Neko 0.9.99
A portable framework for high-order spectral element flow simulations
Loading...
Searching...
No Matches
opr_conv1.cu
Go to the documentation of this file.
1/*
2 Copyright (c) 2021-2023, The Neko Authors
3 All rights reserved.
4
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions
7 are met:
8
9 * Redistributions of source code must retain the above copyright
10 notice, this list of conditions and the following disclaimer.
11
12 * Redistributions in binary form must reproduce the above
13 copyright notice, this list of conditions and the following
14 disclaimer in the documentation and/or other materials provided
15 with the distribution.
16
17 * Neither the name of the authors nor the names of its
18 contributors may be used to endorse or promote products derived
19 from this software without specific prior written permission.
20
21 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
24 FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
25 COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
26 INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
27 BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
28 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
29 CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
31 ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
32 POSSIBILITY OF SUCH DAMAGE.
33*/
34
35#include <string.h>
36#include <stdlib.h>
37#include <stdio.h>
38#include "conv1_kernel.h"
39#include <device/device_config.h>
40#include <device/cuda/check.h>
41
42extern "C" {
43 #include <common/neko_log.h>
44}
45
46template < const int >
47int tune_conv1(void *du, void *u,
48 void *vx, void *vy, void *vz,
49 void *dx, void *dy, void *dz,
50 void *drdx, void *dsdx, void *dtdx,
51 void *drdy, void *dsdy, void *dtdy,
52 void *drdz, void *dsdz, void *dtdz,
53 void *jacinv, int *nel, int *gdim, int *lx);
54
55extern "C" {
56
60 void cuda_conv1(void *du, void *u,
61 void *vx, void *vy, void *vz,
62 void *dx, void *dy, void *dz,
63 void *drdx, void *dsdx, void *dtdx,
64 void *drdy, void *dsdy, void *dtdy,
65 void *drdz, void *dsdz, void *dtdz,
66 void *jacinv, int *nel, int *gdim, int *lx) {
67
68 static int autotune[17] = { 0 };
69
70 const dim3 nthrds_1d(1024, 1, 1);
71 const dim3 nthrds_kstep((*lx), (*lx), 1);
72 const dim3 nblcks((*nel), 1, 1);
73 const cudaStream_t stream = (cudaStream_t) glb_cmd_queue;
74
75#define CASE_1D(LX) \
76 conv1_kernel_1d<real, LX, 1024> \
77 <<<nblcks, nthrds_1d, 0, stream>>> \
78 ((real *) du, (real *) u, \
79 (real *) vx, (real *) vy, (real *) vz, \
80 (real *) dx, (real *) dy, (real *) dz, \
81 (real *) drdx, (real *) dsdx, (real *) dtdx, \
82 (real *) drdy, (real *) dsdy, (real *) dtdy, \
83 (real *) drdz, (real *) dsdz, (real *) dtdz, \
84 (real *) jacinv); \
85 CUDA_CHECK(cudaGetLastError());
86
87#define CASE_KSTEP(LX) \
88 conv1_kernel_kstep<real, LX> \
89 <<<nblcks, nthrds_kstep, 0, stream>>> \
90 ((real *) du, (real *) u, \
91 (real *) vx, (real *) vy, (real *) vz, \
92 (real *) dx, (real *) dy, (real *) dz, \
93 (real *) drdx, (real *) dsdx, (real *) dtdx, \
94 (real *) drdy, (real *) dsdy, (real *) dtdy, \
95 (real *) drdz, (real *) dsdz, (real *) dtdz, \
96 (real *) jacinv); \
97 CUDA_CHECK(cudaGetLastError());
98
99#define CASE(LX) \
100 case LX: \
101 if(autotune[LX] == 0 ) { \
102 autotune[LX]=tune_conv1<LX>(du, u, \
103 vx, vy, vz, \
104 dx, dy, dz, \
105 drdx, dsdx, dtdx, \
106 drdy, dsdy, dtdy, \
107 drdz, dsdz, dtdz, \
108 jacinv, nel, gdim, lx); \
109 } else if (autotune[LX] == 1 ) { \
110 CASE_1D(LX); \
111 } else if (autotune[LX] == 2 ) { \
112 CASE_KSTEP(LX); \
113 } \
114 break
115
116#define CASE_LARGE(LX) \
117 case LX: \
118 CASE_KSTEP(LX); \
119 break
120
121
122 if ((*lx) < 11) {
123 switch(*lx) {
124 CASE(2);
125 CASE(3);
126 CASE(4);
127 CASE(5);
128 CASE(6);
129 CASE(7);
130 CASE(8);
131 CASE(9);
132 CASE(10);
133 default:
134 {
135 fprintf(stderr, __FILE__ ": size not supported: %d\n", *lx);
136 exit(1);
137 }
138 }
139 }
140 else {
141 switch(*lx) {
142 CASE_LARGE(11);
143 CASE_LARGE(12);
144 CASE_LARGE(13);
145 CASE_LARGE(14);
146 CASE_LARGE(15);
147 CASE_LARGE(16);
148 default:
149 {
150 fprintf(stderr, __FILE__ ": size not supported: %d\n", *lx);
151 exit(1);
152 }
153 }
154 }
155 }
156}
157
158template < const int LX >
159int tune_conv1(void *du, void *u,
160 void *vx, void *vy, void *vz,
161 void *dx, void *dy, void *dz,
162 void *drdx, void *dsdx, void *dtdx,
163 void *drdy, void *dsdy, void *dtdy,
164 void *drdz, void *dsdz, void *dtdz,
165 void *jacinv, int *nel, int *gdim, int *lx) {
