1 /***************************************************************************
2 Copyright (c) 2020, The OpenBLAS Project
3 All rights reserved.
4 Redistribution and use in source and binary forms, with or without
5 modification, are permitted provided that the following conditions are
6 met:
7 1. Redistributions of source code must retain the above copyright
8 notice, this list of conditions and the following disclaimer.
9 2. Redistributions in binary form must reproduce the above copyright
10 notice, this list of conditions and the following disclaimer in
11 the documentation and/or other materials provided with the
12 distribution.
13 3. Neither the name of the OpenBLAS project nor the names of
14 its contributors may be used to endorse or promote products
15 derived from this software without specific prior written permission.
16 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
17 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 ARE DISCLAIMED. IN NO EVENT SHALL THE OPENBLAS PROJECT OR CONTRIBUTORS BE
20 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
22 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
23 CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
24 OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
25 USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26 *****************************************************************************/
27
28 #include "common.h"
29 #include <stdio.h>
30 #if !defined(DOUBLE)
31 #define VSETVL(n) vsetvl_e32m8(n)
32 #define VSETVL_MAX vsetvlmax_e32m1()
33 #define FLOAT_V_T vfloat32m8_t
34 #define VLEV_FLOAT vle_v_f32m8
35 #define VLSEV_FLOAT vlse_v_f32m8
36 #define VSEV_FLOAT vse_v_f32m8
37 #define VSSEV_FLOAT vsse_v_f32m8
38 #else
39 #define VSETVL(n) vsetvl_e64m8(n)
40 #define VSETVL_MAX vsetvlmax_e64m1()
41 #define FLOAT_V_T vfloat64m8_t
42 #define VLEV_FLOAT vle_v_f64m8
43 #define VLSEV_FLOAT vlse_v_f64m8
44 #define VSEV_FLOAT vse_v_f64m8
45 #define VSSEV_FLOAT vsse_v_f64m8
46 #endif
47
CNAME(BLASLONG n,BLASLONG dummy0,BLASLONG dummy1,FLOAT dummy3,FLOAT * x,BLASLONG inc_x,FLOAT * y,BLASLONG inc_y,FLOAT * dummy,BLASLONG dummy2)48 int CNAME(BLASLONG n, BLASLONG dummy0, BLASLONG dummy1, FLOAT dummy3, FLOAT *x, BLASLONG inc_x, FLOAT *y, BLASLONG inc_y, FLOAT *dummy, BLASLONG dummy2)
49 {
50 BLASLONG i = 0, j = 0;
51 BLASLONG ix = 0,iy = 0;
52 BLASLONG stride_x, stride_y;
53 FLOAT_V_T vx0, vx1, vy0, vy1;
54 unsigned int gvl = 0;
55
56 if (n < 0) return(0);
57 if(inc_x == 1 && inc_y == 1){
58 gvl = VSETVL(n);
59 if(gvl <= n/2){
60 for(i=0,j=0; i<n/(2*gvl); i++){
61 vx0 = VLEV_FLOAT(&x[j], gvl);
62 vy0 = VLEV_FLOAT(&y[j], gvl);
63 VSEV_FLOAT(&x[j], vy0, gvl);
64 VSEV_FLOAT(&y[j], vx0, gvl);
65
66 vx1 = VLEV_FLOAT(&x[j+gvl], gvl);
67 vy1 = VLEV_FLOAT(&y[j+gvl], gvl);
68 VSEV_FLOAT(&x[j+gvl], vy1, gvl);
69 VSEV_FLOAT(&y[j+gvl], vx1, gvl);
