of file math_common.h.

◆ device::create_sub_devices< info::partition_property::partition_by_affinity_domain >()

template<>

COMPUTECPP_EXPORT vector_class<device> cl::sycl::device::create_sub_devices< info::partition_property::partition_by_affinity_domain > ( info::partition_affinity_domain affinityDomain ) const

Template specialisation of create_sub_devices() for info::partition_property::partition_by_affinity_domain.

◆ device::create_sub_devices< info::partition_property::partition_by_counts >()

template<>

COMPUTECPP_EXPORT vector_class<device> cl::sycl::device::create_sub_devices< info::partition_property::partition_by_counts > ( const vector_class< size_t > & counts ) const

Template specialisation of create_sub_devices() for info::partition_property::partition_by_counts.

◆ device::create_sub_devices< info::partition_property::partition_equally >()

template<>

COMPUTECPP_EXPORT vector_class<device> cl::sycl::device::create_sub_devices< info::partition_property::partition_equally > ( size_t nbSubDev ) const

Template specialisation of create_sub_devices() for info::partition_property::partition_equally.

◆ device::get_info< info::device::name >()

template<>

sub_group_independent_forward_progress info::param_traits<info::device, info::device::name>::return_type cl::sycl::device::get_info< info::device::name > ( ) const

inline

Gets OpenCL information for the underlying cl_device_id.

Template Parameters

param A cl_int specifying the info parameter.

Returns: The retrieved information as the appropriate return type, derived via the get_sycl_info function, defined in info.h.

Definition at line 367 of file device.h.

◆ device::get_info< info::device::vendor >()

template<>

info::param_traits<info::device, info::device::vendor>::return_type cl::sycl::device::get_info< info::device::vendor > ( ) const

inline

Gets OpenCL information for the underlying cl_device_id.

Template Parameters

param A cl_int specifying the info parameter.

Returns: The retrieved information as the appropriate return type, derived via the get_sycl_info function, defined in info.h.

Definition at line 375 of file device.h.

◆ device::get_info< info::device::version >()

template<>

info::param_traits<info::device, info::device::version>::return_type cl::sycl::device::get_info< info::device::version > ( ) const

inline

Gets OpenCL information for the underlying cl_device_id.

Template Parameters

param A cl_int specifying the info parameter.

Returns: The retrieved information as the appropriate return type, derived via the get_sycl_info function, defined in info.h.

Definition at line 383 of file device.h.

◆ distance()

template<typename F1 , typename F2 , typename F = detail::select_float_t<detail::common_return_t<F1, F2>>, COMPUTECPP_REQUIRES((detail::builtin::is_gen_geo_anyfloat< F1 >::value && detail::builtin::is_gen_geo_anyfloat< F2 >::value)) >

F cl::sycl::distance	(	F1	p0,
		F2	p1
	)

noexcept

Returns the distance between p0 and p1.

This is calculated as length(p0 - p1).

Template Parameters

F	must model gengeohalf, gengeofloat, gengeodouble.

Definition at line 79 of file math_geometric.h.

◆ dot()

template<typename F1 , typename F2 , typename F = detail::select_float_t<detail::common_return_t<F1, F2>>, COMPUTECPP_REQUIRES((detail::builtin::is_gen_geo_anyfloat< F1 >::value && detail::builtin::is_gen_geo_anyfloat< F2 >::value)) >

F cl::sycl::dot	(	F1	p0,
		F2	p1
	)

noexcept

Compute dot product.

Template Parameters

F	must model gengeohalf, gengeofloat, gengeodouble.

Definition at line 65 of file math_geometric.h.

◆ erf()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::erf ( F x )

noexcept

Error function encountered in integrating the normal distribution.

Template Parameters

F	must model genfloat.

Definition at line 206 of file math_floating_point.h.

◆ erfc()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::erfc ( F x )

noexcept

Complementary error function.

Template Parameters

F	must model genfloat.

Definition at line 197 of file math_floating_point.h.

◆ exp()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::exp ( F x )

noexcept

Compute the base-e exponential of x.

Template Parameters

F	must model genfloat.

Definition at line 215 of file math_floating_point.h.

◆ exp10()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::exp10 ( F x )

noexcept

Exponential base 10 function.

Template Parameters

F	must model genfloat.

Definition at line 233 of file math_floating_point.h.

◆ exp2()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::exp2 ( F x )

noexcept

Exponential base 2 function.

Template Parameters

F	must model genfloat.

Definition at line 224 of file math_floating_point.h.

◆ expm1()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::expm1 ( F x )

noexcept

Compute exp(x) - 1.0.

Template Parameters

F	must model genfloat.

Definition at line 242 of file math_floating_point.h.

◆ fabs()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::fabs ( F x )

noexcept

Compute absolute value of a floating-point number.

Template Parameters

F	must model genfloat.

Definition at line 251 of file math_floating_point.h.

◆ fast_distance()

template<typename F1 , typename F2 , typename F = detail::scalar_t<detail::common_return_t<F1, F2>>, COMPUTECPP_REQUIRES((computecpp::gsl::or_< detail::builtin::is_gengeohalf< F1 >::value, detail::builtin::is_gengeofloat< F1 >::value >::value) &&(computecpp::gsl::or_< detail::builtin::is_gengeohalf< F2 >::value, detail::builtin::is_gengeofloat< F2 >::value >::value)) >

F cl::sycl::fast_distance	(	F1	p0,
		F2	p1
	)

noexcept

Returns f ast length(p0 - p1).

Template Parameters

F	must model gengeofloat.

Definition at line 116 of file math_geometric.h.

◆ fast_length()

template<typename F , typename return_t = detail::scalar_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_gengeohalf< F >::value, detail::builtin::is_gengeofloat< F >::value >::value) >

return_t cl::sycl::fast_length ( F p )

noexcept

Returns the length of vector p computed as: sqrt((half)(pow(p.x,2) + pow(p.y,2)+ ...)).

Template Parameters

F	must model gengeofloat.

Definition at line 130 of file math_geometric.h.

◆ fast_normalize()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_gengeohalf< F >::value, detail::builtin::is_gengeofloat< F >::value >::value) >

return_t cl::sycl::fast_normalize ( F p )

noexcept

Returns a vector in the same direction as p but with a length of 1.

fast_normalize is computed as: p * rsqrt((half)(pow(p.x, 2) + pow(p.y, 2) + ...)).

The result shall be within 8192 ulps error from the infinitely precise result of

if (all (p == 0.0f))
   result = p;
else
result = p / sqrt (pow(p.x,2)+ pow(p.y,2)+ ... );

with the following exceptions:

If the sum of squares is greater than FLT_MAX then the value of the floating-point values in the result vector are undefined.
If the sum of squares is less than FLT_MIN then the implementation may return back p.
If the device is in "denorms are flushed to zero" mode, individual operand elements with magnitude less than sqrt(FLT_MIN) may be flushed to zero before proceeding with the calculation.
Template Parameters

F must model gengeofloat.

