21.08
|
#include "load_store_utility.h"
Go to the source code of this file.
Data Structures | |
struct | Vector |
Structure to hold Vector information. More... | |
struct | Image |
Structure to hold Image information. More... | |
struct | Tensor3D |
Structure to hold 3D tensor information. More... | |
struct | Tensor4D |
Structure to hold 4D tensor information. More... | |
Macros | ||||||||||
#define | GPU_ARCH_MIDGARD 0x100 | |||||||||
#define | GPU_ARCH_BIFROST 0x200 | |||||||||
#define | CONCAT(a, b) a##b | |||||||||
Concatenate two inputs. More... | ||||||||||
#define | EXPAND(x) x | |||||||||
Expand the given vector. More... | ||||||||||
#define | CLAMP(x, min_val, max_val) min(max(x, min_val), max_val) | |||||||||
Clamp the given value between an upper and lower bound. More... | ||||||||||
#define | VLOAD_STR(size) vload##size | |||||||||
#define | VLOAD(size) VLOAD_STR(size) | |||||||||
#define | PIXEL_UNIT4 1 | |||||||||
#define | PIXEL_UNIT8 2 | |||||||||
#define | PIXEL_UNIT16 4 | |||||||||
#define | read_image2d_floatx1(img, x_coord, y_coord) (float4)(read_imagef(img, (int2)(x_coord, y_coord))); | |||||||||
#define | read_image2d_floatx2(img, x_coord, y_coord) (float8)(read_imagef(img, (int2)(x_coord, y_coord)), read_imagef(img, (int2)(x_coord + 1, y_coord))); | |||||||||
#define | read_image2d_floatx4(img, x_coord, y_coord) (float16)(read_imagef(img, (int2)(x_coord, y_coord)), read_imagef(img, (int2)(x_coord + 1, y_coord)), read_imagef(img, (int2)(x_coord + 2, y_coord)), read_imagef(img, (int2)(x_coord + 3, y_coord))); | |||||||||
#define | convert_float_sat convert_float | |||||||||
#define | convert_float1_sat convert_float | |||||||||
#define | convert_float2_sat convert_float2 | |||||||||
#define | convert_float3_sat convert_float3 | |||||||||
#define | convert_float4_sat convert_float4 | |||||||||
#define | convert_float8_sat convert_float8 | |||||||||
#define | convert_float16_sat convert_float16 | |||||||||
#define | convert_half_sat convert_float | |||||||||
#define | convert_half1_sat convert_half | |||||||||
#define | convert_half2_sat convert_half2 | |||||||||
#define | convert_half3_sat convert_half3 | |||||||||
#define | convert_half4_sat convert_half4 | |||||||||
#define | convert_half8_sat convert_half8 | |||||||||
#define | convert_half16_sat convert_half16 | |||||||||
#define | convert_float1 convert_float | |||||||||
#define | convert_half1 convert_half | |||||||||
#define | convert_char1 convert_char | |||||||||
#define | convert_uchar1 convert_uchar | |||||||||
#define | convert_short1 convert_short | |||||||||
#define | convert_ushort1 convert_ushort | |||||||||
#define | convert_int1 convert_int | |||||||||
#define | convert_uint1 convert_uint | |||||||||
#define | convert_long1 convert_long | |||||||||
#define | convert_ulong1 convert_ulong | |||||||||
#define | convert_double1 convert_double | |||||||||
#define | convert_char1_sat convert_char_sat | |||||||||
#define | convert_uchar1_sat convert_uchar_sat | |||||||||
#define | convert_uchar2_sat convert_uchar2_sat | |||||||||
#define | convert_uchar3_sat convert_uchar3_sat | |||||||||
#define | convert_uchar4_sat convert_uchar4_sat | |||||||||
#define | convert_uchar8_sat convert_uchar8_sat | |||||||||
#define | convert_uchar16_sat convert_uchar16_sat | |||||||||
#define | convert_short1_sat convert_short_sat | |||||||||
#define | convert_ushort1_sat convert_ushort_sat | |||||||||
#define | convert_int1_sat convert_int_sat | |||||||||
#define | convert_uint1_sat convert_uint_sat | |||||||||
#define | convert_long1_sat convert_long_sat | |||||||||
#define | convert_ulong1_sat convert_ulong_sat | |||||||||
#define | convert_double1_sat convert_double_sat | |||||||||
#define | VEC_DATA_TYPE_STR(type, size) type##size | |||||||||
#define | VEC_DATA_TYPE(type, size) VEC_DATA_TYPE_STR(type, size) | |||||||||
#define | CONVERT_STR(x, type) (convert_##type((x))) | |||||||||
#define | CONVERT(x, type) CONVERT_STR(x, type) | |||||||||
#define | CONVERT_SAT_STR(x, type) (convert_##type##_sat((x))) | |||||||||
#define | CONVERT_SAT(x, type) CONVERT_SAT_STR(x, type) | |||||||||
#define | CONVERT_SAT_ROUND_STR(x, type, round) (convert_##type##_sat_##round((x))) | |||||||||
#define | CONVERT_SAT_ROUND(x, type, round) CONVERT_SAT_ROUND_STR(x, type, round) | |||||||||
#define | select_vec_dt_uchar(size) uchar##size | |||||||||
#define | select_vec_dt_char(size) char##size | |||||||||
#define | select_vec_dt_ushort(size) ushort##size | |||||||||
#define | select_vec_dt_short(size) short##size | |||||||||
#define | select_vec_dt_half(size) short##size | |||||||||
#define | select_vec_dt_uint(size) uint##size | |||||||||
#define | select_vec_dt_int(size) int##size | |||||||||
#define | select_vec_dt_float(size) int##size | |||||||||
#define | select_vec_dt_ulong(size) ulong##size | |||||||||
#define | select_vec_dt_long(size) long##size | |||||||||
#define | SELECT_VEC_DATA_TYPE_STR(type, size) select_vec_dt_##type(size) | |||||||||
#define | SELECT_VEC_DATA_TYPE(type, size) SELECT_VEC_DATA_TYPE_STR(type, size) | |||||||||
