blob: 341bb5971599d5dbf579d2442b51ce7053e51179 (
plain)
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
|
#include <cuda.h>
#include <stdint.h>
#include "config.h"
#include "ipedma.h"
__global__ void null(uint32_t *data) {
}
__global__ void ipedma(volatile void *bar, uintptr_t bus_addr, volatile uint64_t *desc, uint32_t *data) {
int i;
clock_t sum = 0, sumk = 0, t1, t2, t3;
for (i = 0; i < GPU_ITERS; i++) {
long wait = 0;
desc[1] = 0;
t1 = clock64();
WR64 (REG_DESCRIPTOR_ADDRESS, bus_addr);
do {
if (++wait > 0x10000) break;
} while (desc[1] == 0);
t2 = clock64();
null<<<1,1>>>(data);
cudaDeviceSynchronize();
t3 = clock64();
sum += t2 - t1;
sumk += t3 - t1;
}
data[0] = sum / GPU_ITERS;
data[1] = sumk / GPU_ITERS;
}
/*
__global__ void do_leet (int *rin) {
*rin = 0x1337;
}
*/
|