Definition at line 164 of file math_geometric.h.

◆ fdim()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::fdim	(	F1	x,
		F2	y
	)

noexcept

x - y if x > y, +0 if x is less than or equal to y.

Template Parameters

F	must model genfloat.

Definition at line 263 of file math_floating_point.h.

◆ floor()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::floor ( F x )

noexcept

Round to integral value using the round to negative infinity rounding mode.

Template Parameters

F	must model genfloat.

Definition at line 273 of file math_floating_point.h.

◆ fma()

template<typename F1 , typename F2 , typename F3 , typename F = detail::common_return_t<F1, F2, F3>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value && detail::builtin::is_genfloat< F3 >::value)) >

F cl::sycl::fma	(	F1	a,
		F2	b,
		F3	c
	)

noexcept

Returns the correctly rounded floating-point representation of the sum of c with the infinitely precise product of a and b.

Rounding of intermediate products shall not occur. Edge case behavior is per the IEEE 754-2008 standard.

Template Parameters

F	must model genfloat.

Definition at line 289 of file math_floating_point.h.

◆ fmax()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES( detail::builtin::is_genfloat< F1 >::value &&(detail::builtin::is_genfloat< F2 >::value|| detail::builtin::is_sgenfloat< F2, detail::scalar_t< F >>::value)) >

F cl::sycl::fmax	(	F1	x,
		F2	y
	)

noexcept

Returns y if x < y, otherwise it returns x.

If one argument is a NaN, fmax() returns the other argument. If both arguments are NaNs, fmax() returns a NaN.

Template Parameters

F	must model genfloat.

Definition at line 305 of file math_floating_point.h.

◆ fmin()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES( detail::builtin::is_genfloat< F1 >::value &&(detail::builtin::is_genfloat< F2 >::value|| detail::builtin::is_sgenfloat< F2, detail::scalar_t< F >>::value)) >

F cl::sycl::fmin	(	F1	x,
		F2	y
	)

noexcept

Returns y if y < x, otherwise it returns x.

If one argument is a NaN, fmin() returns the other argument. If both arguments are NaNs, fmin() returns a NaN.

Template Parameters

F	must model genfloat.

Definition at line 321 of file math_floating_point.h.

◆ fmod()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::fmod	(	F1	x,
		F2	y
	)

noexcept

Modulus.

Returns x * y ∗ trunc(x/y).

Template Parameters

F	must model genfloat.

Definition at line 333 of file math_floating_point.h.

◆ fract()

template<typename F , access::address_space AddressSpace, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

F cl::sycl::fract	(	F	x,
		detail::builtin_ptr< F, AddressSpace >	iptr
	)

noexcept

Returns fmin(x - floor (x), 0x1.fffffep-1f).

floor(x) is returned in iptr.

Template Parameters

F	must model genfloat.
AddressSpace	The address space for which the pointer may address. Cannot be constant_space.

Definition at line 359 of file math_floating_point.h.

◆ frexp()

template<typename F , typename Integral , access::address_space AddressSpace, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >

F cl::sycl::frexp	(	F	x,
		detail::builtin_ptr< Integral, AddressSpace >	exp
	)

noexcept

Extract mantissa and exponent from x.

For each component the mantissa returned is a float with magnitude in the interval [1/2, 1) or 0. Each component of x equals mantissa returned * 2exp.

Template Parameters

F	must model genfloat.
AddressSpace	The address space for which the pointer may address. Cannot be constant_space.

Definition at line 374 of file math_floating_point.h.

◆ get_access()

accessor<elemT, kDims, kMode, kTarget, access::placeholder::false_t> cl::sycl::get_access ( handler & commandHandler ) const

Obtains a normal accessor from the placeholder accessor.

Parameters

commandHandler Command group handler where the accessor will be used

Returns: Normal accessor that does not need to be registered

Definition at line 1139 of file buffer_accessor.h.

◆ get_accessor_with_added_offset()

accessor<elemT, 1, kMode, kTarget, access::placeholder::true_t> cl::sycl::get_accessor_with_added_offset ( int addedOffset ) const

protected

returns an accessor similar to *this, with the offset changed to this->get_offset()[0] + addedOffset.

A call to this function will fail with ACCESSOR_ARGUMENTS_ERROR if the created accessor has an offset less than 0 or if the new offset plus the current access range exceed this->get_range()[0].

Definition at line 942 of file buffer_accessor.h.

◆ hadd()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::hadd	(	Integral	x,
		Integral	y
	)

noexcept

Returns (x + y) >> 1.

The intermediate sum does not modulo overflow.

Template Parameters

Integral must model geninteger.

Definition at line 134 of file math_integral.h.

◆ hypot()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::hypot	(	F1	x,
		F2	y
	)

noexcept

Compute the value of the square root of x2 + y2 without undue overflow or underflow.

Template Parameters

F	must model genfloat.

Definition at line 387 of file math_floating_point.h.

◆ ilogb()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::correct_int_t<F> cl::sycl::ilogb ( F x )

noexcept

Return the exponent as an integer value.

Template Parameters

F	must model genfloat.

Definition at line 411 of file math_floating_point.h.

◆ isequal()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::isequal	(	F	x,
		F	y
	)

noexcept

Returns the component-wise compare of x == y.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 37 of file math_relational.h.

◆ isfinite()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::isfinite ( F x )

noexcept

Test for finite value.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 109 of file math_relational.h.

◆ isgreater()

template<typename F1 , typename F2 , typename F = detail::matching_integral_t<detail::common_return_t<F1, F2>>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::isgreater	(	F1	x,
		F2	y
	)

noexcept

Returns the component-wise compare of x > y.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 60 of file math_relational.h.

◆ isgreaterequal()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::isgreaterequal	(	F	x,
		F	y
	)

noexcept

Returns the component-wise compare of x >= y.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 69 of file math_relational.h.

◆ isinf()

template<typename F , typename return_t = detail::matching_integral_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::isinf ( F x )

noexcept

Test for infinity value (positive or negative).

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 121 of file math_relational.h.

◆ isless()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::isless	(	F	x,
		F	y
	)

noexcept

Returns the component-wise compare of x < y.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 79 of file math_relational.h.

◆ islessequal()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::islessequal	(	F	x,
		F	y
	)

noexcept

Returns the component-wise compare of x <= y.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 89 of file math_relational.h.

◆ islessgreater()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::islessgreater	(	F	x,
		F	y
	)

noexcept

Returns the component-wise compare of (x < y) || (x > y).

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 99 of file math_relational.h.

◆ isnan()

template<typename F , typename return_t = detail::matching_integral_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::isnan ( F x )

noexcept

Test for a NaN.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 132 of file math_relational.h.

