amdgpuintrin.h File Reference (original) (raw)

Go to the source code of this file.

Macros
#define	bool _Bool

#define	__gpu_private __attribute__((address_space(5)))

#define	__gpu_constant __attribute__((address_space(4)))

#define	__gpu_local __attribute__((address_space(3)))

#define	__gpu_global __attribute__((address_space(1)))

#define	__gpu_generic __attribute__((address_space(0)))

#define	__gpu_kernel __attribute__((amdgpu_kernel, visibility("protected")))

Functions
	_Pragma ("push_macro(\"bool\")")

	_Pragma ("omp begin declare target device_type(nohost)")

	_Pragma ("omp begin declare variant match(device = {arch(amdgcn)})")

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_num_blocks_x (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_num_blocks_y (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_num_blocks_z (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_block_id_x (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_block_id_y (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_block_id_z (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_num_threads_x (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_num_threads_y (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_num_threads_z (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_thread_id_x (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_thread_id_y (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_thread_id_z (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_num_lanes (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_lane_id (void)

static _DEFAULT_FN_ATTRS __inline__ uint64_t	__gpu_lane_mask (void)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_read_first_lane_u32 (uint64_t __lane_mask, uint32_t __x)

_DEFAULT_FN_ATTRS __inline__ uint64_t	__gpu_read_first_lane_u64 (uint64_t __lane_mask, uint64_t __x)

static _DEFAULT_FN_ATTRS __inline__ uint64_t	__gpu_ballot (uint64_t __lane_mask, bool __x)

static _DEFAULT_FN_ATTRS __inline__ void	__gpu_sync_threads (void)

static _DEFAULT_FN_ATTRS __inline__ void	__gpu_sync_lane (uint64_t __lane_mask)

static _DEFAULT_FN_ATTRS __inline__ uint32_t	__gpu_shuffle_idx_u32 (uint64_t __lane_mask, uint32_t __idx, uint32_t __x)

static _DEFAULT_FN_ATTRS __inline__ uint64_t	__gpu_shuffle_idx_u64 (uint64_t __lane_mask, uint32_t __idx, uint64_t __x)

static _DEFAULT_FN_ATTRS __inline__ bool	__gpu_is_ptr_local (void *ptr)

static _DEFAULT_FN_ATTRS __inline__ bool	__gpu_is_ptr_private (void *ptr)

static _DEFAULT_FN_ATTRS __inline__ void	__gpu_exit (void)

static _DEFAULT_FN_ATTRS __inline__ void	__gpu_thread_suspend (void)

	_Pragma ("omp end declare variant")

	_Pragma ("omp end declare target")

	_Pragma ("pop_macro(\"bool\")")

◆ __gpu_constant

◆ __gpu_generic

◆ __gpu_global

◆ __gpu_kernel

#define __gpu_kernel __attribute__((amdgpu_kernel, visibility("protected")))

◆ __gpu_local

◆ __gpu_private

◆ bool

◆ __gpu_ballot()

◆ __gpu_block_id_x()

◆ __gpu_block_id_y()

◆ __gpu_block_id_z()

◆ __gpu_exit()

◆ __gpu_is_ptr_local()

◆ __gpu_is_ptr_private()

◆ __gpu_lane_id()

◆ __gpu_lane_mask()

◆ __gpu_num_blocks_x()

◆ __gpu_num_blocks_y()

◆ __gpu_num_blocks_z()

◆ __gpu_num_lanes()

◆ __gpu_num_threads_x()

◆ __gpu_num_threads_y()

◆ __gpu_num_threads_z()

◆ __gpu_read_first_lane_u32()

static _DEFAULT_FN_ATTRS __inline__ uint32_t __gpu_read_first_lane_u32 ( uint64_t __lane_mask, uint32_t __x )	static

◆ __gpu_read_first_lane_u64()

_DEFAULT_FN_ATTRS __inline__ uint64_t __gpu_read_first_lane_u64	(	uint64_t	__lane_mask,
uint64_t	__x
)

◆ __gpu_shuffle_idx_u32()

static _DEFAULT_FN_ATTRS __inline__ uint32_t __gpu_shuffle_idx_u32 ( uint64_t __lane_mask, uint32_t __idx, uint32_t __x )	static

◆ __gpu_shuffle_idx_u64()

static _DEFAULT_FN_ATTRS __inline__ uint64_t __gpu_shuffle_idx_u64 ( uint64_t __lane_mask, uint32_t __idx, uint64_t __x )	static

◆ __gpu_sync_lane()

static _DEFAULT_FN_ATTRS __inline__ void __gpu_sync_lane ( uint64_t __lane_mask)	static

◆ __gpu_sync_threads()

◆ __gpu_thread_id_x()

◆ __gpu_thread_id_y()

◆ __gpu_thread_id_z()

◆ __gpu_thread_suspend()

◆ _Pragma() [1/6]

_Pragma	(	"omp begin declare target device_type(nohost)"	)

◆ _Pragma() [2/6]

_Pragma	(	"omp begin declare variant match(device = {arch(amdgcn)})"	)

◆ _Pragma() [3/6]

_Pragma	(	"omp end declare target"	)

◆ _Pragma() [4/6]

_Pragma	(	"omp end declare variant"	)

◆ _Pragma() [5/6]

_Pragma	(	"pop_macro(\"bool\")"	)

◆ _Pragma() [6/6]

_Pragma	(	"push_macro(\"bool\")"	)