Gitiles
Code Review Sign In
realtimeroboticsgroup.org / RealtimeRoboticsGroup / test / 9a24b370cbc07492fa6c205f1027367fa55b6742 / . / test_problems / test_s_strmat.c
blob: 456db87051a3e653464308c0b0a6d6f1a2301846 [file] [log] [blame]
Austin Schuh9a24b372018-01-28 16:12:29 -0800[diff] [blame^]1/**************************************************************************************************
2* *
3* This file is part of BLASFEO. *
4* *
5* BLASFEO -- BLAS For Embedded Optimization. *
6* Copyright (C) 2016-2017 by Gianluca Frison. *
7* Developed at IMTEK (University of Freiburg) under the supervision of Moritz Diehl. *
8* All rights reserved. *
9* *
10* HPMPC is free software; you can redistribute it and/or *
11* modify it under the terms of the GNU Lesser General Public *
12* License as published by the Free Software Foundation; either *
13* version 2.1 of the License, or (at your option) any later version. *
14* *
15* HPMPC is distributed in the hope that it will be useful, *
16* but WITHOUT ANY WARRANTY; without even the implied warranty of *
17* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. *
18* See the GNU Lesser General Public License for more details. *
19* *
20* You should have received a copy of the GNU Lesser General Public *
21* License along with HPMPC; if not, write to the Free Software *
22* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA *
23* *
24* Author: Gianluca Frison, giaf (at) dtu.dk *
25* gianluca.frison (at) imtek.uni-freiburg.de *
26* *
27**************************************************************************************************/
28
29#include <stdlib.h>
30#include <stdio.h>
31#include <sys/time.h>
32
33#include "../include/blasfeo_common.h"
34#include "../include/blasfeo_i_aux_ext_dep.h"
35#include "../include/blasfeo_s_aux_ext_dep.h"
36#include "../include/blasfeo_s_aux.h"
37#include "../include/blasfeo_s_kernel.h"
38#include "../include/blasfeo_s_blas.h"
39
40
41int main()
42 {
43
44#if defined(LA_HIGH_PERFORMANCE)
45
46 printf("\nLA provided by HIGH_PERFORMANCE\n\n");
47
48#elif defined(LA_REFERENCE)
49
50 printf("\nLA provided by REFERENCE\n\n");
51
52#elif defined(LA_BLAS)
53
54 printf("\nLA provided by BLAS\n\n");
55
56#else
57
58 printf("\nLA provided by ???\n\n");
59 exit(2);
60
61#endif
62
63 int ii, jj;
64
65 int n = 16;
66
67 //
68 // matrices in column-major format
69 //
70 float *A; s_zeros(&A, n, n);
71 for(ii=0; ii<n*n; ii++) A[ii] = ii;
72// for(jj=0; jj<n; jj++)