◆ isnormal()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::isnormal ( F x )

noexcept

Test for a normal value.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 141 of file math_relational.h.

◆ isnotequal()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::isnotequal	(	F	x,
		F	y
	)

noexcept

Returns the component-wise compare of x != y.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 47 of file math_relational.h.

◆ isordered()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::isordered	(	F	x,
		F	y
	)

noexcept

Test if arguments are ordered.

isordered() takes arguments x and y, and returns the result isequal(x, x) && isequal(y, y).

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 154 of file math_relational.h.

◆ isunordered()

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

detail::matching_integral_t<F> cl::sycl::isunordered	(	F	x,
		F	y
	)

noexcept

Test if arguments are unordered.

isunordered() takes arguments x and y, returning non-zero if x or y is NaN, and zero otherwise.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 167 of file math_relational.h.

◆ ldexp() [1/2]

template<typename F , typename Integral , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >

F cl::sycl::ldexp	(	F	x,
		Integral	k
	)

noexcept

Multiply x by 2 to the power k.

Template Parameters

F	must model genfloat.
Integral	must model genint.

Definition at line 422 of file math_floating_point.h.

◆ ldexp() [2/2]

template<typename F , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

F cl::sycl::ldexp	(	F	x,
		const int	k
	)

noexcept

Multiply x by 2 to the power k.

Template Parameters

F	must model genfloat.

Definition at line 431 of file math_floating_point.h.

◆ length()

template<typename F , typename return_t = detail::select_float_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES(detail::builtin::is_gen_geo_anyfloat< F >::value) >

return_t cl::sycl::length ( F p )

noexcept

Return the length of vector p, i.e.

sqrt((p.x * p.x) + (p.y * p.y) + ...)

Template Parameters

F	must model gengeohalf, gengeofloat, gengeodouble.

Definition at line 91 of file math_geometric.h.

◆ lgamma()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::lgamma ( F x )

noexcept

Log gamma function.

Returns the natural logarithm of the absolute value of the gamma function. The sign of the gamma function is returned in the signp argument of lgamma_r.

Template Parameters

F	must model genfloat.

Definition at line 443 of file math_floating_point.h.

◆ lgamma_r()

template<typename F , typename Integral , access::address_space AddressSpace, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >

F cl::sycl::lgamma_r	(	F	x,
		detail::builtin_ptr< Integral, AddressSpace >	signp
	)

noexcept

Log gamma function.

Returns the natural logarithm of the absolute value of the gamma function. The sign of the gamma function is returned in the signp argument of lgamma_r.

Template Parameters

F	must model genfloat.

Definition at line 456 of file math_floating_point.h.

◆ log()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::log ( F x )

noexcept

Compute natural logarithm.

Template Parameters

F	must model genfloat.

Definition at line 465 of file math_floating_point.h.

◆ log10()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::log10 ( F x )

noexcept

Compute a base 10 logarithm.

Template Parameters

F	must model genfloat.

Definition at line 483 of file math_floating_point.h.

◆ log1p()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::log1p ( F x )

noexcept

Compute loge(1.0 + x).

Template Parameters

F	must model genfloat.

Definition at line 492 of file math_floating_point.h.

◆ log2()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::log2 ( F x )

noexcept

Compute a base 2 logarithm.

Template Parameters

F	must model genfloat.

Definition at line 474 of file math_floating_point.h.

◆ logb()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::logb ( F x )

noexcept

Compute the exponent of x, which is the integral part of logr (|x|).

Template Parameters

F	must model genfloat.

Definition at line 501 of file math_floating_point.h.

◆ mad()

template<typename F1 , typename F2 , typename F3 , typename F = detail::common_return_t<F1, F2, F3>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value && detail::builtin::is_genfloat< F3 >::value)) >

F cl::sycl::mad	(	F1	a,
		F2	b,
		F3	c
	)

noexcept

mad approximates a * b + c.

Whether or how the product of a * b is rounded and how supernormal or subnormal intermediate products are handled is not defined. mad is intended to be used where speed is preferred over accuracy.

Template Parameters

F	must model genfloat.

Definition at line 517 of file math_floating_point.h.

◆ mad24()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger32bit< Integral >::value) >

Integral cl::sycl::mad24	(	Integral	x,
		Integral	y,
		Integral	z
	)

noexcept

Multipy two 24-bit integer values x and y and add the 32-bit integer result to the 32-bit integer z.

Refer to definition of mul24 to see how the 24-bit integer multiplication is performed.

Template Parameters

Integral must model geninteger32bit.

Definition at line 341 of file math_integral.h.

◆ mad_hi()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::mad_hi	(	Integral	a,
		Integral	b,
		Integral	c
	)

noexcept

Returns mul_hi(a, b) + c.

Template Parameters

Integral must model geninteger.

Definition at line 211 of file math_integral.h.

◆ mad_sat()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::mad_sat	(	Integral	a,
		Integral	b,
		Integral	c
	)

noexcept

Returns a * b + c and saturates the result.

Template Parameters

Integral must model geninteger.

Definition at line 220 of file math_integral.h.

◆ make_ptr() [1/2]

template<typename dataType , cl::sycl::access::address_space Space>

std::enable_if< !std::is_same<typename multi_ptr<dataType, Space>::ptr_t, dataType*>::value, multi_ptr<dataType, Space> >::type cl::sycl::make_ptr ( dataType * ptr )

Create a multi_ptr object from a raw pointer.

Template Parameters

dataType	Type of the pointed values.
Space	Address space in which the data live.

Parameters

ptr	The raw pointer from which to create the multi_ptr.

Returns: A multi_ptr object pointing to the same address pointed as by ptr.

Definition at line 1182 of file multi_pointer.h.

◆ make_ptr() [2/2]

template<typename dataType , cl::sycl::access::address_space Space>

multi_ptr<dataType, Space> cl::sycl::make_ptr ( typename multi_ptr< dataType, Space >::ptr_t ptr )

Definition at line 1187 of file multi_pointer.h.

◆ max() [1/3]

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)||(detail::builtin::is_genfloath< F1 >::value && detail::builtin::is_sgenfloat< F2, half >::value)||(detail::builtin::is_genfloatf< F1 >::value && detail::builtin::is_sgenfloat< F2, float >::value)||(detail::builtin::is_genfloatd< F1 >::value && detail::builtin::is_sgenfloat< F2, double >::value)) >

F cl::sycl::max	(	F1	x,
		F2	y
	)

noexcept

Returns y if x < y, otherwise it returns x.

If x or y are infinite or NaN, the return values are undefined.

Template Parameters

F1	must model genfloat.
F2	must model either genfloat or be the scalar equivalent of F1.

Definition at line 93 of file math_common.h.

◆ max() [2/3]

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::max	(	Integral	x,
		Integral	y
	)

noexcept

Returns y if x < y, otherwise it returns x.

Template Parameters

Integral must model geninteger.

