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
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
|
/**
* app.c
* CPU-DPU Communication Host Application Source File
*
*/
#include <stdio.h>
#include <stdlib.h>
#include <stdbool.h>
#include <string.h>
#include <dpu.h>
#include <dpu_log.h>
#include <unistd.h>
#include <getopt.h>
#include <assert.h>
#include "../support/common.h"
#include "../support/timer.h"
#include "../support/params.h"
// Define the DPU Binary path as DPU_BINARY here
#ifndef DPU_BINARY
#define DPU_BINARY "./bin/dpu_code"
#endif
#define XSTR(x) STR(x)
#define STR(x) #x
// Pointer declaration
static T* A;
static T* B;
static T* C;
static T* C2;
static const char transfer_mode[] =
#if SERIAL
"SERIAL"
#elif BROADCAST
"BROADCAST"
#else
"PUSH"
#endif
;
// Create input arrays
static void read_input(T* A, T* B, unsigned int nr_elements) {
srand(0);
//printf("nr_elements\t%u\t", nr_elements);
for (unsigned int i = 0; i < nr_elements; i++) {
A[i] = (T) (rand());
B[i] = A[i];
}
}
// Main of the Host Application
int main(int argc, char **argv) {
struct Params p = input_params(argc, argv);
struct dpu_set_t dpu_set, dpu;
uint32_t nr_of_dpus;
uint32_t nr_of_ranks;
char ntpp[24];
// Timer declaration
Timer timer;
snprintf(ntpp, 24, "nrThreadPerPool=%d", p.n_threads);
// Allocate DPUs and load binary
start(&timer, 4, 0);
#if NR_DPUS
DPU_ASSERT(dpu_alloc(NR_DPUS, ntpp, &dpu_set));
#elif NR_RANKS
DPU_ASSERT(dpu_alloc_ranks(NR_RANKS, ntpp, &dpu_set));
#else
#error "NR_DPUS o NR_RANKS must be set"
#endif
stop(&timer, 4);
start(&timer, 5, 0);
DPU_ASSERT(dpu_load(dpu_set, DPU_BINARY, NULL));
stop(&timer, 5);
start(&timer, 6, 0);
DPU_ASSERT(dpu_get_nr_dpus(dpu_set, &nr_of_dpus));
DPU_ASSERT(dpu_get_nr_ranks(dpu_set, &nr_of_ranks));
stop(&timer, 6);
//printf("Allocated %d DPU(s)\n", nr_of_dpus);
unsigned int i = 0;
uint64_t input_size = p.exp == 0 ? p.input_size * nr_of_dpus : p.input_size;
//printf("Load input data\n");
// Input arguments
const uint64_t input_size_dpu = input_size / nr_of_dpus;
#ifdef BROADCAST
const uint64_t transfer_size = input_size;
#else
const uint64_t transfer_size = input_size;
#endif
// Input/output allocation
A = malloc(input_size * sizeof(T));
B = malloc(input_size * sizeof(T));
C = malloc(input_size * sizeof(T));
C2 = malloc(input_size * sizeof(T));
T *bufferA = A;
T *bufferC = C;
// Create an input file with arbitrary data
read_input(A, B, input_size);
//printf("NR_TASKLETS\t%d\tBL\t%d\n", NR_TASKLETS, BL);
printf("[::] NMC reconfiguration | n_dpus=%d n_ranks=%d n_tasklets=%d n_nops=%d n_instr=%d e_type=%s n_elements=%lu e_mode=%s"
" | latency_dpu_alloc_ns=%lu latency_dpu_load_ns=%lu latency_dpu_get_ns=%lu\n",
nr_of_dpus, nr_of_ranks, NR_TASKLETS, p.n_nops, p.n_instr, XSTR(T), transfer_size, transfer_mode,
timer.nanoseconds[4], timer.nanoseconds[5], timer.nanoseconds[6]);
// Loop over main kernel
for(int rep = 0; rep < p.n_warmup + p.n_reps; rep++) {
// Copy input arrays
if(rep >= p.n_warmup)
start(&timer, 1, 0);
i = 0;
#ifdef SERIAL
DPU_FOREACH (dpu_set, dpu) {