#define | SELECT_DATA_TYPE(type) SELECT_VEC_DATA_TYPE_STR(type, 1) | |||||||||
#define | signed_int_vec_dt_uchar(size) char##size | |||||||||
#define | signed_int_vec_dt_char(size) char##size | |||||||||
#define | signed_int_vec_dt_ushort(size) short##size | |||||||||
#define | signed_int_vec_dt_short(size) short##size | |||||||||
#define | signed_int_vec_dt_half(size) short##size | |||||||||
#define | signed_int_vec_dt_uint(size) int##size | |||||||||
#define | signed_int_vec_dt_int(size) int##size | |||||||||
#define | signed_int_vec_dt_float(size) int##size | |||||||||
#define | signed_int_vec_dt_ulong(size) long##size | |||||||||
#define | signed_int_vec_dt_long(size) long##size | |||||||||
#define | SIGNED_INT_VEC_DATA_TYPE_STR(type, size) signed_int_vec_dt_##type(size) | |||||||||
#define | SIGNED_INT_VEC_DATA_TYPE(type, size) SIGNED_INT_VEC_DATA_TYPE_STR(type, size) | |||||||||
#define | SIGNED_INT_DATA_TYPE(type) SIGNED_INT_VEC_DATA_TYPE_STR(type, 1) | |||||||||
#define | sum_reduce_1(x) (x) | |||||||||
#define | sum_reduce_2(x) ((x).s0) + ((x).s1) | |||||||||
#define | sum_reduce_3(x) sum_reduce_2((x).s01) + ((x).s2) | |||||||||
#define | sum_reduce_4(x) sum_reduce_2((x).s01) + sum_reduce_2((x).s23) | |||||||||
#define | sum_reduce_8(x) sum_reduce_4((x).s0123) + sum_reduce_4((x).s4567) | |||||||||
#define | sum_reduce_16(x) sum_reduce_8((x).s01234567) + sum_reduce_8((x).s89ABCDEF) | |||||||||
#define | SUM_REDUCE_STR(x, size) sum_reduce_##size(x) | |||||||||
#define | SUM_REDUCE(x, size) SUM_REDUCE_STR(x, size) | |||||||||
#define | prod_reduce_1(x) (x) | |||||||||
#define | prod_reduce_2(x) ((x).s0) * ((x).s1) | |||||||||
#define | prod_reduce_3(x) prod_reduce_2((x).s01) * ((x).s2) | |||||||||
#define | prod_reduce_4(x) prod_reduce_2((x).s01) * prod_reduce_2((x).s23) | |||||||||
#define | prod_reduce_8(x) prod_reduce_4((x).s0123) * prod_reduce_4((x).s4567) | |||||||||
#define | prod_reduce_16(x) prod_reduce_8((x).s01234567) * prod_reduce_8((x).s89ABCDEF) | |||||||||
#define | PROD_REDUCE_STR(x, size) prod_reduce_##size(x) | |||||||||
#define | PROD_REDUCE(x, size) PROD_REDUCE_STR(x, size) | |||||||||
#define | max_reduce_1(x) (x) | |||||||||
#define | max_reduce_2(x) max(((x).s0), ((x).s1)) | |||||||||
#define | max_reduce_3(x) max(max_reduce_2((x).s01), ((x).s2)) | |||||||||
#define | max_reduce_4(x) max(max_reduce_2((x).s01), max_reduce_2((x).s23)) | |||||||||
#define | max_reduce_8(x) max(max_reduce_4((x).s0123), max_reduce_4((x).s4567)) | |||||||||
#define | max_reduce_16(x) max(max_reduce_8((x).s01234567), max_reduce_8((x).s89ABCDEF)) | |||||||||
#define | MAX_REDUCE_STR(x, size) max_reduce_##size(x) | |||||||||
#define | MAX_REDUCE(x, size) MAX_REDUCE_STR(x, size) | |||||||||
#define | VECTOR_DECLARATION(name) | |||||||||
#define | IMAGE_DECLARATION(name) | |||||||||
#define | TENSOR3D_DECLARATION(name) | |||||||||
#define | TENSOR4D_DECLARATION(name) | |||||||||
#define | CONVERT_TO_VECTOR_STRUCT(name) update_vector_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x) | |||||||||
#define | CONVERT_TO_VECTOR_STRUCT_NO_STEP(name) update_vector_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0) | |||||||||
#define | CONVERT_TO_IMAGE_STRUCT(name) update_image_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x, name##_stride_y, name##_step_y) | |||||||||
#define | CONVERT_TO_IMAGE_STRUCT_NO_STEP(name) update_image_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0) | |||||||||
#define | CONVERT_TENSOR3D_TO_IMAGE_STRUCT(name) update_image_from_tensor3D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x, name##_stride_y, name##_step_y, name##_stride_z, name##_step_z) | |||||||||
#define | CONVERT_TENSOR3D_TO_IMAGE_STRUCT_NO_STEP(name) update_image_from_tensor3D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, name##_step_z) | |||||||||
#define | CONVERT_TENSOR3D_TO_IMAGE_STRUCT(name) update_image_from_tensor3D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x, name##_stride_y, name##_step_y, name##_stride_z, name##_step_z) | |||||||||
#define | CONVERT_TO_TENSOR3D_STRUCT(name) | |||||||||
#define | CONVERT_TO_TENSOR3D_STRUCT_NO_STEP(name) update_tensor3D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, 0) | |||||||||
#define | CONVERT_TO_TENSOR4D_STRUCT(name, mod_size) | |||||||||
#define | CONVERT_TO_TENSOR4D_STRUCT_NO_STEP(name, mod_size) update_tensor4D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, 0, name##_stride_w, 0, mod_size) | |||||||||
#define | CONVERT_TO_TENSOR3D_STRUCT_NO_UPDATE_PTR(name) | |||||||||
REVn | ||||||||||
REVn reverses the given vector whose size is n.
| ||||||||||
#define | REV1(x) ((x)) | |||||||||
#define | REV2(x) ((x).s10) | |||||||||
#define | REV3(x) ((x).s210) | |||||||||
#define | REV4(x) ((x).s3210) | |||||||||
#define | REV8(x) ((x).s76543210) | |||||||||
#define | REV16(x) ((x).sFEDCBA9876543210) | |||||||||
REVERSE | ||||||||||
Reverse the given vector.
| ||||||||||
#define | REVERSE_STR(x, s) REV##s((x)) | |||||||||
#define | REVERSE(x, s) REVERSE_STR(x, s) | |||||||||
ROTs_n | ||||||||||
Circular-right-shift (rotate-right) the vector of size s by the amount of n.