Definition at line 229 of file math_integral.h.

◆ max() [3/3]

template<typename Integral , typename S , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value && detail::builtin::is_sgeninteger< S >::value)) >

Integral cl::sycl::max	(	Integral	x,
		S	y
	)

noexcept

Returns y if x < y, otherwise it returns x.

Template Parameters

Integral	must model geninteger.
S	must model sgeninteger.

Definition at line 241 of file math_integral.h.

◆ maxmag()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::maxmag	(	F1	x,
		F2	y
	)

noexcept

Returns x if |x| > |y|, y if |y| > |x|, otherwise fmax(x, y).

Template Parameters

F	must model genfloat.

Definition at line 528 of file math_floating_point.h.

◆ min() [1/3]

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)||(detail::builtin::is_genfloath< F1 >::value && detail::builtin::is_sgenfloat< F2, half >::value)||(detail::builtin::is_genfloatf< F1 >::value && detail::builtin::is_sgenfloat< F2, float >::value)||(detail::builtin::is_genfloatd< F1 >::value && detail::builtin::is_sgenfloat< F2, double >::value)) >

F cl::sycl::min	(	F1	x,
		F2	y
	)

noexcept

Returns y if x > y, otherwise it returns x.

If x or y are infinite or NaN, the return values are undefined.

Template Parameters

F1	must model genfloat.
F2	must model either genfloat or be the scalar equivalent of F1.

Definition at line 112 of file math_common.h.

◆ min() [2/3]

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::min	(	Integral	x,
		Integral	y
	)

noexcept

Returns y if y < x, otherwise it returns x.

Template Parameters

Integral must model geninteger.

Definition at line 250 of file math_integral.h.

◆ min() [3/3]

template<typename Integral , typename S , COMPUTECPP_REQUIRES((detail::builtin::is_geninteger< Integral >::value && detail::builtin::is_sgeninteger< S >::value)) >

Integral cl::sycl::min	(	Integral	x,
		S	y
	)

noexcept

Returns y if y < x, otherwise it returns x.

Template Parameters

Integral	must model geninteger.
S	must model sgeninteger.

Definition at line 262 of file math_integral.h.

◆ minmag()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::minmag	(	F1	x,
		F2	y
	)

noexcept

Returns x if |x| < |y|, y if |y| < |x|, otherwise fmin(x, y).

Template Parameters

F	must model genfloat.

Definition at line 540 of file math_floating_point.h.

◆ mix()

template<typename F1 , typename F2 , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && std::is_same< F1, F2 >::value)||(detail::builtin::is_genfloath< F1 >::value && std::is_same< F2, half >::value)||(detail::builtin::is_genfloatf< F1 >::value && std::is_same< F2, float >::value)||(detail::builtin::is_genfloatd< F1 >::value && std::is_same< F2, double >::value)) >

F1 cl::sycl::mix	(	F1	x,
		F1	y,
		F2	a
	)

noexcept

Returns the linear blend of x&y implemented as: x + (y - x) * a.

a must be a value in the range 0.0 ... 1.0. If a is not in the range 0.0 ... 1.0, the return values are undefined.

Template Parameters

F1	must model genfloat.
F2	must model either genfloat or be the scalar equivalent of T1.

Precondition: 0.0 <= a <= 1.0.

Definition at line 132 of file math_common.h.

◆ modf()

template<typename F , access::address_space AddressSpace, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

F cl::sycl::modf	(	F	x,
		detail::builtin_ptr< F, AddressSpace >	iptr
	)

noexcept

Decompose a floating-point number.

The modf function breaks the argument x into integral and fractional parts, each of which has the same sign as the argument. It stores the integral part in the object pointed to by iptr.

Template Parameters

F	must model genfloat.
AddressSpace	The address space for which the pointer may address. Cannot be constant_space.

Definition at line 555 of file math_floating_point.h.

◆ mul24()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger32bit< Integral >::value) >

Integral cl::sycl::mul24	(	Integral	x,
		Integral	y
	)

noexcept

Multiply two 24-bit integer values x and y.

x and y are 32-bit integers but only the low 24-bits are used to perform the multiplication.

mul24 should only be used when values in x and y are in the range [-223, 223 - 1] if x and y are signed integers and in the range [0, 224 - 1] if x and y are unsigned integers. If x and y are not in this range, the multiplication result is implementation-defined.

Template Parameters

Integral must model geninteger32bit.

Definition at line 358 of file math_integral.h.

◆ mul_hi()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::mul_hi	(	Integral	x,
		Integral	y
	)

noexcept

Computes x * y and returns the high half of the product of x and y.

Template Parameters

Integral must model geninteger.

Definition at line 272 of file math_integral.h.

◆ nan() [1/2]

template<typename Integral , COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_ugenint< Integral >::value, detail::builtin::is_ugenlonginteger< Integral >::value >::value) >

auto cl::sycl::nan ( Integral nancode ) -> detail::matching_float_t<Integral>

noexcept

Returns x if |x| < |y|, y if |y| < |x|, otherwise fmin(x, y).

Template Parameters

F	must model genfloatf or genfloatd.
Integral	If F models genfloatf, then Integral must model ugenint. If F models genfloatd, then Integral must model ugenlonginteger.

Definition at line 612 of file math_floating_point.h.

◆ nan() [2/2]

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_ushortn< Integral >::value) >

detail::matching_float_t<Integral> cl::sycl::nan ( Integral nancode )

noexcept

Definition at line 619 of file math_floating_point.h.

◆ nextafter()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::nextafter	(	F1	x,
		F2	y
	)

noexcept

Computes the next representable single-precision floating-point value following x in the direction of y.

Thus, if y is less than x, nextafter() returns the largest representable floating-point number less than x.

Template Parameters

F	must model genfloat.

Definition at line 634 of file math_floating_point.h.

◆ normalize()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_gen_geo_anyfloat< F >::value) >

return_t cl::sycl::normalize ( F p )

noexcept

Returns a vector in the same direction as p but with a length of 1.

Template Parameters

F	must model gengeohalf, gengeofloat, gengeodouble.

Definition at line 100 of file math_geometric.h.

◆ operator+()

template<COMPUTECPP_ENABLE_IF(elemT,(kDims==1)) >

accessor<elemT, 1, kMode, kTarget, access::placeholder::true_t> cl::sycl::operator+ ( int rhs ) const

Creates and returns a new accessor with its offset changed by rhs.

Parameters

rhs	the offset to add to the current accessor's offset in the new accessor

Returns: A new accessor with its get_offset()[0] equal to this->get_offset()[0] + rhs

Definition at line 1152 of file buffer_accessor.h.

◆ operator+=()

template<COMPUTECPP_ENABLE_IF(elemT,(kDims==1)) >

accessor<elemT, 1, kMode, kTarget, access::placeholder::true_t>& cl::sycl::operator+= ( int rhs )

Changes this->offset by rhs as this->get_offset()[0] + rhs.