DPU_ASSERT(dpu_copy_to(dpu, DPU_MRAM_HEAP_POINTER_NAME, 0, bufferA + input_size_dpu * i, input_size_dpu * sizeof(T)));
i++;
}
#elif BROADCAST
DPU_ASSERT(dpu_broadcast_to(dpu_set, DPU_MRAM_HEAP_POINTER_NAME, 0, bufferA, input_size * sizeof(T), DPU_XFER_DEFAULT));
#else
DPU_FOREACH(dpu_set, dpu, i) {
DPU_ASSERT(dpu_prepare_xfer(dpu, bufferA + input_size_dpu * i));
}
DPU_ASSERT(dpu_push_xfer(dpu_set, DPU_XFER_TO_DPU, DPU_MRAM_HEAP_POINTER_NAME, 0, input_size_dpu * sizeof(T), DPU_XFER_DEFAULT));
#endif
if(rep >= p.n_warmup)
stop(&timer, 1);
//printf("Run program on DPU(s) \n");
// Run DPU kernel
if(rep >= p.n_warmup)
start(&timer, 2, 0);
// empty kernel -> measure communication overhead
DPU_ASSERT(dpu_launch(dpu_set, DPU_SYNCHRONOUS));
if(rep >= p.n_warmup)
stop(&timer, 2);
#if PRINT
{
unsigned int each_dpu = 0;
printf("Display DPU Logs\n");
DPU_FOREACH (dpu_set, dpu) {
printf("DPU#%d:\n", each_dpu);
DPU_ASSERT(dpulog_read_for_dpu(dpu.dpu, stdout));
each_dpu++;
}
}
#endif
//printf("Retrieve results\n");
if(rep >= p.n_warmup)
start(&timer, 3, 0);
i = 0;
#ifdef SERIAL
DPU_FOREACH (dpu_set, dpu) {
DPU_ASSERT(dpu_copy_from(dpu, DPU_MRAM_HEAP_POINTER_NAME, 0, bufferC + input_size_dpu * i, input_size_dpu * sizeof(T)));
i++;
}
#else
DPU_FOREACH(dpu_set, dpu, i) {
DPU_ASSERT(dpu_prepare_xfer(dpu, bufferC + input_size_dpu * i));
}
DPU_ASSERT(dpu_push_xfer(dpu_set, DPU_XFER_FROM_DPU, DPU_MRAM_HEAP_POINTER_NAME, 0, input_size_dpu * sizeof(T), DPU_XFER_DEFAULT));
#endif
if(rep >= p.n_warmup)
stop(&timer, 3);
if (rep >= p.n_warmup) {
printf("[::] transfer UPMEM | n_dpus=%d n_ranks=%d n_tasklets=%d n_nops=%d n_instr=%d e_type=%s n_elements=%lu n_elements_per_dpu=%lu e_mode=%s"
" | latency_dram_mram_ns=%lu latency_mram_dram_ns=%lu throughput_dram_mram_Bps=%f throughput_mram_dram_Bps=%f",
#ifdef BROADCAST
nr_of_dpus, nr_of_ranks, NR_TASKLETS, p.n_nops, p.n_instr, XSTR(T), transfer_size, transfer_size, transfer_mode,
#else
nr_of_dpus, nr_of_ranks, NR_TASKLETS, p.n_nops, p.n_instr, XSTR(T), transfer_size, transfer_size / nr_of_dpus, transfer_mode,
#endif
timer.nanoseconds[1], timer.nanoseconds[3],
transfer_size * sizeof(T) * 1e9 / timer.nanoseconds[1],
transfer_size * sizeof(T) * 1e9 / timer.nanoseconds[3]);
printf(" throughput_dram_mram_Opps=%f throughput_mram_dram_Opps=%f",
transfer_size * 1e9 / timer.nanoseconds[1],
transfer_size * 1e9 / timer.nanoseconds[3]);
printf(" latency_dpu_launch_ns=%lu\n",
timer.nanoseconds[2]);
}
}
// Check output
bool status = true;
#ifdef BROADCAST
for (i = 0; i < input_size/nr_of_dpus; i++) {
if(B[i] != bufferC[i]){
status = false;
#if PRINT
printf("%d: %u -- %u\n", i, B[i], bufferA[i]);
#endif
}
}
#else
for (i = 0; i < input_size; i++) {
if(B[i] != bufferC[i]){
status = false;
#if PRINT
printf("%d: %u -- %u\n", i, B[i], bufferA[i]);
#endif
}
}
#endif
if (status) {
//printf("[" ANSI_COLOR_GREEN "OK" ANSI_COLOR_RESET "] Outputs are equal\n");
} else {
printf("[" ANSI_COLOR_RED "ERROR" ANSI_COLOR_RESET "] Outputs differ!\n");
}
// Deallocation
free(A);
free(B);
free(C);
free(C2);
DPU_ASSERT(dpu_free(dpu_set));
return 0;
}
|