| ||||||||||
#define | ROT1_0(x) ((x)) | |||||||||
#define | ROT1_1(x) ((x)) | |||||||||
#define | ROT2_0(x) ((x)) | |||||||||
#define | ROT2_1(x) ((x).s10) | |||||||||
#define | ROT2_2(x) ((x)) | |||||||||
#define | ROT3_0(x) ((x)) | |||||||||
#define | ROT3_1(x) ((x).s201) | |||||||||
#define | ROT3_2(x) ((x).s120) | |||||||||
#define | ROT3_3(x) ((x)) | |||||||||
#define | ROT4_0(x) ((x)) | |||||||||
#define | ROT4_1(x) ((x).s3012) | |||||||||
#define | ROT4_2(x) ((x).s2301) | |||||||||
#define | ROT4_3(x) ((x).s1230) | |||||||||
#define | ROT4_4(x) ((x)) | |||||||||
#define | ROT8_0(x) ((x)) | |||||||||
#define | ROT8_1(x) ((x).s70123456) | |||||||||
#define | ROT8_2(x) ((x).s67012345) | |||||||||
#define | ROT8_3(x) ((x).s56701234) | |||||||||
#define | ROT8_4(x) ((x).s45670123) | |||||||||
#define | ROT8_5(x) ((x).s34567012) | |||||||||
#define | ROT8_6(x) ((x).s23456701) | |||||||||
#define | ROT8_7(x) ((x).s12345670) | |||||||||
#define | ROT8_8(x) ((x)) | |||||||||
#define | ROT16_0(x) ((x)) | |||||||||
#define | ROT16_1(x) ((x).sF0123456789ABCDE) | |||||||||
#define | ROT16_2(x) ((x).sEF0123456789ABCD) | |||||||||
#define | ROT16_3(x) ((x).sDEF0123456789ABC) | |||||||||
#define | ROT16_4(x) ((x).sCDEF0123456789AB) | |||||||||
#define | ROT16_5(x) ((x).sBCDEF0123456789A) | |||||||||
#define | ROT16_6(x) ((x).sABCDEF0123456789) | |||||||||
#define | ROT16_7(x) ((x).s9ABCDEF012345678) | |||||||||
#define | ROT16_8(x) ((x).s89ABCDEF01234567) | |||||||||
#define | ROT16_9(x) ((x).s789ABCDEF0123456) | |||||||||
#define | ROT16_10(x) ((x).s6789ABCDEF012345) | |||||||||
#define | ROT16_11(x) ((x).s56789ABCDEF01234) | |||||||||
#define | ROT16_12(x) ((x).s456789ABCDEF0123) | |||||||||
#define | ROT16_13(x) ((x).s3456789ABCDEF012) | |||||||||
#define | ROT16_14(x) ((x).s23456789ABCDEF01) | |||||||||
#define | ROT16_15(x) ((x).s123456789ABCDEF0) | |||||||||
#define | ROT16_16(x) ((x)) | |||||||||
ROTATE | ||||||||||
Circular-right-shift (rotate-right) the given vector by the given amount.
| ||||||||||
#define | ROTATE_STR(x, s, n) ROT##s##_##n(x) | |||||||||
#define | ROTATE(x, s, n) ROTATE_STR(x, s, n) | |||||||||
V_OFFSn | ||||||||||
Creates a vector of size n filled with offset values corresponding to the location of each element.
| ||||||||||
#define | V_OFFS1(dt) (dt##1)(0) | |||||||||
#define | V_OFFS2(dt) (dt##2)(0, 1) | |||||||||
#define | V_OFFS3(dt) (dt##3)(0, 1, 2) | |||||||||
#define | V_OFFS4(dt) (dt##4)(0, 1, 2, 3) | |||||||||
#define | V_OFFS8(dt) (dt##8)(0, 1, 2, 3, 4, 5, 6, 7) | |||||||||
#define | V_OFFS16(dt) (dt##16)(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15) | |||||||||
VEC_OFFS | ||||||||||
Create a vector filled with offset values corresponding to the location of each element.
| ||||||||||
#define | VEC_OFFS_STR(dt, s) V_OFFS##s(dt) | |||||||||
#define | VEC_OFFS(dt, s) VEC_OFFS_STR(dt, s) | |||||||||
CONVERT_VECTOR_SIZE_TO_PIXEL_UNIT | ||||||||||
Utility macro to convert a vector size in pixel unit.
| ||||||||||
#define | CONVERT_VECTOR_SIZE_TO_PIXEL_UNIT_STR(vec_size) PIXEL_UNIT##vec_size | |||||||||
#define | CONVERT_VECTOR_SIZE_TO_PIXEL_UNIT(vec_size) CONVERT_VECTOR_SIZE_TO_PIXEL_UNIT_STR(vec_size) | |||||||||
#define | READ_IMAGE2D_STR(data_type, n0, img, x_coord, y_coord) read_image2d_##data_type##x##n0(img, x_coord, y_coord) | |||||||||
Utility macro to read a 2D OpenCL image object. More... | ||||||||||
#define | READ_IMAGE2D(data_type, n0, img, x_coord, y_coord) READ_IMAGE2D_STR(data_type, n0, img, x_coord, y_coord) | |||||||||
#define | VSTORE_STR(size) vstore##size | |||||||||
#define | VSTORE(size) VSTORE_STR(size) | |||||||||
#define | float1 float | |||||||||
#define | half1 half | |||||||||
#define | char1 char | |||||||||
#define | uchar1 uchar | |||||||||
#define | short1 short | |||||||||
#define | ushort1 ushort | |||||||||
#define | int1 int | |||||||||
#define | uint1 uint | |||||||||
#define | long1 long | |||||||||
#define | ulong1 ulong | |||||||||
#define | double1 double | |||||||||
#define | vload1(OFFSET, PTR) *(OFFSET + PTR) | |||||||||
#define | vstore1(DATA, OFFSET, PTR) *(OFFSET + PTR) = DATA | |||||||||
VSTORE_PARTIAL | ||||||||||
Extended partial vstore that correctly handles scalar values as well. Store the lower 0 to (n-1)th elements of the given vector while minimising the amount of vstore ops
| ||||||||||
#define | VSTORE_PARTIAL_STR(size, store_size) vstore_partial_##size##_##store_size | |||||||||
#define | VSTORE_PARTIAL(size, store_size) VSTORE_PARTIAL_STR(size, store_size) | |||||||||
#define | NO_STORE(data, offs, ptr) | |||||||||
#define | vstore_partial_1_0 NO_STORE | |||||||||
#define | vstore_partial_1_1 vstore1 | |||||||||
#define | vstore_partial_1_2 NO_STORE | |||||||||
#define | vstore_partial_1_3 NO_STORE | |||||||||
#define | vstore_partial_1_4 NO_STORE | |||||||||
#define | vstore_partial_1_5 NO_STORE | |||||||||
#define | vstore_partial_1_6 NO_STORE | |||||||||
#define | vstore_partial_1_7 NO_STORE | |||||||||
#define | vstore_partial_1_8 NO_STORE | |||||||||
#define | vstore_partial_1_9 NO_STORE | |||||||||
#define | vstore_partial_1_10 NO_STORE | |||||||||
#define | vstore_partial_1_11 NO_STORE | |||||||||
#define | vstore_partial_1_12 NO_STORE | |||||||||
#define | vstore_partial_1_13 NO_STORE | |||||||||
#define | vstore_partial_1_14 NO_STORE | |||||||||
#define | vstore_partial_1_15 NO_STORE | |||||||||
#define | vstore_partial_1_16 NO_STORE | |||||||||
#define | vstore_partial_2_0 NO_STORE | |||||||||