73// for(ii=0; ii<jj; ii++)
74// A[ii+n*jj] = 0.0/0.0;
75// s_print_mat(n, n, A, n);
76
77 float *B; s_zeros(&B, n, n);
78 for(ii=0; ii<n; ii++) B[ii*(n+1)] = 1.0;
79// s_print_mat(n, n, B, n);
80
81 float *D; s_zeros(&D, n, n);
82 for(ii=0; ii<n*n; ii++) D[ii] = -1.0;
83// s_print_mat(n, n, B, n);
84
85
86 //
87 // matrices in matrix struct format
88 //
89
90 struct s_strmat sA;
91 s_allocate_strmat(n, n, &sA);
92 s_cvt_mat2strmat(n, n, A, n, &sA, 0, 0);
93 s_print_strmat(n, n, &sA, 0, 0);
94
95 struct s_strmat sB;
96 s_allocate_strmat(n, n, &sB);
97 s_cvt_mat2strmat(n, n, B, n, &sB, 0, 0);
98 s_print_strmat(n, n, &sB, 0, 0);
99
100 struct s_strmat sD;
101 s_allocate_strmat(n, n, &sD);
102 s_cvt_mat2strmat(n, n, D, n, &sD, 0, 0);
103
104 struct s_strvec sx;
105 s_allocate_strvec(n, &sx);
106 sx.pa[7] = 1.0;
107 s_print_tran_strvec(n, &sx, 0);
108
109 struct s_strvec sz0;
110 s_allocate_strvec(n, &sz0);
111
112 struct s_strvec sz1;
113 s_allocate_strvec(n, &sz1);
114
115 //
116 // tests
117 //
118
119 float alpha = 1.0;
120 float beta = 0.0;
121// kernel_sgemm_nt_24x4_lib8(4, &alpha, sA.pA, sA.cn, sB.pA, &beta, sD.pA, sD.cn, sD.pA, sD.cn);
122// kernel_sgemm_nt_16x4_lib8(4, &alpha, sA.pA, sA.cn, sB.pA, &beta, sD.pA, sD.cn, sD.pA, sD.cn);
123// kernel_sgemm_nt_8x8_lib8(5, &alpha, sA.pA, sB.pA, &beta, sD.pA, sD.pA);
124// kernel_sgemm_nt_8x4_lib8(5, &alpha, sA.pA, sB.pA, &beta, sD.pA, sD.pA);
125// kernel_sgemm_nt_4x8_gen_lib8(8, &alpha, sA.pA, sB.pA, &beta, 0, sD.pA, sD.cn, 0, sD.pA, sD.cn, 0, 4, 0, 8);
126// kernel_sgemm_nt_8x4_vs_lib8(8, &alpha, sA.pA, sB.pA, &beta, sD.pA, sD.pA, 7, 4);
127// kernel_sgemm_nt_8x4_lib8(8, &alpha, sB.pA, sA.pA+4, &beta, sA.pA+4*8, sD.pA+4*8);
128// kernel_sgemm_nn_16x4_lib8(4, &alpha, sA.pA, sA.cn, 0, sB.pA, sB.cn, &beta, sD.pA, sD.cn, sD.pA, sD.cn);
129// kernel_sgemm_nt_12x4_lib4(4, &alpha, sA.pA, sA.cn, sB.pA, &beta, sD.pA, sD.cn, sD.pA, sD.cn);
130// kernel_sgemm_nt_8x8_lib4(8, &alpha, sA.pA, sA.cn, sB.pA, sB.cn, &beta, sD.pA, sD.cn, sD.pA, sD.cn);
131// kernel_sgemm_nt_8x4_lib4(2, &alpha, sA.pA, sA.cn, sB.pA, &beta, sD.pA, sD.cn, sD.pA, sD.cn);
132// s_print_strmat(n, n, &sD, 0, 0);
133// return 0;
134// sgemm_nt_libstr(n, n, 5, 1.0, &sA, 0, 0, &sB, 0, 0, 0.0, &sB, 0, 0, &sD, 0, 0);
135// ssyrk_ln_libstr(n, n, 1.0, &sA, 0, 0, &sB, 0, 0, 0.0, &sB, 0, 0, &sD, 0, 0);
136// ssyrk_ln_mn_libstr(n, n, n, 1.0, &sA, 0, 0, &sB, 0, 0, 0.0, &sB, 0, 0, &sD, 0, 0);