Parameters

rhs	the offset to add to the current accessor's offset

Returns: This accessor

Definition at line 1174 of file buffer_accessor.h.

◆ operator-()

template<COMPUTECPP_ENABLE_IF(elemT,(kDims==1)) >

accessor<elemT, 1, kMode, kTarget, access::placeholder::true_t> cl::sycl::operator- ( int rhs ) const

Creates and returns a new accessor with its offset changed by rhs.

Parameters

rhs	the offset to subtract to the current accessor's offset in the new accessor

Returns: A new accessor with its get_offset()[0] equal to this->get_offset()[0] - rhs

Definition at line 1164 of file buffer_accessor.h.

◆ operator-=()

template<COMPUTECPP_ENABLE_IF(elemT,(kDims==1)) >

accessor<elemT, 1, kMode, kTarget, access::placeholder::true_t>& cl::sycl::operator-= ( int rhs )

Changes this->offset by rhs as this->get_offset()[0] - rhs.

Parameters

rhs	the offset to subtract to the current accessor's offset

Returns: This accessor

Definition at line 1185 of file buffer_accessor.h.

◆ operator<<() [1/4]

COMPUTECPP_EXPORT std::ostream& cl::sycl::operator<<	(	std::ostream &	out,
		const stream &	rhs
	)

Stream operator that streams a stream object to an ostream.

Friend function declaration to give global stream operator access to private members.

Parameters

out	The output stream.
rhs	The stream object to be streamed into the output.

◆ operator<<() [2/4]

template<typename valueT >

const stream& cl::sycl::operator<<	(	const stream &	os,
		valueT	rhs
	)

inline

Template stream operator that streams a value to a stream object.

Template Parameters

The	type of the value to be streamed.

Parameters

os	The stream object to be streamed to.
rhs	The value to be streamed.

Definition at line 1457 of file stream.h.

◆ operator<<() [3/4]

template<typename elementT , int kDimensions>

const stream& cl::sycl::operator<<	(	const stream &	os,
		const vec< elementT, kDimensions >	rhs
	)

Template stream operator that streams a vec object to a stream object.

Template Parameters

elementT	The element type of the vec.
kDimensions	The number of dimensions of the vec.

Parameters

os	The stream object to be streamed to.
rhs	The vec object to be streamed.

Definition at line 1470 of file stream.h.

◆ operator<<() [4/4]

template<typename elementT , int kDimensions, int... swizzleIndexes>

const stream& cl::sycl::operator<<	(	const stream &	os,
		const swizzled_vec< elementT, kDimensions, swizzleIndexes... >	rhs
	)

Streams a swizzled_vec object to a stream object.

Template Parameters

elementT	The element type of the swizzled_vec
kDimensions	The number of dimensions of the swizzled_vec

Parameters

os	The stream object to be streamed to
rhs	The swizzled_vec object to be streamed

Returns: Original stream object

Definition at line 1510 of file stream.h.

◆ platform::get_info< info::platform::name >()

template<>

info::param_traits<info::platform, info::platform::name>::return_type cl::sycl::platform::get_info< info::platform::name > ( ) const

inline

Get OpenCL information for the underlying cl_platform_id.

Template Parameters

param The info::platform descriptor parameter.

Returns: The retrieved information as the appropriate return type, derived via the param_traits struct, defined in param_traits.h.

Definition at line 303 of file platform.h.

◆ platform::get_info< info::platform::vendor >()

template<>

info::param_traits<info::platform, info::platform::vendor>::return_type cl::sycl::platform::get_info< info::platform::vendor > ( ) const

inline

Get OpenCL information for the underlying cl_platform_id.

Template Parameters

param The info::platform descriptor parameter.

Returns: The retrieved information as the appropriate return type, derived via the param_traits struct, defined in param_traits.h.

Definition at line 311 of file platform.h.

◆ popcount()

template<typename Integral , typename return_t = detail::collapse_swizzled_vec_t<Integral>, COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

return_t cl::sycl::popcount ( Integral x )

noexcept

Returns the number of non-zero bits in x.

Template Parameters

Integral must model geninteger.

Definition at line 327 of file math_integral.h.

◆ pow()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::pow	(	F1	x,
		F2	y
	)

noexcept

Compute x to the power y.

Template Parameters

F	must model genfloat.

Definition at line 646 of file math_floating_point.h.

◆ pown()

template<typename F , typename Integral , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >

F cl::sycl::pown	(	F	x,
		Integral	y
	)

noexcept

Compute x to the power y, where y is an integer.

Template Parameters

F	must model genfloat.
Integral	must model genint.

Definition at line 657 of file math_floating_point.h.

◆ powr()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::powr	(	F1	x,
		F2	y
	)

noexcept

Compute x to the power y, where x >= 0.

Template Parameters

F	must model genfloat.

Precondition: x >= 0

Definition at line 669 of file math_floating_point.h.

◆ radians()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::radians ( const F x )

noexcept

Converts degrees to radians, i.e.

(π / 180) * x.

Template Parameters

F	must model genfloat.

Definition at line 141 of file math_common.h.

◆ remainder()

template<typename F1 , typename F2 , typename F = detail::common_return_t<F1, F2>, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && detail::builtin::is_genfloat< F2 >::value)) >

F cl::sycl::remainder	(	F1	x,
		F2	y
	)

noexcept

Compute the value r such that r = x - n * y, where n is the integer nearest the exact value of x / y.

If there are two integers closest to x / y, n shall be the even one. If r is zero, it is given the same sign as x.

Template Parameters

F	must model genfloat.

Definition at line 686 of file math_floating_point.h.

◆ remquo()

template<typename F , typename Integral , access::address_space AddressSpace, COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F >::value && detail::builtin::is_genint< Integral >::value)) >

F cl::sycl::remquo	(	F	x,
		F	y,
		detail::builtin_ptr< Integral, AddressSpace >	quo
	)

noexcept

The remquo function computes the value r such that r = x - k * y, where k is the integer nearest the exact value of x/y.

If there are two integers closest to x/y, k shall be the even one. If r is zero, it is given the same sign as x. This is the same value that is returned by the remainder function. remquo also calculates the lower seven bits of the integral quotient x/y, and gives that value the same sign as x/y. It stores this signed value in the object pointed to by quo.

Template Parameters

F	must model genfloat.
Integral	must model genint.
AddressSpace	The address space for which the pointer may address. Cannot be constant_space.

Definition at line 706 of file math_floating_point.h.

◆ rhadd()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::rhadd	(	Integral	x,
		Integral	y
	)

noexcept

Returns (x + y + 1) >> 1.

The intermediate sum does not modulo overflow.

Template Parameters

Integral must model geninteger.

Definition at line 144 of file math_integral.h.