70 j+=gvl * 2;
71 }
72 }
73 for(;j<n;){
74 gvl = VSETVL(n-j);
75 vx0 = VLEV_FLOAT(&x[j], gvl);
76 vy0 = VLEV_FLOAT(&y[j], gvl);
77 VSEV_FLOAT(&x[j], vy0, gvl);
78 VSEV_FLOAT(&y[j], vx0, gvl);
79 j+=gvl;
80 }
81 }else if (inc_y == 1){
82 gvl = VSETVL(n);
83 stride_x = inc_x * sizeof(FLOAT);
84 if(gvl <= n/2){
85 BLASLONG inc_xv = inc_x * gvl;
86 for(i=0,j=0; i<n/(2*gvl); i++){
87 vx0 = VLSEV_FLOAT(&x[ix], stride_x, gvl);
88 vy0 = VLEV_FLOAT(&y[j], gvl);
89 VSSEV_FLOAT(&x[ix], stride_x, vy0, gvl);
90 VSEV_FLOAT(&y[j], vx0, gvl);
91
92 vx1 = VLSEV_FLOAT(&x[ix+inc_xv], stride_x, gvl);
93 vy1 = VLEV_FLOAT(&y[j+gvl], gvl);
94 VSSEV_FLOAT(&x[ix+inc_xv], stride_x, vy1, gvl);
95 VSEV_FLOAT(&y[j+gvl], vx1, gvl);
96 j += gvl * 2;
97 ix += inc_xv * 2;
98 }
99 }
100 for(;j<n;){
101 gvl = VSETVL(n-j);
102 vx0 = VLSEV_FLOAT(&x[ix], stride_x, gvl);
103 vy0 = VLEV_FLOAT(&y[j], gvl);
104 VSSEV_FLOAT(&x[ix], stride_x, vy0, gvl);
105 VSEV_FLOAT(&y[j], vx0, gvl);
106 j += gvl;
107 ix += inc_x * gvl;
108 }
109 }else if(inc_x == 1){
110 gvl = VSETVL(n);
111 stride_y = inc_y * sizeof(FLOAT);
112 if(gvl <= n/2){
113 BLASLONG inc_yv = inc_y * gvl;
114 for(i=0,j=0; i<n/(2*gvl); i++){
115 vx0 = VLEV_FLOAT(&x[j], gvl);
116 vy0 = VLSEV_FLOAT(&y[iy], stride_y, gvl);
117 VSEV_FLOAT(&x[j], vy0, gvl);
118 VSSEV_FLOAT(&y[iy], stride_y, vx0, gvl);
119
120 vx1 = VLEV_FLOAT(&x[j+gvl], gvl);
121 vy1 = VLSEV_FLOAT(&y[iy+inc_yv], stride_y, gvl);
122 VSEV_FLOAT(&x[j+gvl], vy1, gvl);
123 VSSEV_FLOAT(&y[iy+inc_yv], stride_y, vx1, gvl);
124 j += gvl * 2;
125 iy += inc_yv * 2;
126 }
127 }
128 for(;j<n;){
129 gvl = VSETVL(n-j);
130 vx0 = VLEV_FLOAT(&x[j], gvl);
131 vy0 = VLSEV_FLOAT(&y[iy], stride_y, gvl);
132 VSEV_FLOAT(&x[j], vy0, gvl);
133 VSSEV_FLOAT(&y[iy], stride_y, vx0, gvl);
134 j += gvl;
135 iy += inc_y * gvl;
136 }
137 }else{
138 gvl = VSETVL(n);
139 stride_x = inc_x * sizeof(FLOAT);
140 stride_y = inc_y * sizeof(FLOAT);
141 if(gvl <= n/2){
142 BLASLONG inc_xv = inc_x * gvl;
143 BLASLONG inc_yv = inc_y * gvl;
144 for(i=0,j=0; i<n/(2*gvl); i++){
145 vx0 = VLSEV_FLOAT(&x[ix], stride_x, gvl);
146 vy0 = VLSEV_FLOAT(&y[iy], stride_y, gvl);
147 VSSEV_FLOAT(&x[ix], stride_x, vy0, gvl);
148 VSSEV_FLOAT(&y[iy], stride_y, vx0, gvl);
149
150 vx1 = VLSEV_FLOAT(&x[ix+inc_xv], stride_x, gvl);
151 vy1 = VLSEV_FLOAT(&y[iy+inc_yv], stride_y, gvl);
152 VSSEV_FLOAT(&x[ix+inc_xv], stride_x, vy1, gvl);
153 VSSEV_FLOAT(&y[iy+inc_yv], stride_y, vx1, gvl);
154 j += gvl * 2;
155 ix += inc_xv * 2;
156 iy += inc_yv * 2;
157 }
158 }
159 for(;j<n;){
160 gvl = VSETVL(n-j);
161 vx0 = VLSEV_FLOAT(&x[ix], stride_x, gvl);
162 vy0 = VLSEV_FLOAT(&y[iy], stride_y, gvl);
163 VSSEV_FLOAT(&x[ix], stride_x, vy0, gvl);
164 VSSEV_FLOAT(&y[iy], stride_y, vx0, gvl);
165 j += gvl;
166 ix += inc_x * gvl;
167 iy += inc_y * gvl;
168 }
169 }
170 return(0);
171 }
172
173
174