#define | vstore_partial_2_1 vstore_partial_1 | |||||||||
#define | vstore_partial_2_2 vstore_partial_2 | |||||||||
#define | vstore_partial_2_3 NO_STORE | |||||||||
#define | vstore_partial_2_4 NO_STORE | |||||||||
#define | vstore_partial_2_5 NO_STORE | |||||||||
#define | vstore_partial_2_6 NO_STORE | |||||||||
#define | vstore_partial_2_7 NO_STORE | |||||||||
#define | vstore_partial_2_8 NO_STORE | |||||||||
#define | vstore_partial_2_9 NO_STORE | |||||||||
#define | vstore_partial_2_10 NO_STORE | |||||||||
#define | vstore_partial_2_11 NO_STORE | |||||||||
#define | vstore_partial_2_12 NO_STORE | |||||||||
#define | vstore_partial_2_13 NO_STORE | |||||||||
#define | vstore_partial_2_14 NO_STORE | |||||||||
#define | vstore_partial_2_15 NO_STORE | |||||||||
#define | vstore_partial_2_16 NO_STORE | |||||||||
#define | vstore_partial_3_0 NO_STORE | |||||||||
#define | vstore_partial_3_1 vstore_partial_1 | |||||||||
#define | vstore_partial_3_2 vstore_partial_2 | |||||||||
#define | vstore_partial_3_3 vstore_partial_3 | |||||||||
#define | vstore_partial_3_4 NO_STORE | |||||||||
#define | vstore_partial_3_5 NO_STORE | |||||||||
#define | vstore_partial_3_6 NO_STORE | |||||||||
#define | vstore_partial_3_7 NO_STORE | |||||||||
#define | vstore_partial_3_8 NO_STORE | |||||||||
#define | vstore_partial_3_9 NO_STORE | |||||||||
#define | vstore_partial_3_10 NO_STORE | |||||||||
#define | vstore_partial_3_11 NO_STORE | |||||||||
#define | vstore_partial_3_12 NO_STORE | |||||||||
#define | vstore_partial_3_13 NO_STORE | |||||||||
#define | vstore_partial_3_14 NO_STORE | |||||||||
#define | vstore_partial_3_15 NO_STORE | |||||||||
#define | vstore_partial_3_16 NO_STORE | |||||||||
#define | vstore_partial_4_0 NO_STORE | |||||||||
#define | vstore_partial_4_1 vstore_partial_1 | |||||||||
#define | vstore_partial_4_2 vstore_partial_2 | |||||||||
#define | vstore_partial_4_3 vstore_partial_3 | |||||||||
#define | vstore_partial_4_4 vstore_partial_4 | |||||||||
#define | vstore_partial_4_5 NO_STORE | |||||||||
#define | vstore_partial_4_6 NO_STORE | |||||||||
#define | vstore_partial_4_7 NO_STORE | |||||||||
#define | vstore_partial_4_8 NO_STORE | |||||||||
#define | vstore_partial_4_9 NO_STORE | |||||||||
#define | vstore_partial_4_10 NO_STORE | |||||||||
#define | vstore_partial_4_11 NO_STORE | |||||||||
#define | vstore_partial_4_12 NO_STORE | |||||||||
#define | vstore_partial_4_13 NO_STORE | |||||||||
#define | vstore_partial_4_14 NO_STORE | |||||||||
#define | vstore_partial_4_15 NO_STORE | |||||||||
#define | vstore_partial_4_16 NO_STORE | |||||||||
#define | vstore_partial_8_0 NO_STORE | |||||||||
#define | vstore_partial_8_1 vstore_partial_1 | |||||||||
#define | vstore_partial_8_2 vstore_partial_2 | |||||||||
#define | vstore_partial_8_3 vstore_partial_3 | |||||||||
#define | vstore_partial_8_4 vstore_partial_4 | |||||||||
#define | vstore_partial_8_5 vstore_partial_5 | |||||||||
#define | vstore_partial_8_6 vstore_partial_6 | |||||||||
#define | vstore_partial_8_7 vstore_partial_7 | |||||||||
#define | vstore_partial_8_8 vstore_partial_8 | |||||||||
#define | vstore_partial_8_9 NO_STORE | |||||||||
#define | vstore_partial_8_10 NO_STORE | |||||||||
#define | vstore_partial_8_11 NO_STORE | |||||||||
#define | vstore_partial_8_12 NO_STORE | |||||||||
#define | vstore_partial_8_13 NO_STORE | |||||||||
#define | vstore_partial_8_14 NO_STORE | |||||||||
#define | vstore_partial_8_15 NO_STORE | |||||||||
#define | vstore_partial_8_16 NO_STORE | |||||||||
#define | vstore_partial_16_0 NO_STORE | |||||||||
#define | vstore_partial_16_1 vstore_partial_1 | |||||||||
#define | vstore_partial_16_2 vstore_partial_2 | |||||||||
#define | vstore_partial_16_3 vstore_partial_3 | |||||||||
#define | vstore_partial_16_4 vstore_partial_4 | |||||||||
#define | vstore_partial_16_5 vstore_partial_5 | |||||||||
#define | vstore_partial_16_6 vstore_partial_6 | |||||||||
#define | vstore_partial_16_7 vstore_partial_7 | |||||||||
#define | vstore_partial_16_8 vstore_partial_8 | |||||||||
#define | vstore_partial_16_9 vstore_partial_9 | |||||||||
#define | vstore_partial_16_10 vstore_partial_10 | |||||||||
#define | vstore_partial_16_11 vstore_partial_11 | |||||||||
#define | vstore_partial_16_12 vstore_partial_12 | |||||||||
#define | vstore_partial_16_13 vstore_partial_13 | |||||||||
#define | vstore_partial_16_14 vstore_partial_14 | |||||||||
#define | vstore_partial_16_15 vstore_partial_15 | |||||||||
#define | vstore_partial_16_16 vstore_partial_16 | |||||||||
vstore_partial_n | ||||||||||
Partial vstore. Store the lower 0 to (n-1)th elements of the given vector while minimising the amount of vstore ops
| ||||||||||
#define | vstore_partial_1(DATA, OFFSET, PTR) vstore1(DATA.s0, OFFSET, PTR); | |||||||||
#define | vstore_partial_2(DATA, OFFSET, PTR) vstore2(DATA.s01, OFFSET, PTR); | |||||||||
#define | vstore_partial_3(DATA, OFFSET, PTR) vstore3(DATA.s012, OFFSET, PTR); | |||||||||
#define | vstore_partial_4(DATA, OFFSET, PTR) vstore4(DATA.s0123, OFFSET, PTR); | |||||||||
#define | vstore_partial_5(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_6(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_7(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_8(DATA, OFFSET, PTR) vstore8(DATA.s01234567, OFFSET, PTR); | |||||||||