◆ rint()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::rint ( F x )

noexcept

Round to integral value (using round to nearest even rounding mode) in floating-point format.

Refer to section 7.1 of the OpenCL 1.2 specification document for description of rounding modes.

Template Parameters

F	must model genfloat.

Definition at line 717 of file math_floating_point.h.

◆ rootn()

template<typename F , typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value, detail::builtin::is_genint< Integral >::value) >

F cl::sycl::rootn	(	F	x,
		Integral	y
	)

noexcept

Compute x to the power 1/y.

Template Parameters

F	must model genfloat.
Integral	must model genint.

Definition at line 728 of file math_floating_point.h.

◆ rotate()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::rotate	(	Integral	v,
		Integral	i
	)

noexcept

For each element in v, the bits are shifted left by the number of bits given by the corresponding element in i (subject to usual shift modulo rules described in section 6.3).

Bits shifted off the left side of the element are shifted back in from the right.

Template Parameters

Integral must model geninteger.

Definition at line 286 of file math_integral.h.

◆ round()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::round ( F x )

noexcept

Return the integral value nearest to x rounding halfway cases away from zero, regardless of the current rounding direction.

Template Parameters

F	must model genfloat.

Definition at line 738 of file math_floating_point.h.

◆ rsqrt()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::rsqrt ( F x )

noexcept

Compute inverse square root.

Template Parameters

F	must model genfloat.

Definition at line 747 of file math_floating_point.h.

◆ select()

template<typename T1 , typename T2 , COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_geninteger< T1 >::value && detail::builtin::is_igeninteger< T2 >::value, detail::builtin::is_geninteger< T1 >::value && detail::builtin::is_ugeninteger< T2 >::value, detail::builtin::is_genfloatf< T1 >::value && detail::builtin::is_genint< T2 >::value, detail::builtin::is_genfloatf< T1 >::value && detail::builtin::is_ugenint< T2 >::value, detail::builtin::is_genfloatd< T1 >::value && detail::builtin::is_igeninteger64bit< T2 >::value, detail::builtin::is_genfloatd< T1 >::value && detail::builtin::is_ugeninteger64bit< T2 >::value >::value) >

T1 cl::sycl::select	(	T1	a,
		T1	b,
		T2	c
	)

noexcept

For each component of a vector type: result[i] = (MSB of c[i] is set)? b[i] : a[i] For a scalar type: result = c ? b : a.

geninteger must have the same number of elements and bits as gentype.

Template Parameters

T1	must model geninteger, genfloatf, or genfloatd.
T2	If:

T1 models geninteger, then T2 must model either igeninteger or ugeninteger.
T1 models genfloatf, then T2 must model either igenint or ugenint.
T1 models genfloatd, then T2 must model either igeninteger64bit or ugeninteger64bit.

Definition at line 248 of file math_relational.h.

◆ setprecision()

precision_manipulator cl::sycl::setprecision ( int precision )

inline

Inline function that constructs a precision_manipulator object, used in conjunction with stream operators.

Parameters

precision The precision value to construct the precision_manipulator with.

Returns: The constructed precision_manipulator.

Definition at line 247 of file stream.h.

◆ setw()

detail::width_manipulator cl::sycl::setw ( const int width )

inline

Creates a stream manipulator for setting the stream width.

Parameters

width Desired stream width

Returns: Stream width manipulator object

Definition at line 255 of file stream.h.

◆ signbit()

template<typename F , typename return_t = detail::matching_integral_t<detail::collapse_swizzled_vec_t<F>>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::signbit ( F x )

noexcept

Test for sign bit.

The scalar version of the function returns a 1 if the sign bit in the float is set else returns 0.

The vector version of the function returns the following for each component in floatn: -1 (i.e all bits set) if the sign bit in the float is set else returns 0.

Template Parameters

F	must model genfloatf or genfloatd.

Definition at line 184 of file math_relational.h.

◆ sin()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::sin ( F x )

noexcept

Compute sine.

Template Parameters

F	must model genfloat.

Definition at line 756 of file math_floating_point.h.

◆ sincos()

template<typename F , access::address_space AddressSpace, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

F cl::sycl::sincos	(	F	x,
		detail::builtin_ptr< F, AddressSpace >	cosval
	)

noexcept

Compute sine and cosine of x.

The computed sine is the return value and computed cosine is returned in cosval.

Template Parameters

F	must model geninteger.
AddressSpace	The address space for which the pointer may address. Cannot be constant_space.

Definition at line 770 of file math_floating_point.h.

◆ sinh()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::sinh ( F x )

noexcept

Compute hyperbolic sine.

Template Parameters

F	must model genfloat.

Definition at line 779 of file math_floating_point.h.

◆ sinpi()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::sinpi ( F x )

noexcept

Compute sin(π * x).

Template Parameters

F	must model genfloat.

Definition at line 788 of file math_floating_point.h.

◆ sqrt()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::sqrt ( F x )

noexcept

Compute square root.

Template Parameters

F	must model genfloat.

Definition at line 797 of file math_floating_point.h.

◆ step()

template<typename F1 , typename F2 , COMPUTECPP_REQUIRES((detail::builtin::is_genfloat< F1 >::value && std::is_same< F1, F2 >::value)||(std::is_same< F1, half >::value && detail::builtin::is_genfloath< F2 >::value)||(std::is_same< F1, float >::value && detail::builtin::is_genfloatf< F2 >::value)||(std::is_same< F1, double >::value && detail::builtin::is_genfloatd< F2 >::value)) >

F2 cl::sycl::step	(	F1	edge,
		F2	x
	)

noexcept

Returns 0.0 if x < edge, otherwise it returns 1.0.

Template Parameters

F1	must model genfloat.
F2	must model either genfloat or be the scalar equivalent of T1.

Definition at line 158 of file math_common.h.

◆ stream_value() [1/15]

template<typename pointerT >

const stream& cl::sycl::stream_value	(	const stream &	os,
		pointerT *	rhs
	)

inline

Template function that streams a pointer value to a stream object.

Template Parameters

pointerT The type of a pointer value to be streamed.

Parameters

os	The stream object to be streamed to.
rhs	The pointer value to be streamed.

Definition at line 876 of file stream.h.

◆ stream_value() [2/15]

template<typename dataType , cl::sycl::access::address_space asp>

const stream& cl::sycl::stream_value	(	const stream &	os,
		const multi_ptr< dataType, asp > &	rhs
	)

inline

Streams a multi_ptr object into a stream.

Template Parameters

dataType	Underlying type of the multi_ptr pointer
asp	Address space of the pointer

Parameters

os	The stream object to be streamed to
rhs	multi_ptr object to display in the stream

Returns: Original stream object

Definition at line 889 of file stream.h.

◆ stream_value() [3/15]

template<class dataType , COMPUTECPP_ENABLE_IF(dataType,(std::is_integral< dataType >::value)) >

const stream& cl::sycl::stream_value	(	const stream &	os,
		dataType	rhs
	)

inline

Streams an integral type into a stream.