137// kernel_ssyrk_nt_l_8x8_lib8(n, &alpha, sA.pA, sA.pA, &beta, sB.pA, sD.pA);
138// sgecp_libstr(16, 16, &sA, 2, 0, &sD, 1, 0);
139// sgetr_libstr(16, 16, &sA, 2, 0, &sD, 2, 0);
140// s_print_strmat(n, n, &sD, 0, 0);
141// sgemv_n_libstr(6, 6, 1.0, &sA, 1, 0, &sx, 0, 0.0, &sz0, 0, &sz0, 0);
142// sgemv_t_libstr(11, 8, 1.0, &sA, 0, 0, &sx, 0, 0.0, &sz0, 0, &sz0, 0);
143// strmv_lnn_libstr(6, 6, &sA, 1, 0, &sx, 0, &sz0, 0);
144// strmv_ltn_libstr(10, 10, &sA, 1, 0, &sx, 0, &sz0, 0);
145// sA.pA[0] = 1.0;
146// strsv_lnn_libstr(10, &sA, 0, 0, &sx, 0, &sz0, 0);
147// for(ii=0; ii<8; ii++) sA.dA[ii] = 1.0/sgeex1_libstr(&sA, ii, ii);
148// kernel_strsv_lt_inv_8_lib8(0, sA.pA, sA.cn, sA.dA, sx.pa, sx.pa, sz0.pa);
149// kernel_strsv_lt_inv_8_vs_lib8(0, sA.pA, sA.cn, sA.dA, sx.pa, sx.pa, sz0.pa, 3);
150// s_print_strmat(n, n, &sA, 0, 0);
151// strsv_ltn_libstr(12, &sA, 0, 0, &sx, 0, &sz0, 0);
152// strsv_ltn_mn_libstr(11, 3, &sA, 0, 0, &sx, 0, &sz0, 0);
153// s_print_strmat(n, n, &sA, 0, 0);
154// kernel_sgemv_nt_4_lib8(n, &alpha, &alpha, sA.pA, sA.cn, sx.pa, sx.pa, &beta, sz1.pa, sz0.pa, sz1.pa);
155// kernel_sgemv_nt_4_vs_lib8(n, &alpha, &alpha, sA.pA, sA.cn, sx.pa, sx.pa, &beta, sz1.pa, sz0.pa, sz1.pa, 3);
156// sgemv_nt_libstr(5, 2, alpha, alpha, &sA, 0, 0, &sx, 0, &sx, 0, beta, beta, &sz0, 0, &sz1, 0, &sz0, 0, &sz1, 0);
157// ssymv_l_libstr(10, 10, alpha, &sA, 1, 0, &sx, 0, beta, &sz0, 0, &sz1, 0);
158// s_print_tran_strvec(n, &sz0, 0);
159// s_print_tran_strvec(n, &sz1, 0);
160// return 0;
161// sgesc_libstr(16, 9, 2.0, &sD, 0, 0);
162// s_print_strmat(n, n, &sD, 0, 0);
163// kernel_spotrf_nt_l_8x8_lib8(0, sD.pA, sD.pA, sD.pA, sD.pA, sx.pa);
164// s_print_strmat(n, n, &sD, 0, 0);
165// s_print_tran_strvec(n, &sx, 0);
166// kernel_strsm_nt_rl_inv_8x8_lib8(0, sD.pA, sD.pA, sD.pA+8*sD.cn, sD.pA+8*sD.cn, sD.pA, sx.pa);
167// s_print_strmat(n, n, &sD, 0, 0);
168// kernel_spotrf_nt_l_8x8_lib8(8, sD.pA+8*sD.cn, sD.pA+8*sD.cn, sD.pA+8*sD.cn+8*8, sD.pA+8*sD.cn+8*8, sx.pa+8);
169// spotrf_l_mn_libstr(23, 17, &sD, 0, 0, &sD, 0, 0);
170// spotrf_l_libstr(n, &sD, 0, 0, &sD, 0, 0);
171// kernel_strmm_nn_rl_8x4_lib8(3, &alpha, sB.pA, 7, sA.pA, sA.cn, sD.pA);
172 strmm_rlnn_libstr(12, 8, 1.0, &sA, 0, 0, &sB, 0, 0, &sD, 0, 0);
173 s_print_strmat(n, n, &sD, 0, 0);
174 return 0;
175
176
177
178 //
179 // free memory
180 //
181
182 free(A);
183 free(B);
184 free(D);
185 s_free_strmat(&sA);
186 s_free_strmat(&sB);
187 s_free_strmat(&sD);
188
189 return 0;
190
191 }