#define | vstore_partial_9(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_10(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_11(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_12(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_13(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_14(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_15(DATA, OFFSET, PTR) | |||||||||
#define | vstore_partial_16(DATA, OFFSET, PTR) vstore16(DATA, OFFSET, PTR); | |||||||||
Typedefs | |
typedef struct Vector | Vector |
Structure to hold Vector information. More... | |
typedef struct Image | Image |
Structure to hold Image information. More... | |
typedef struct Tensor3D | Tensor3D |
Structure to hold 3D tensor information. More... | |
typedef struct Tensor4D | Tensor4D |
Structure to hold 4D tensor information. More... | |
Functions | |
Vector | update_vector_workitem_ptr (__global uchar *ptr, uint offset_first_element_in_bytes, uint stride_x, uint step_x) |
Wrap vector information into an Vector structure, and make the pointer point at this workitem's data. More... | |
Image | update_image_workitem_ptr (__global uchar *ptr, uint offset_first_element_in_bytes, uint stride_x, uint step_x, uint stride_y, uint step_y) |
Wrap image information into an Image structure, and make the pointer point at this workitem's data. More... | |
Image | update_image_from_tensor3D_workitem_ptr (__global uchar *ptr, uint offset_first_element_in_bytes, uint stride_x, uint step_x, uint stride_y, uint step_y, uint stride_z, uint step_z) |
Wrap 3D tensor information into an image structure, and make the pointer point at this workitem's data. More... | |
Tensor3D | update_tensor3D_workitem_ptr (__global uchar *ptr, uint offset_first_element_in_bytes, uint stride_x, uint step_x, uint stride_y, uint step_y, uint stride_z, uint step_z) |
Wrap 3D tensor information into an tensor structure, and make the pointer point at this workitem's data. More... | |
Tensor3D | tensor3D_ptr_no_update (__global uchar *ptr, uint offset_first_element_in_bytes, uint stride_x, uint step_x, uint stride_y, uint step_y, uint stride_z, uint step_z) |
Wrap 3D tensor information into an tensor structure. More... | |
Tensor4D | update_tensor4D_workitem_ptr (__global uchar *ptr, uint offset_first_element_in_bytes, uint stride_x, uint step_x, uint stride_y, uint step_y, uint stride_z, uint step_z, uint stride_w, uint step_w, uint mod_size) |
__global const uchar * | vector_offset (const Vector *vec, int x) |
Get the pointer position of a Vector. More... | |
__global uchar * | offset (const Image *img, int x, int y) |
Get the pointer position of a Image. More... | |
__global const uchar * | tensor3D_offset (const Tensor3D *tensor, int x, int y, int z) |
Get the pointer position of a Tensor3D. More... | |
__global const uchar * | tensor4D_offset (const Tensor4D *tensor, int x, int y, int z, int w) |
Get the pointer position of a Tensor4D. More... | |
__global const uchar * | tensor3D_index2ptr (const Tensor3D *tensor, uint width, uint height, uint depth, uint index) |
Get the offset for a given linear index of a Tensor3D. More... | |
#define CLAMP | ( | x, | |
min_val, | |||
max_val | |||
) | min(max(x, min_val), max_val) |
#define CONCAT | ( | a, | |
b | |||
) | a##b |
#define CONVERT | ( | x, | |
type | |||
) | CONVERT_STR(x, type) |
Definition at line 522 of file helpers.h.
Referenced by bilinear_interpolate_with_border(), get_invsqrt_quantized_multiplier_exp(), and pooling_layer_3().
#define CONVERT_SAT | ( | x, | |
type | |||
) | CONVERT_SAT_STR(x, type) |
Definition at line 525 of file helpers.h.
Referenced by bilinear_interpolate_with_border_quantized(), and quantize_qasymm8().
#define CONVERT_SAT_ROUND | ( | x, | |
type, | |||
round | |||
) | CONVERT_SAT_ROUND_STR(x, type, round) |
#define CONVERT_TENSOR3D_TO_IMAGE_STRUCT_NO_STEP | ( | name | ) | update_image_from_tensor3D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, name##_step_z) |
Definition at line 641 of file helpers.h.
Referenced by fill_image_borders_constant(), and fill_image_borders_replicate().
#define CONVERT_TO_IMAGE_STRUCT | ( | name | ) | update_image_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x, name##_stride_y, name##_step_y) |
Definition at line 632 of file helpers.h.
Referenced by get_invsqrt_quantized_multiplier_exp(), non_max_suppression(), remap_bilinear_nchw(), remap_nearest_neighbour_nchw(), scale_bilinear_nchw(), scale_bilinear_quantized_nchw(), and scale_nearest_neighbour_nchw().
#define CONVERT_TO_IMAGE_STRUCT_NO_STEP | ( | name | ) | update_image_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0) |
Definition at line 635 of file helpers.h.
Referenced by remap_bilinear_nchw(), remap_nearest_neighbour_nchw(), roi_pooling_layer(), scale_bilinear_nchw(), scale_bilinear_quantized_nchw(), and scale_nearest_neighbour_nchw().
#define CONVERT_TO_TENSOR3D_STRUCT | ( | name | ) |
Definition at line 647 of file helpers.h.
Referenced by calculate_avg_scale(), deconvolution_upsample(), max_unpooling_layer_2(), pooling_layer_2(), pooling_layer_2_nchw_indices_fp16(), pooling_layer_2_nchw_indices_fp32(), pooling_layer_3(), reshape_layer(), upsample_layer_nchw(), and upsample_layer_nhwc().
#define CONVERT_TO_TENSOR3D_STRUCT_NO_STEP | ( | name | ) | update_tensor3D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, 0) |
Definition at line 651 of file helpers.h.
Referenced by deconvolution_upsample(), reshape_layer(), and roi_pooling_layer().
#define CONVERT_TO_TENSOR3D_STRUCT_NO_UPDATE_PTR | ( | name | ) |
Definition at line 661 of file helpers.h.
Referenced by max_unpooling_layer_2().
