1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
|
// =================================================================================================
// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. This
// project loosely follows the Google C++ styleguide and uses a tab-size of two spaces and a max-
// width of 100 characters per line.
//
// Author(s):
// Cedric Nugteren <www.cedricnugteren.nl>
//
// This file implements the Xher2 class (see the header for information about the class).
//
// =================================================================================================
#include "internal/routines/level2/xher2.h"
#include <string>
namespace clblast {
// =================================================================================================
// Specific implementations to get the memory-type based on a template argument
template <> const Precision Xher2<float>::precision_ = Precision::kSingle;
template <> const Precision Xher2<double>::precision_ = Precision::kDouble;
template <> const Precision Xher2<float2>::precision_ = Precision::kComplexSingle;
template <> const Precision Xher2<double2>::precision_ = Precision::kComplexDouble;
// =================================================================================================
// Constructor: forwards to base class constructor
template <typename T>
Xher2<T>::Xher2(Queue &queue, Event &event, const std::string &name):
Routine<T>(queue, event, name, {"Xger"}, precision_) {
source_string_ =
#include "../../kernels/level2/level2.opencl"
#include "../../kernels/level2/xher2.opencl"
;
}
// =================================================================================================
// The main routine
template <typename T>
StatusCode Xher2<T>::DoHer2(const Layout layout, const Triangle triangle,
const size_t n,
const T alpha,
const Buffer<T> &x_buffer, const size_t x_offset, const size_t x_inc,
const Buffer<T> &y_buffer, const size_t y_offset, const size_t y_inc,
const Buffer<T> &a_buffer, const size_t a_offset, const size_t a_ld,
const bool packed) {
// Makes sure the dimensions are larger than zero
if (n == 0) { return StatusCode::kInvalidDimension; }
// The data is either in the upper or lower triangle
const auto is_upper = ((triangle == Triangle::kUpper && layout != Layout::kRowMajor) ||
(triangle == Triangle::kLower && layout == Layout::kRowMajor));
const auto is_rowmajor = (layout == Layout::kRowMajor);
// Tests the matrix and the vectors for validity
auto status = StatusCode::kSuccess;
if (packed) { status = TestMatrixAP(n, a_buffer, a_offset, sizeof(T)); }
else { status = TestMatrixA(n, n, a_buffer, a_offset, a_ld, sizeof(T)); }
if (ErrorIn(status)) { return status; }
status = TestVectorX(n, x_buffer, x_offset, x_inc, sizeof(T));
if (ErrorIn(status)) { return status; }
status = TestVectorY(n, y_buffer, y_offset, y_inc, sizeof(T));
if (ErrorIn(status)) { return status; }
// Retrieves the Xgemv kernel from the compiled binary
try {
auto& program = GetProgramFromCache();
auto kernel = Kernel(program, "Xher2");
// Sets the kernel arguments
kernel.SetArgument(0, static_cast<int>(n));
kernel.SetArgument(1, alpha);
kernel.SetArgument(2, x_buffer());
kernel.SetArgument(3, static_cast<int>(x_offset));
kernel.SetArgument(4, static_cast<int>(x_inc));
kernel.SetArgument(5, y_buffer());
kernel.SetArgument(6, static_cast<int>(y_offset));
kernel.SetArgument(7, static_cast<int>(y_inc));
kernel.SetArgument(8, a_buffer());
kernel.SetArgument(9, static_cast<int>(a_offset));
kernel.SetArgument(10, static_cast<int>(a_ld));
kernel.SetArgument(11, static_cast<int>(is_upper));
kernel.SetArgument(12, static_cast<int>(is_rowmajor));
// Launches the kernel
auto global_one = Ceil(CeilDiv(n, db_["WPT"]), db_["WGS1"]);
auto global_two = Ceil(CeilDiv(n, db_["WPT"]), db_["WGS2"]);
auto global = std::vector<size_t>{global_one, global_two};
auto local = std::vector<size_t>{db_["WGS1"], db_["WGS2"]};
status = RunKernel(kernel, global, local);
if (ErrorIn(status)) { return status; }
// Waits for all kernels to finish
queue_.Finish();
// Succesfully finished the computation
return StatusCode::kSuccess;
} catch (...) { return StatusCode::kInvalidKernel; }
}
// =================================================================================================
// Compiles the templated class
template class Xher2<float>;
template class Xher2<double>;
template class Xher2<float2>;
template class Xher2<double2>;
// =================================================================================================
} // namespace clblast
|
