Blame - src/core/GLES_COMPUTE/cs_shaders/concatenate.cs - ml/ComputeLibrary

Anthony Barbier

7068f99

2017-10-26 15:23:08 +0100

[diff] [blame^]

/*

*

* SPDX-License-Identifier: MIT

5

*

6

* Permission is hereby granted, free of charge, to any person obtaining a copy

7

* of this software and associated documentation files (the "Software"), to

8

* deal in the Software without restriction, including without limitation the

9

* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or

10

* sell copies of the Software, and to permit persons to whom the Software is

11

* furnished to do so, subject to the following conditions:

12

*

13

* The above copyright notice and this permission notice shall be included in all

14

* copies or substantial portions of the Software.

15

*

16

* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR

17

* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

18

* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE

19

* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER

20

* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,

21

* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE

22

* SOFTWARE.

23

*/

24

layout(local_size_x = LOCAL_SIZE_X, local_size_y = LOCAL_SIZE_Y, local_size_z = LOCAL_SIZE_Z) in;

25

#include "helpers.h"

26

27

#ifdef DATA_TYPE_FP32

28

precision highp float;

29

30

layout(std140) uniform shader_params

31

{

32

TENSOR3D_PARAM_DECLARATION(src);

33

TENSOR3D_PARAM_DECLARATION(dst);

34

};

35

36

BUFFER_DECLARATION(src, 1, float, readonly);

37

BUFFER_DECLARATION(dst, 2, float, writeonly);

38

39

/** This kernel concatenates the input tensor into the output tensor along the third dimension

40

*

41

* @param[in] src_ptr Pointer to the source tensor. Supported data types: F32

42

* @param[in] src_stride_x Stride of the source tensor in X dimension (in bytes)

43

* @param[in] src_step_x src_stride_x * number of elements along X processed per workitem(in bytes)

44

* @param[in] src_stride_y Stride of the source tensor in Y dimension (in bytes)

45

* @param[in] src_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)

46

* @param[in] src_stride_z Stride of the source tensor in Z dimension (in bytes)

47

* @param[in] src_step_z src_stride_z * number of elements along Z processed per workitem(in bytes)

48

* @param[in] src_offset_first_element_in_bytes The offset of the first element in the source tensor

49

* @param[out] dst_ptr Pointer to the destination tensor. Supported data types: same as @p src_ptr

50

* @param[in] dst_stride_x Stride of the destination tensor in X dimension (in bytes)

51

* @param[in] dst_step_x dst_stride_x * number of elements along X processed per workitem(in bytes)

52

* @param[in] dst_stride_y Stride of the destination tensor in Y dimension (in bytes)

53

* @param[in] dst_step_y dst_stride_y * number of elements along Y processed per workitem(in bytes)

54

* @param[in] dst_stride_z Stride of the source tensor in Z dimension (in bytes)

55

* @param[in] dst_step_z dst_stride_z * number of elements along Z processed per workitem(in bytes)

56

* @param[in] dst_offset_first_element_in_bytes The offset of the first element in the destination tensor

*/

void main(void)

{

Tensor3D src = CONVERT_TO_TENSOR3D_STRUCT(src);

61

Tensor3D dst = CONVERT_TO_TENSOR3D_STRUCT(dst);

62

63

dst_ptr[dst.current_offset + uint(OFFSETS_Z >> 2)] = src_ptr[tensor3D_offset(src, -OFFSETS_X, -OFFSETS_Y, 0)];

64

}

65

66

#elif defined(DATA_TYPE_FP16)

67

precision mediump float;

68

69

layout(std140) uniform shader_params

70

{

71

TENSOR3D_PARAM_DECLARATION(src);

72

TENSOR3D_PARAM_DECLARATION(dst);

73

};

74

75

BUFFER_DECLARATION(src, 1, uvec2, readonly);

76

BUFFER_DECLARATION(dst, 2, uvec2, writeonly);

77

78