#define CONVERT_TO_TENSOR4D_STRUCT | ( | name, | |
mod_size | |||
) |
Definition at line 654 of file helpers.h.
Referenced by strided_slice().
#define CONVERT_TO_TENSOR4D_STRUCT_NO_STEP | ( | name, | |
mod_size | |||
) | update_tensor4D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, 0, name##_stride_w, 0, mod_size) |
Definition at line 658 of file helpers.h.
Referenced by strided_slice().
#define CONVERT_TO_VECTOR_STRUCT | ( | name | ) | update_vector_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x) |
Definition at line 626 of file helpers.h.
Referenced by get_invsqrt_quantized_multiplier_exp().
#define CONVERT_TO_VECTOR_STRUCT_NO_STEP | ( | name | ) | update_vector_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0) |
Definition at line 629 of file helpers.h.
Referenced by deconvolution_upsample().
#define CONVERT_VECTOR_SIZE_TO_PIXEL_UNIT | ( | vec_size | ) | CONVERT_VECTOR_SIZE_TO_PIXEL_UNIT_STR(vec_size) |
#define CONVERT_VECTOR_SIZE_TO_PIXEL_UNIT_STR | ( | vec_size | ) | PIXEL_UNIT##vec_size |
#define EXPAND | ( | x | ) | x |
#define IMAGE_DECLARATION | ( | name | ) |
#define MAX_REDUCE | ( | x, | |
size | |||
) | MAX_REDUCE_STR(x, size) |
#define max_reduce_16 | ( | x | ) | max(max_reduce_8((x).s01234567), max_reduce_8((x).s89ABCDEF)) |
#define max_reduce_3 | ( | x | ) | max(max_reduce_2((x).s01), ((x).s2)) |
#define max_reduce_4 | ( | x | ) | max(max_reduce_2((x).s01), max_reduce_2((x).s23)) |
#define max_reduce_8 | ( | x | ) | max(max_reduce_4((x).s0123), max_reduce_4((x).s4567)) |
#define PROD_REDUCE | ( | x, | |
size | |||
) | PROD_REDUCE_STR(x, size) |
#define prod_reduce_16 | ( | x | ) | prod_reduce_8((x).s01234567) * prod_reduce_8((x).s89ABCDEF) |
#define prod_reduce_3 | ( | x | ) | prod_reduce_2((x).s01) * ((x).s2) |
#define prod_reduce_4 | ( | x | ) | prod_reduce_2((x).s01) * prod_reduce_2((x).s23) |
#define prod_reduce_8 | ( | x | ) | prod_reduce_4((x).s0123) * prod_reduce_4((x).s4567) |
#define READ_IMAGE2D | ( | data_type, | |
n0, | |||
img, | |||
x_coord, | |||
y_coord | |||
) | READ_IMAGE2D_STR(data_type, n0, img, x_coord, y_coord) |
#define read_image2d_floatx1 | ( | img, | |
x_coord, | |||
y_coord | |||
) | (float4)(read_imagef(img, (int2)(x_coord, y_coord))); |
#define read_image2d_floatx2 | ( | img, | |
x_coord, | |||
y_coord | |||
) | (float8)(read_imagef(img, (int2)(x_coord, y_coord)), read_imagef(img, (int2)(x_coord + 1, y_coord))); |
#define read_image2d_floatx4 | ( | img, | |
x_coord, | |||
y_coord | |||
) | (float16)(read_imagef(img, (int2)(x_coord, y_coord)), read_imagef(img, (int2)(x_coord + 1, y_coord)), read_imagef(img, (int2)(x_coord + 2, y_coord)), read_imagef(img, (int2)(x_coord + 3, y_coord))); |
#define READ_IMAGE2D_STR | ( | data_type, | |
n0, | |||
img, | |||
x_coord, | |||
y_coord | |||
) | read_image2d_##data_type##x##n0(img, x_coord, y_coord) |
Utility macro to read a 2D OpenCL image object.
[in] | data_type | Data type |
[in] | n0 | Number of pixel to read. Only 1,2 and 4 is supported |
[in] | img | OpenCL image object |
[in] | x_coord | The x coordinate for the top-left pixel |
[in] | y_coord | The y coordinate for the top-left pixel |
#define REVERSE | ( | x, | |
s | |||
) | REVERSE_STR(x, s) |
#define ROTATE | ( | x, | |
s, | |||
n | |||
) | ROTATE_STR(x, s, n) |
#define SELECT_DATA_TYPE | ( | type | ) | SELECT_VEC_DATA_TYPE_STR(type, 1) |
#define SELECT_VEC_DATA_TYPE | ( | type, | |
size | |||
) | SELECT_VEC_DATA_TYPE_STR(type, size) |
#define SIGNED_INT_DATA_TYPE | ( | type | ) | SIGNED_INT_VEC_DATA_TYPE_STR(type, 1) |
#define SIGNED_INT_VEC_DATA_TYPE | ( | type, | |
size | |||
) | SIGNED_INT_VEC_DATA_TYPE_STR(type, size) |
#define SUM_REDUCE | ( | x, | |
size | |||
) | SUM_REDUCE_STR(x, size) |
#define sum_reduce_16 | ( | x | ) | sum_reduce_8((x).s01234567) + sum_reduce_8((x).s89ABCDEF) |
#define sum_reduce_3 | ( | x | ) | sum_reduce_2((x).s01) + ((x).s2) |
#define sum_reduce_4 | ( | x | ) | sum_reduce_2((x).s01) + sum_reduce_2((x).s23) |
#define sum_reduce_8 | ( | x | ) | sum_reduce_4((x).s0123) + sum_reduce_4((x).s4567) |
#define TENSOR3D_DECLARATION | ( | name | ) |
Definition at line 604 of file helpers.h.
Referenced by calculate_avg_scale(), deconvolution_upsample(), and pooling_layer_3().
#define TENSOR4D_DECLARATION | ( | name | ) |
#define VEC_DATA_TYPE | ( | type, | |
size | |||
) | VEC_DATA_TYPE_STR(type, size) |
Definition at line 519 of file helpers.h.
Referenced by bilinear_interpolate_with_border(), bilinear_interpolate_with_border_quantized(), calculate_avg_scale(), cast_down(), cast_up(), get_invsqrt_quantized_multiplier_exp(), non_max_suppression(), pooling_layer_2(), pooling_layer_3(), read_texels4(), roi_pool_1x1(), roi_pooling_layer(), strided_slice(), upsample_layer_nchw(), upsample_layer_nhwc(), vec16_max(), vec4_max(), and vec8_max().
