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
115
116
117
|
// =================================================================================================
// 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 Xher class (see the header for information about the class).
//
// =================================================================================================
#include "routines/level2/xher.hpp"
#include <string>
namespace clblast {
// =================================================================================================
// Constructor: forwards to base class constructor
template <typename T, typename U>
Xher<T,U>::Xher(Queue &queue, EventPointer event, const std::string &name):
Routine(queue, event, name, {"Xger"}, PrecisionValue<T>()) {
source_string_ =
#include "../../kernels/level2/level2.opencl"
#include "../../kernels/level2/xher.opencl"
;
}
// =================================================================================================
// Specializations to compute alpha of type 'T'
template <> float2 Xher<float2,float>::GetAlpha(const float alpha) { return float2{alpha, 0.0f}; }
template <> double2 Xher<double2,double>::GetAlpha(const double alpha) { return double2{alpha, 0.0}; }
template <> float Xher<float,float>::GetAlpha(const float alpha) { return alpha; }
template <> double Xher<double,double>::GetAlpha(const double alpha) { return alpha; }
template <> half Xher<half,half>::GetAlpha(const half alpha) { return alpha; }
// =================================================================================================
// The main routine
template <typename T, typename U>
StatusCode Xher<T,U>::DoHer(const Layout layout, const Triangle triangle,
const size_t n,
const U alpha,
const Buffer<T> &x_buffer, const size_t x_offset, const size_t x_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); }
else { status = TestMatrixA(n, n, a_buffer, a_offset, a_ld); }
if (ErrorIn(status)) { return status; }
status = TestVectorX(n, x_buffer, x_offset, x_inc);
if (ErrorIn(status)) { return status; }
// If alpha is zero an update is not required
if (alpha == U{0}) { return StatusCode::kSuccess; }
// Creates a matching version of alpha
const auto matching_alpha = GetAlpha(alpha);
// Upload the scalar argument as a constant buffer to the device (needed for half-precision)
auto alpha_buffer = Buffer<T>(context_, 1);
alpha_buffer.Write(queue_, 1, &matching_alpha);
// Retrieves the kernel from the compiled binary
try {
const auto program = GetProgramFromCache(context_, PrecisionValue<T>(), routine_name_);
auto kernel = Kernel(program, "Xher");
// Sets the kernel arguments
kernel.SetArgument(0, static_cast<int>(n));
kernel.SetArgument(1, alpha_buffer());
kernel.SetArgument(2, x_buffer());
kernel.SetArgument(3, static_cast<int>(x_offset));
kernel.SetArgument(4, static_cast<int>(x_inc));
kernel.SetArgument(5, a_buffer());
kernel.SetArgument(6, static_cast<int>(a_offset));
kernel.SetArgument(7, static_cast<int>(a_ld));
kernel.SetArgument(8, static_cast<int>(is_upper));
kernel.SetArgument(9, 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, queue_, device_, global, local, event_);
if (ErrorIn(status)) { return status; }
// Succesfully finished the computation
return StatusCode::kSuccess;
} catch (...) { return StatusCode::kInvalidKernel; }
}
// =================================================================================================
// Compiles the templated class
template class Xher<half, half>;
template class Xher<float, float>;
template class Xher<double, double>;
template class Xher<float2, float>;
template class Xher<double2, double>;
// =================================================================================================
} // namespace clblast
|