Template Parameters

dataType	Type of the value to be streamed
COMPUTECPP_ENABLE_IF	Only enabled for integral types

Parameters

os	The stream object to be streamed to
rhs	Integral value

Returns: Original stream object

Definition at line 904 of file stream.h.

◆ stream_value() [4/15]

const stream& cl::sycl::stream_value	(	const stream &	os,
		const half &	rhs
	)

inline

Streams a half value to a stream object.

Parameters

os	The stream object to be streamed to
rhs	The half value to be streamed

Definition at line 913 of file stream.h.

◆ stream_value() [5/15]

const stream& cl::sycl::stream_value	(	const stream &	os,
		const float &	rhs
	)

inline

Inline function that streams a float value to a stream object.

Parameters

os	The stream object to be streamed to.
rhs	The float value to be streamed.

Definition at line 923 of file stream.h.

◆ stream_value() [6/15]

const stream& cl::sycl::stream_value	(	const stream &	os,
		const double &	rhs
	)

inline

Inline function that streams a double value to a stream object.

Parameters

os	The stream object to be streamed to.
rhs	The double value to be streamed.

Definition at line 933 of file stream.h.

◆ stream_value() [7/15]

const stream& cl::sycl::stream_value	(	const stream &	os,
		const bool &	rhs
	)

inline

Inline function that streams a bool value to a stream object.

Parameters

os	The stream object to be streamed to.
rhs	The bool value to be streamed.

Definition at line 943 of file stream.h.

◆ stream_value() [8/15]

const stream& cl::sycl::stream_value	(	const stream &	os,
		const char &	rhs
	)

inline

Inline function that streams a const char & value to a stream object.

Parameters

os	The stream object to be streamed to.
rhs	The const char & value to be streamed.

Definition at line 953 of file stream.h.

◆ stream_value() [9/15]

const stream& cl::sycl::stream_value	(	const stream &	os,
		const char *	rhs
	)

inline

Inline function that streams a const char * value to a stream object.

Parameters

os	The stream object to be streamed to.
rhs	The const char * value to be streamed.

Definition at line 963 of file stream.h.

◆ stream_value() [10/15]

const stream& cl::sycl::stream_value	(	const stream &	os,
		const stream_manipulator &	rhs
	)

inline

Inline function that streams a manipulator to a stream object, applying the appropriate action.

Parameters

os	The stream object to be streamed to.
rhs	The stream_manipulator be applied to the stream object.

Definition at line 974 of file stream.h.

◆ stream_value() [11/15]

const stream& cl::sycl::stream_value	(	const stream &	os,
		const precision_manipulator &	rhs
	)

inline

Inline function that streams a precision manipulator to a stream object, applying the precision value.

Parameters

os	The stream object to be streamed to.
rhs	The precision_manipulator be applied to the stream object.

Definition at line 990 of file stream.h.

◆ stream_value() [12/15]

const stream& cl::sycl::stream_value	(	const stream &	os,
		const detail::width_manipulator &	rhs
	)

inline

Sets the width used for streaming data to the value specified by rhs.

Parameters

os	The stream object to be streamed to
rhs	Width manipulator specifying the new width

Returns: Original stream object

Definition at line 1001 of file stream.h.

◆ stream_value() [13/15]

template<int dimensions>

const stream& cl::sycl::stream_value	(	const stream &	os,
		const id< dimensions > &	rhs,
		char *	typeStr = `nullptr`
	)

inline

Streams an id object into a stream object.

Template Parameters

dimensions Number of dimensions of the id

Parameters

os	The stream object to be streamed to
rhs	id object
typeStr	Optional type name

Returns: Original stream object

Definition at line 1015 of file stream.h.

◆ stream_value() [14/15]

template<int dimensions>

const stream& cl::sycl::stream_value	(	const stream &	os,
		const range< dimensions > &	rhs,
		char *	typeStr = `nullptr`
	)

inline

Streams a range object into a stream object.

Template Parameters

dimensions Number of dimensions of the range

Parameters

os	The stream object to be streamed to
rhs	range object
typeStr	Optional type name

Returns: Original stream object

Definition at line 1035 of file stream.h.

◆ stream_value() [15/15]

template<int dimensions, bool with_offset>

const stream& cl::sycl::stream_value	(	const stream &	os,
		const item< dimensions, with_offset > &	rhs,
		char *	typeStr = `nullptr`
	)

inline

Streams an item object into a stream object.

Template Parameters

dimensions	Number of dimensions of the item
with_offset	Whether the item object has an offset or not

Parameters

os	The stream object to be streamed to
rhs	item object

Returns: Original stream object

Definition at line 1086 of file stream.h.

◆ sub_sat()

template<typename Integral , COMPUTECPP_REQUIRES(detail::builtin::is_geninteger< Integral >::value) >

Integral cl::sycl::sub_sat	(	Integral	x,
		Integral	y
	)

noexcept

Returns x - y and saturates the result.

Template Parameters

Integral must model geninteger.

Definition at line 295 of file math_integral.h.

◆ tan()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::tan ( F x )

noexcept

Compute tangent.

Template Parameters

F	must model genfloat.

Definition at line 806 of file math_floating_point.h.

◆ tanh()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::tanh ( F x )

noexcept

Compute hyperbolic tangent.

Template Parameters

F	must model genfloat.

Definition at line 815 of file math_floating_point.h.

◆ tanpi()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::tanpi ( F x )

noexcept

Compute tan(π * x).

Template Parameters

F	must model genfloat.

Definition at line 824 of file math_floating_point.h.

◆ tgamma()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::tgamma ( F x )

noexcept

Compute the gamma function.

Template Parameters

F	must model genfloat.

Definition at line 833 of file math_floating_point.h.

◆ trunc()

template<typename F , typename return_t = detail::collapse_swizzled_vec_t<F>, COMPUTECPP_REQUIRES(detail::builtin::is_genfloat< F >::value) >

return_t cl::sycl::trunc ( F x )

noexcept

Round to integral value using the round to zero rounding mode.

Template Parameters

F	must model genfloat.

Definition at line 842 of file math_floating_point.h.

◆ upsample()

template<typename I1 , typename I2 , COMPUTECPP_REQUIRES( computecpp::gsl::or_< detail::builtin::is_geninteger8bit< I1 >::value && detail::builtin::is_ugeninteger8bit< I2 >::value, detail::builtin::is_geninteger16bit< I1 >::value && detail::builtin::is_ugeninteger16bit< I2 >::value, detail::builtin::is_geninteger32bit< I1 >::value && detail::builtin::is_ugeninteger32bit< I2 >::value >::value) >

auto cl::sycl::upsample	(	I1	hi,
		I2	lo
	)		-> decltype(::cl::sycl::detail::double_width_cast(std::declval<I1>()))

noexcept

Returns result[i] = (hi[i] << 8) | lo[i]

Template Parameters

I1	must model geninteger. The integral size of I1 must be double the integral size of I2.
I2	must model geninteger.