#define VEC_OFFS | ( | dt, | |
s | |||
) | VEC_OFFS_STR(dt, s) |
#define VECTOR_DECLARATION | ( | name | ) |
Definition at line 590 of file helpers.h.
Referenced by deconvolution_upsample(), and get_invsqrt_quantized_multiplier_exp().
#define VLOAD | ( | size | ) | VLOAD_STR(size) |
Definition at line 203 of file helpers.h.
Referenced by cast_down(), cast_up(), get_invsqrt_quantized_multiplier_exp(), pooling_layer_2_nchw_indices_fp16(), pooling_layer_2_nchw_indices_fp32(), roi_pool_1x1(), and strided_slice().
#define VSTORE | ( | size | ) | VSTORE_STR(size) |
Definition at line 249 of file helpers.h.
Referenced by get_invsqrt_quantized_multiplier_exp(), and strided_slice().
#define vstore1 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) | *(OFFSET + PTR) = DATA |
#define VSTORE_PARTIAL | ( | size, | |
store_size | |||
) | VSTORE_PARTIAL_STR(size, store_size) |
#define vstore_partial_1 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) | vstore1(DATA.s0, OFFSET, PTR); |
#define vstore_partial_10 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define vstore_partial_11 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define vstore_partial_12 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define vstore_partial_13 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define vstore_partial_14 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define vstore_partial_15 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define vstore_partial_16 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) | vstore16(DATA, OFFSET, PTR); |
#define vstore_partial_16_1 vstore_partial_1 |
#define vstore_partial_16_10 vstore_partial_10 |
#define vstore_partial_16_11 vstore_partial_11 |
#define vstore_partial_16_12 vstore_partial_12 |
#define vstore_partial_16_13 vstore_partial_13 |
#define vstore_partial_16_14 vstore_partial_14 |
#define vstore_partial_16_15 vstore_partial_15 |
#define vstore_partial_16_16 vstore_partial_16 |
#define vstore_partial_16_2 vstore_partial_2 |
#define vstore_partial_16_3 vstore_partial_3 |
#define vstore_partial_16_4 vstore_partial_4 |
#define vstore_partial_16_5 vstore_partial_5 |
#define vstore_partial_16_6 vstore_partial_6 |
#define vstore_partial_16_7 vstore_partial_7 |
#define vstore_partial_16_8 vstore_partial_8 |
#define vstore_partial_16_9 vstore_partial_9 |
#define vstore_partial_2 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) | vstore2(DATA.s01, OFFSET, PTR); |
#define vstore_partial_2_1 vstore_partial_1 |
#define vstore_partial_2_2 vstore_partial_2 |
#define vstore_partial_3 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) | vstore3(DATA.s012, OFFSET, PTR); |
#define vstore_partial_3_1 vstore_partial_1 |
#define vstore_partial_3_2 vstore_partial_2 |
#define vstore_partial_3_3 vstore_partial_3 |
#define vstore_partial_4 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) | vstore4(DATA.s0123, OFFSET, PTR); |
#define vstore_partial_4_1 vstore_partial_1 |
#define vstore_partial_4_2 vstore_partial_2 |
#define vstore_partial_4_3 vstore_partial_3 |
#define vstore_partial_4_4 vstore_partial_4 |
#define vstore_partial_5 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define vstore_partial_6 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define vstore_partial_7 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define vstore_partial_8 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) | vstore8(DATA.s01234567, OFFSET, PTR); |
#define vstore_partial_8_1 vstore_partial_1 |
#define vstore_partial_8_2 vstore_partial_2 |
#define vstore_partial_8_3 vstore_partial_3 |
#define vstore_partial_8_4 vstore_partial_4 |
#define vstore_partial_8_5 vstore_partial_5 |
#define vstore_partial_8_6 vstore_partial_6 |
#define vstore_partial_8_7 vstore_partial_7 |
#define vstore_partial_8_8 vstore_partial_8 |
#define vstore_partial_9 | ( | DATA, | |
OFFSET, | |||
PTR | |||
) |
#define VSTORE_PARTIAL_STR | ( | size, | |
store_size | |||
) | vstore_partial_##size##_##store_size |
|
inline |
Get the pointer position of a Image.
[in] | img | Pointer to the starting position of the buffer |
[in] | x | Relative X position |
[in] | y | Relative Y position |
Definition at line 861 of file helpers.h.
References Image::ptr, Image::stride_x, and Image::stride_y.
Referenced by bilinear_interpolate_with_border(), bilinear_interpolate_with_border_quantized(), arm_compute::cpu::bilinear_neon_scale(), arm_compute::test::validation::reference::bounding_box_transform(), arm_compute::scale_utils::calculate_resize_ratio(), arm_compute::test::colorconvert_helper::detail::colorconvert_nv12_to_rgb(), arm_compute::test::colorconvert_helper::detail::colorconvert_yuyv_to_iyuv(), arm_compute::test::colorconvert_helper::detail::colorconvert_yuyv_to_nv12(), arm_compute::test::colorconvert_helper::detail::colorconvert_yuyv_to_rgb(), CpuConcatenate::configure(), ClConcatenate::configure(), CLTileKernel::configure(), arm_gemm::ConvolutionInterleave(), ITensor::copy_from(), arm_compute::test::validation::DATA_TEST_CASE(), arm_compute::test::validation::reference::deconvolution_layer(), GemmHybridIndirect< strategy, To, Tr, OutputStage, SeparateQuantize >::execute(), MemoryRegion::extract_subregion(), AssetsLibrary::fill(), fill_image_borders_constant(), fill_image_borders_replicate(), arm_compute::test::validation::reference::gather(), get_invsqrt_quantized_multiplier_exp(), IMemoryRegion::IMemoryRegion(), IndirectOutputArg< T >::IndirectOutputArg(), MMappedFile::map(), OffsetLifetimeManager::mapping_type(), NEPriorBoxLayerKernel::name(), arm_compute::cpu::nearest_neon_scale(), non_max_suppression(), arm_compute::test::colorconvert_helper::detail::nv_to_iyuv(), arm_compute::test::colorconvert_helper::detail::nv_to_yuv4(), SubTensorInfo::offset_element_in_bytes(), TensorInfo::offset_element_in_bytes(), arm_compute::offset_int_vec(), SimpleTensor< uint8_t >::operator[](), PaddingCalculator::PaddingCalculator(), quantize_qasymm8(), RandomAccessor::RandomAccessor(), read_texels4(), arm_compute::test::validation::reference::reorg_layer(), roi_pooling_layer(), NEReorgLayerKernel::run(), CPPBoxWithNonMaximaSuppressionLimitKernel::run_nmslimit(), arm_compute::test::validation::reference::select(), PaddingCalculator::set_access_offset(), arm_compute::test::validation::reference::slice(), arm_compute::test::validation::reference::softmax_layer_generic(), strided_slice(), arm_compute::test::validation::reference::strided_slice(), arm_compute::test::validation::TEST_CASE(), CpuConcatenate::validate(), ClConcatenate::validate(), NEBoundingBoxTransformKernel::validate(), arm_compute::vdequantize(), arm_compute::vquantize(), and arm_compute::vquantize_signed().
|
inline |
Get the offset for a given linear index of a Tensor3D.