166 cudaEvent_t start,stop;
167 float time1,time2;
168 int retval;
169
170 const dim3 nthrds_1d(1024, 1, 1);
171 const dim3 nthrds_kstep((*lx), (*lx), 1);
172 const dim3 nblcks((*nel), 1, 1);
173 const cudaStream_t stream = (cudaStream_t) glb_cmd_queue;
174
175 char *env_value = NULL;
176 char neko_log_buf[80];
177
178 env_value=getenv("NEKO_AUTOTUNE");
179
180 sprintf(neko_log_buf, "Autotune conv1 (lx: %d)", *lx);
181 log_section(neko_log_buf);
182
183 if(env_value) {
184 if( !strcmp(env_value,"1D") ) {
185 CASE_1D(LX);
186 sprintf(neko_log_buf,"Set by env : 1 (1D)");
187 log_message(neko_log_buf);
188 log_end_section();
189 return 1;
190 } else if( !strcmp(env_value,"KSTEP") ) {
191 CASE_KSTEP(LX);
192 sprintf(neko_log_buf,"Set by env : 2 (KSTEP)");
193 log_message(neko_log_buf);
194 log_end_section();
195 return 2;
196 } else {
197 sprintf(neko_log_buf, "Invalid value set for NEKO_AUTOTUNE");
198 log_error(neko_log_buf);
199 }
200 }
201
202 cudaEventCreate(&start);
203 cudaEventCreate(&stop);
204
205 cudaEventRecord(start,0);
206
207 for(int i = 0; i < 100; i++) {
208 CASE_1D(LX);
209 }
210
211 cudaEventRecord(stop,0);
212 cudaEventSynchronize(stop);
213 cudaEventElapsedTime(&time1, start, stop);
214
215 cudaEventRecord(start,0);
216
217 for(int i = 0; i < 100; i++) {
218 CASE_KSTEP(LX);
219 }
220
221 cudaEventRecord(stop,0);
222 cudaEventSynchronize(stop);
223 cudaEventElapsedTime(&time2, start, stop);
224
225 if(time1 < time2) {
226 retval = 1;
227 } else {
228 retval = 2;
229 }
230
231 sprintf(neko_log_buf, "Chose : %d (%s)", retval,
232 (retval > 1 ? "KSTEP" : "1D"));
233 log_message(neko_log_buf);
234 log_end_section();
235 return retval;
236}
drdy
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ drdy
Definition ax_helm_full_kernel.h:54
drdz
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ drdz
Definition ax_helm_full_kernel.h:55
dsdz
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dsdz
Definition ax_helm_full_kernel.h:58
dsdy
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dsdy
Definition ax_helm_full_kernel.h:57
i
const int i
Definition ax_helm_full_kernel.h:95
dtdy
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dtdy
Definition ax_helm_full_kernel.h:60
u
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ u
Definition ax_helm_full_kernel.h:46
dx
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dx
Definition ax_helm_full_kernel.h:49
drdx
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ drdx
Definition ax_helm_full_kernel.h:53
dtdz
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dtdz
Definition ax_helm_full_kernel.h:61
dsdx
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dsdx
Definition ax_helm_full_kernel.h:56
dtdx
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dtdx
Definition ax_helm_full_kernel.h:59
dz
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dz
Definition ax_helm_full_kernel.h:51
dy
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ dy
Definition ax_helm_full_kernel.h:50
jacinv
__global__ void T *__restrict__ T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ jacinv
Definition ax_helm_full_kernel.h:62
vz
__global__ void const T *__restrict__ const T *__restrict__ const T *__restrict__ const T *__restrict__ vz
Definition conv1_kernel.h:135
vx
__global__ void const T *__restrict__ const T *__restrict__ vx
Definition conv1_kernel.h:133
vy
__global__ void const T *__restrict__ const T *__restrict__ const T *__restrict__ vy
Definition conv1_kernel.h:134
dirichlet_apply_scalar_kernel
__global__ void dirichlet_apply_scalar_kernel(const int *__restrict__ msk, T *__restrict__ x, const T g, const int m)
Definition dirichlet_kernel.h:42
device_config.h
conv1_kernel.h
neko_log.h
log_error
void log_error(char *msg)
log_message
void log_message(char *msg)
log_end_section
void log_end_section()
log_section
void log_section(char *msg)
CASE
#define CASE(LX)
cuda_conv1
void cuda_conv1(void *du, void *u, void *vx, void *vy, void *vz, void *dx, void *dy, void *dz, void *drdx, void *dsdx, void *dtdx, void *drdy, void *dsdy, void *dtdy, void *drdz, void *dsdz, void *dtdz, void *jacinv, int *nel, int *gdim, int *lx)
Definition opr_conv1.cu:60
CASE_KSTEP
#define CASE_KSTEP(LX)
CASE_LARGE
#define CASE_LARGE(LX)
CASE_1D
#define CASE_1D(LX)
tune_conv1
int tune_conv1(void *du, void *u, void *vx, void *vy, void *vz, void *dx, void *dy, void *dz, void *drdx, void *dsdx, void *dtdx, void *drdy, void *dsdy, void *dtdy, void *drdz, void *dsdz, void *dtdz, void *jacinv, int *nel, int *gdim, int *lx)
Definition opr_conv1.cu:159