Definition at line 313 of file math_integral.h.

Variable Documentation

◆ accessOffset

AllocatorT id<kDims> cl::sycl::accessOffset

Definition at line 999 of file buffer_accessor.h.

◆ accessRange

AllocatorT id<kDims> range<kDims> cl::sycl::accessRange

Definition at line 999 of file buffer_accessor.h.

◆ bufferRef

AllocatorT& cl::sycl::bufferRef

Definition at line 999 of file buffer_accessor.h.

◆ COMPUTECPP_CONVERT_ATTR

cl::sycl::accessor< elemT, kDims, kMode, access::target::subgroup_local, access::placeholder::false_t > cl::sycl::COMPUTECPP_CONVERT_ATTR

Definition at line 876 of file buffer_accessor.h.

◆ COMPUTECPP_CONVERT_ATTR_PLACEHOLDER

COMPUTECPP_ACCESSOR_LINUX_ALIGNMENT cl::sycl::COMPUTECPP_CONVERT_ATTR_PLACEHOLDER

Definition at line 1191 of file buffer_accessor.h.

◆ COMPUTECPP_CONVERT_ATTR_SAMPLER

class COMPUTECPP_EXPORT COMPUTECPP_SAMPLER_WINDOWS_ALIGNMENT cl::sycl::sampler cl::sycl::COMPUTECPP_CONVERT_ATTR_SAMPLER

◆ COMPUTECPP_PRIVATE_MEMORY_ATTR

cl::sycl::private_memory cl::sycl::COMPUTECPP_PRIVATE_MEMORY_ATTR

◆ dec

const stream_manipulator cl::sycl::dec = stream_manipulator::dec

Stream manipulator for displaying numbers in base 10.

Definition at line 183 of file stream.h.

◆ defaultfloat

const stream_manipulator cl::sycl::defaultfloat = stream_manipulator::defaultfloat

Stream manipulator that reverts the display of floating point numbers back to the default representation.

Definition at line 235 of file stream.h.

◆ endl

const stream_manipulator cl::sycl::endl = stream_manipulator::endl

Const value defined to the endl stream manipulator, used in conjunction with stream operators.

Definition at line 221 of file stream.h.

◆ fixed

const stream_manipulator cl::sycl::fixed = stream_manipulator::fixed

Stream manipulator that forces floating point numbers to be displayed using fixed precision.

Definition at line 226 of file stream.h.

◆ flush

const stream_manipulator cl::sycl::flush = stream_manipulator::flush

Stream manipulator that flushes the current stream buffer.

Definition at line 238 of file stream.h.

◆ functor

range< dimensions > const functorT & cl::sycl::functor

Initial value:

{
  group<dimensions>(groupID).parallel_for_work_item(functor)

Definition at line 1681 of file apis.h.

◆ hex

const stream_manipulator cl::sycl::hex = stream_manipulator::hex

Const value defined to the hex stream manipulator, used in conjunction with stream operators.

Definition at line 189 of file stream.h.

◆ hexfloat

const stream_manipulator cl::sycl::hexfloat = stream_manipulator::hexfloat

Stream manipulator that displays floating point numbers in base 16.

Definition at line 230 of file stream.h.

◆ kDims

COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT cl::sycl::kDims

Definition at line 913 of file buffer_accessor.h.

◆ kMode

COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT cl::sycl::kMode

Definition at line 913 of file buffer_accessor.h.

◆ kTarget

COMPUTECPP_SET_IGNORE_DEPRECATED_API class COMPUTECPP_ACCESSOR_WINDOWS_ALIGNMENT cl::sycl::kTarget

Definition at line 913 of file buffer_accessor.h.

◆ localRange

range<dimensions> cl::sycl::localRange

Definition at line 1693 of file apis.h.

◆ noshowbase

const stream_manipulator cl::sycl::noshowbase = stream_manipulator::noshowbase

Stream manipulator for reverting showbase.

Definition at line 205 of file stream.h.

◆ noshowpos

const stream_manipulator cl::sycl::noshowpos = stream_manipulator::noshowpos

Stream manipulator for reverting showpos.

Definition at line 215 of file stream.h.

◆ oct

const stream_manipulator cl::sycl::oct = stream_manipulator::oct

Const value defined to the oct stream manipulator, used in conjunction with stream operators.

Definition at line 195 of file stream.h.

◆ scientific

const stream_manipulator cl::sycl::scientific = stream_manipulator::scientific

Const value defined to the scientific stream manipulator, used in conjunction with stream operators.

Definition at line 179 of file stream.h.

◆ showbase

const stream_manipulator cl::sycl::showbase = stream_manipulator::showbase

Const value defined to the showbase stream manipulator, used in conjunction with stream operators.

Definition at line 201 of file stream.h.

◆ showpos

const stream_manipulator cl::sycl::showpos = stream_manipulator::showpos

Const value defined to the showpos stream manipulator, used in conjunction with stream operators.

Definition at line 211 of file stream.h.

menu_bookReference

Namespaces

Classes

Typedefs

Enumerations

Functions

Variables

Typedef Documentation

◆ async_handler

◆ atomic_float

◆ atomic_int

◆ atomic_uint

◆ backend_input_t

◆ backend_return_t

◆ base_t

◆ bitset_class

◆ byte

◆ char16

◆ char2

◆ char3

◆ char4

◆ char8

◆ cl_bool

◆ cl_char

◆ cl_char16

◆ cl_char2

◆ cl_char3

◆ cl_char4

◆ cl_char8

◆ cl_double

◆ cl_double16

◆ cl_double2

◆ cl_double3

◆ cl_double4

◆ cl_double8

◆ cl_float

◆ cl_float16

◆ cl_float2

◆ cl_float3

◆ cl_float4

◆ cl_float8

◆ cl_half

◆ cl_half16

◆ cl_half2

◆ cl_half3

◆ cl_half4

◆ cl_half8

◆ cl_int

◆ cl_int16

◆ cl_int2

◆ cl_int3

◆ cl_int4

◆ cl_int8

◆ cl_long

◆ cl_long16

◆ cl_long2

◆ cl_long3

◆ cl_long4

◆ cl_long8

◆ cl_short

◆ cl_short16

◆ cl_short2

◆ cl_short3

◆ cl_short4

◆ cl_short8

◆ cl_uchar

◆ cl_uchar16

◆ cl_uchar2

◆ cl_uchar3

◆ cl_uchar4

◆ cl_uchar8

◆ cl_uint

◆ cl_uint16

◆ cl_uint2

◆ cl_uint3

◆ cl_uint4

◆ cl_uint8

◆ cl_ulong

◆ cl_ulong16

◆ cl_ulong2