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
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
|
/* ************************************************************************
* Copyright (C) 2016-2024 Advanced Micro Devices, Inc. All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*
* ************************************************************************ */
#include <fstream>
#include <iostream>
#include <stdlib.h>
#include <vector>
#include "testing_common.hpp"
using hipblasGetrfModel = ArgumentModel<e_a_type, e_N, e_lda>;
inline void testname_getrf(const Arguments& arg, std::string& name)
{
hipblasGetrfModel{}.test_name(arg, name);
}
template <typename T>
void testing_getrf_bad_arg(const Arguments& arg)
{
bool FORTRAN = arg.api == hipblas_client_api::FORTRAN;
auto hipblasGetrfFn = FORTRAN ? hipblasGetrf<T, true> : hipblasGetrf<T, false>;
hipblasLocalHandle handle(arg);
int64_t N = 101;
int64_t M = N;
int64_t lda = 102;
int64_t Ipiv_size = std::min(M, N);
host_matrix<T> hA(M, N, lda);
// Allocate device memory
device_matrix<T> dA(M, N, lda);
device_matrix<int> dIpiv(1, Ipiv_size, 1);
device_vector<int> dInfo(1);
EXPECT_HIPBLAS_STATUS(hipblasGetrfFn(nullptr, N, dA, lda, dIpiv, dInfo),
HIPBLAS_STATUS_NOT_INITIALIZED);
EXPECT_HIPBLAS_STATUS(hipblasGetrfFn(handle, -1, dA, lda, dIpiv, dInfo),
HIPBLAS_STATUS_INVALID_VALUE);
EXPECT_HIPBLAS_STATUS(hipblasGetrfFn(handle, N, dA, N - 1, dIpiv, dInfo),
HIPBLAS_STATUS_INVALID_VALUE);
// If N == 0, A and ipiv can be nullptr
CHECK_HIPBLAS_ERROR(hipblasGetrfFn(handle, 0, nullptr, lda, nullptr, dInfo));
if(arg.bad_arg_all)
{
EXPECT_HIPBLAS_STATUS(hipblasGetrfFn(handle, N, nullptr, lda, dIpiv, dInfo),
HIPBLAS_STATUS_INVALID_VALUE);
EXPECT_HIPBLAS_STATUS(hipblasGetrfFn(handle, N, dA, lda, dIpiv, nullptr),
HIPBLAS_STATUS_INVALID_VALUE);
}
}
template <typename T>
void testing_getrf(const Arguments& arg)
{
using U = real_t<T>;
bool FORTRAN = arg.api == hipblas_client_api::FORTRAN;
auto hipblasGetrfFn = FORTRAN ? hipblasGetrf<T, true> : hipblasGetrf<T, false>;
int M = arg.N;
int N = arg.N;
int lda = arg.lda;
int Ipiv_size = std::min(M, N);
// Check to prevent memory allocation error
if(M < 0 || N < 0 || lda < M)
{
return;
}
// Naming: dK is in GPU (device) memory. hK is in CPU (host) memory
host_matrix<T> hA(M, N, lda);
host_matrix<T> hA1(M, N, lda);
host_matrix<int> hIpiv(1, Ipiv_size, 1);
host_matrix<int64_t> hIpiv64(1, Ipiv_size, 1);
host_vector<int> hInfo(1);
host_vector<int> hInfo1(1);
// Allocate device memory
device_matrix<T> dA(M, N, lda);
device_matrix<int> dIpiv(1, Ipiv_size, 1);
device_vector<int> dInfo(1);
// Check device memory allocation
CHECK_DEVICE_ALLOCATION(dA.memcheck());
CHECK_DEVICE_ALLOCATION(dIpiv.memcheck());
CHECK_DEVICE_ALLOCATION(dInfo.memcheck());
double gpu_time_used, hipblas_error;
hipblasLocalHandle handle(arg);
// Initial hA on CPU
hipblas_init_matrix(hA, arg, hipblas_client_never_set_nan, hipblas_general_matrix, true);
T* A = (T*)hA;
// scale A to avoid singularities
for(int i = 0; i < M; i++)
{
for(int j = 0; j < N; j++)
{
if(i == j)
A[i + j * lda] += 400;
else
A[i + j * lda] -= 4;
}
}
// Copy data from CPU to device
CHECK_HIP_ERROR(dA.transfer_from(hA));
CHECK_HIP_ERROR(hipMemset(dIpiv, 0, Ipiv_size * sizeof(int)));
CHECK_HIP_ERROR(hipMemset(dInfo, 0, sizeof(int)));
if(arg.unit_check || arg.norm_check)
{
/* =====================================================================
HIPBLAS
=================================================================== */
CHECK_HIPBLAS_ERROR(hipblasGetrfFn(handle, N, dA, lda, dIpiv, dInfo));
// Copy output from device to CPU
CHECK_HIP_ERROR(hA1.transfer_from(dA));
CHECK_HIP_ERROR(hIpiv.transfer_from(dIpiv));
CHECK_HIP_ERROR(hipMemcpy(hInfo1, dInfo, sizeof(int), hipMemcpyDeviceToHost));
/* =====================================================================
CPU LAPACK
=================================================================== */
for(int i = 0; i < Ipiv_size; i++)
hIpiv64[0][i] = hIpiv[0][i];
hInfo[0] = ref_getrf(M, N, hA.data(), lda, hIpiv64.data());
hipblas_error = norm_check_general<T>('F', M, N, lda, hA, hA1);
if(arg.unit_check)
{
U eps = std::numeric_limits<U>::epsilon();
double tolerance = eps * 2000;
unit_check_error(hipblas_error, tolerance);
}
}
if(arg.timing)
{
hipStream_t stream;
CHECK_HIPBLAS_ERROR(hipblasGetStream(handle, &stream));
int runs = arg.cold_iters + arg.iters;
for(int iter = 0; iter < runs; iter++)
{
if(iter == arg.cold_iters)
gpu_time_used = get_time_us_sync(stream);
CHECK_HIPBLAS_ERROR(hipblasGetrfFn(handle, N, dA, lda, dIpiv, dInfo));
}
gpu_time_used = get_time_us_sync(stream) - gpu_time_used;
hipblasGetrfModel{}.log_args<T>(std::cout,
arg,
gpu_time_used,
getrf_gflop_count<T>(N, M),
ArgumentLogging::NA_value,
hipblas_error);
}
}
|