[in] | tensor | Pointer to the starting position of the buffer |
[in] | width | Width of the input tensor |
[in] | height | Height of the input tensor |
[in] | depth | Depth of the input tensor |
[in] | index | Linear index |
Definition at line 899 of file helpers.h.
References Tensor3D::offset_first_element_in_bytes, Tensor3D::ptr, Tensor3D::stride_x, Tensor3D::stride_y, and Tensor3D::stride_z.
Referenced by max_unpooling_layer_2().
|
inline |
Get the pointer position of a Tensor3D.
[in] | tensor | Pointer to the starting position of the buffer |
[in] | x | Relative X position |
[in] | y | Relative Y position |
[in] | z | Relative Z position |
Definition at line 873 of file helpers.h.
References Tensor3D::ptr, Tensor3D::stride_x, Tensor3D::stride_y, and Tensor3D::stride_z.
Referenced by calculate_avg_scale(), deconvolution_upsample(), pooling_layer_2(), pooling_layer_2_nchw_indices_fp16(), pooling_layer_2_nchw_indices_fp32(), pooling_layer_3(), reshape_layer(), roi_pool_1x1(), roi_pooling_layer(), upsample_layer_nchw(), and upsample_layer_nhwc().
|
inline |
Wrap 3D tensor information into an tensor structure.
[in] | ptr | Pointer to the starting postion of the buffer |
[in] | offset_first_element_in_bytes | The offset of the first element in the source image |
[in] | stride_x | Stride of the image in X dimension (in bytes) |
[in] | step_x | stride_x * number of elements along X processed per workitem(in bytes) |
[in] | stride_y | Stride of the image in Y dimension (in bytes) |
[in] | step_y | stride_y * number of elements along Y processed per workitem(in bytes) |
[in] | stride_z | Stride of the image in Z dimension (in bytes) |
[in] | step_z | stride_z * number of elements along Z processed per workitem(in bytes) |
Definition at line 814 of file helpers.h.
References Vector::offset_first_element_in_bytes, Vector::ptr, Tensor3D::ptr, and Vector::stride_x.
|
inline |
Get the pointer position of a Tensor4D.
[in] | tensor | Pointer to the starting position of the buffer |
[in] | x | Relative X position |
[in] | y | Relative Y position |
[in] | z | Relative Z position |
[in] | w | Relative W position |
Definition at line 886 of file helpers.h.
References Tensor4D::ptr, Tensor4D::stride_w, Tensor4D::stride_x, Tensor4D::stride_y, and Tensor4D::stride_z.
|
inline |
Wrap 3D tensor information into an image structure, and make the pointer point at this workitem's data.
[in] | ptr | Pointer to the starting postion of the buffer |
[in] | offset_first_element_in_bytes | The offset of the first element in the source image |
[in] | stride_x | Stride of the image in X dimension (in bytes) |
[in] | step_x | stride_x * number of elements along X processed per workitem(in bytes) |
[in] | stride_y | Stride of the image in Y dimension (in bytes) |
[in] | step_y | stride_y * number of elements along Y processed per workitem(in bytes) |
[in] | stride_z | Stride of the image in Z dimension (in bytes) |
[in] | step_z | stride_z * number of elements along Z processed per workitem(in bytes) |
Definition at line 761 of file helpers.h.
References Vector::offset_first_element_in_bytes, Image::offset_first_element_in_bytes, Vector::ptr, Image::ptr, and Vector::stride_x.
|
inline |
Wrap image information into an Image structure, and make the pointer point at this workitem's data.
[in] | ptr | Pointer to the starting postion of the buffer |
[in] | offset_first_element_in_bytes | The offset of the first element in the source image |
[in] | stride_x | Stride of the image in X dimension (in bytes) |
[in] | step_x | stride_x * number of elements along X processed per workitem(in bytes) |
[in] | stride_y | Stride of the image in Y dimension (in bytes) |
[in] | step_y | stride_y * number of elements along Y processed per workitem(in bytes) |
Definition at line 735 of file helpers.h.
References Vector::offset_first_element_in_bytes, Image::offset_first_element_in_bytes, Vector::ptr, Image::ptr, and Vector::stride_x.
|
inline |
Wrap 3D tensor information into an tensor structure, and make the pointer point at this workitem's data.
[in] | ptr | Pointer to the starting postion of the buffer |
[in] | offset_first_element_in_bytes | The offset of the first element in the source image |
[in] | stride_x | Stride of the image in X dimension (in bytes) |
[in] | step_x | stride_x * number of elements along X processed per workitem(in bytes) |
[in] | stride_y | Stride of the image in Y dimension (in bytes) |
[in] | step_y | stride_y * number of elements along Y processed per workitem(in bytes) |
[in] | stride_z | Stride of the image in Z dimension (in bytes) |
[in] | step_z | stride_z * number of elements along Z processed per workitem(in bytes) |
Definition at line 787 of file helpers.h.
References Vector::offset_first_element_in_bytes, Tensor3D::offset_first_element_in_bytes, Vector::ptr, Tensor3D::ptr, and Vector::stride_x.
|
inline |
Definition at line 827 of file helpers.h.
References Vector::offset_first_element_in_bytes, Tensor4D::offset_first_element_in_bytes, Vector::ptr, Tensor4D::ptr, and Vector::stride_x.
|
inline |
Wrap vector information into an Vector structure, and make the pointer point at this workitem's data.
[in] | ptr | Pointer to the starting postion of the buffer |
[in] | offset_first_element_in_bytes | The offset of the first element in the source vector |
[in] | stride_x | Stride of the vector in X dimension (in bytes) |
[in] | step_x | stride_x * number of elements along X processed per workitem(in bytes) |
Definition at line 712 of file helpers.h.
References Vector::offset_first_element_in_bytes, Vector::ptr, and Vector::stride_x.
|
inline |
Get the pointer position of a Vector.
[in] | vec | Pointer to the starting position of the buffer |
[in] | x | Relative X position |
Definition at line 850 of file helpers.h.
References Vector::ptr, and Vector::stride_x.
Referenced by deconvolution_upsample(), and get_invsqrt_quantized_multiplier_exp().