int main(int argc, char *argv[])
{
#if _POSIX_CPUTIME == -1
printf("_POSIX_CPUTIME unsupported\n");
return PTS_UNSUPPORTED;
#else
#ifdef CLOCK_PROCESS_CPUTIME_ID
struct timespec ts1, ts2, ts3, ts4;
if (sysconf(_SC_CPUTIME) == -1) {
printf("_POSIX_CPUTIME unsupported\n");
return PTS_UNSUPPORTED;
}
if (clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &ts1) != 0) {
printf("clock_gettime() failed: errno %d\n", errno);
return PTS_UNRESOLVED;
}
dosomething();
if (clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &ts2) != 0) {
printf("clock_gettime() failed: errno %d\n", errno);
return PTS_UNRESOLVED;
}
dosomething();
if (clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &ts3) != 0) {
printf("clock_gettime() failed: errno %d\n", errno);
return PTS_UNRESOLVED;
}
dosomething();
if (clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &ts4) != 0) {
printf("clock_gettime() failed: errno %d\n", errno);
return PTS_UNRESOLVED;
}
if ((ts1.tv_sec <= ts2.tv_sec) &&
(ts2.tv_sec <= ts3.tv_sec) && (ts3.tv_sec <= ts4.tv_sec)) {
printf("Test PASSED\n");
return PTS_PASS;
} else {
printf("Test FAILED - ts1=%ld,ts2=%ld,ts3=%ld,ts4=%ld\n",
ts1.tv_sec, ts2.tv_sec, ts3.tv_sec, ts4.tv_sec);
return PTS_FAIL;
}
printf("This code should not be executed.\n");
return PTS_UNRESOLVED;
#else
printf("CLOCK_PROCESS_CPUTIME_ID unsupported\n");
return PTS_UNSUPPORTED;
#endif
#endif
}
int main(int argc, char *argv[])
{
#ifndef _POSIX_CPUTIME
printf("_POSIX_CPUTIME unsupported\n");
return PTS_UNSUPPORTED;
#else
clockid_t clockid;
struct timespec tp1 = {.tv_sec = 0, .tv_nsec = 0};
dosomething();
if (clock_getcpuclockid(getpid(), &clockid) != 0) {
printf("clock_getcpuclockid() failed\n");
return PTS_FAIL;
}
/* Verify that it returned a valid clockid_t that can be used in other functions */
if (clock_gettime(clockid, &tp1) != 0) {
printf("clock_getcpuclockid() returned an invalid clockid_t: %d\n", clockid);
return PTS_FAIL;
}
printf("Test PASSED\n");
return PTS_PASS;
#endif
}
int main() {
int i = 0;
int a = 1234;
//for (i = 0; i < 100000000L; i++) {
for (i = 0; i < 10; i++) {
printf("PRE : a %d b %d c %d, flag %s\n", a,b,c, flag & 0x01 ? "SET":"NOT SET");
a = dosomething(PASS_A(a));
printf("POST: a %d b %d c %d, flag %s\n", a,b,c, flag & 0x01 ? "SET":"NOT SET");
// a = ntohs(a);
}
exit(0);
}
void process_or_do_work(int fd1, int fd2) {
char buf[1024];
ssize_t bytesread;
for ( ; ; ) {
bytesread = r_read(fd1, buf, sizeof(buf));
if ((bytesread == -1) && (errno != EAGAIN))
return; /* a real error on fd1 */
else if (bytesread > 0) {
docommand(buf, bytesread);
continue;
}
bytesread = r_read(fd2, buf, sizeof(buf));
if ((bytesread == -1) && (errno != EAGAIN))
return; /* a real error on fd2 */
else if (bytesread > 0)
docommand(buf, bytesread);
else
dosomething(); /* input not available, do something else */
}
}
// returns 1 if a key is pressed.
// the key value/index is stored in the global variable NewKey.
char kbhit(){
unsigned char KeyDATA[] =
{'1','2','3','A','4','5','6','B','7','8','9','C','*','0','#','D'};
for(ColCnt = 0, colnum = 0x04; colnum < 0x40; ColCnt = ColCnt +4){
PORTD = colnum; // set column value
dosomething(); // delay
rownum = PORTE; // get row value
rownum = (rownum & 0x0f); // mask all but rows
if(rownum){ // if a row in this column is high
if(rownum == 0x04) rownum = 3;
if(rownum == 0x08) rownum = 4; /* figure row number */
rownum = rownum - 1;
keybuf = ColCnt + rownum; /* calculate key index */
if(keybuf == LastKey) return 0; /* ignore if same key pressed */
LastKey = keybuf; /* save this key for repeat */
NewKey = KeyDATA[keybuf];
return 1; /* return YES */
}
colnum = colnum << 1; // shift column left
}
LastKey = 0xff; /* initialize key repeat compare value*/
return 0; /* no key pressed, return 0 */
}
int main(int argc, char *argv[])
{
int i;
double *rbuf;
FILE *fp = stdout;
char fn[BUFSIZ];
/* for an output in the json format */
char strbuf[BUFSIZ];
char *ptr;
int pos, remain;
int len;
double timeout_sec = 600.0, st;
int opt;
/* If seq_rank >=0, only seq_rank does a real work
* other ranks do nothing; waiting until the next sync point
*/
int seq_rank = -1;
int cnt = 0;
fn[0] = 0;
MPI_Init(NULL, NULL);
while ((opt = getopt(argc, argv, "ht:s:o:")) != -1) {
switch (opt) {
case 'h':
usage(argv[0]);
exit(0);
case 't':
timeout_sec = atoi(optarg);
break;
case 's':
seq_rank = atoi(optarg);
break;
case 'o':
snprintf(fn, BUFSIZ, "%s", optarg);
break;
}
}
MPI_Comm_rank(MPI_COMM_WORLD, &rank);
MPI_Comm_size(MPI_COMM_WORLD, &size);
if (rank == 0) {
printf("# timeout_sec=%lf\n", timeout_sec);
printf("# mpisize=%d\n", size);
if (strlen(fn) > 0) {
fp = fopen(fn, "w");
if (!fp) {
fprintf(stderr, "Unable to open %s", fn);
exit(1);
}
}
}
set_strict_affinity(size, rank);
if (rank == 0)
printf("# set affinity\n");
if (rank == 0) {
rbuf = (double *)malloc(size*sizeof(double));
assert(rbuf);
}
MPI_Barrier(MPI_COMM_WORLD);
initprd();
st = MPI_Wtime();
while (1) {
MPI_Barrier(MPI_COMM_WORLD);
if ((MPI_Wtime()-st) >= timeout_sec)
break;
dosomething(rank, seq_rank);
MPI_Gather(&prd.gflops, 1, MPI_DOUBLE,
rbuf, 1, MPI_DOUBLE, 0, MPI_COMM_WORLD);
if (rank == 0) {
int j;
double agg;
ptr = strbuf;
remain = BUFSIZ - 1;
snprintf(ptr, remain, "{\"sample\":\"dgemm\",\"time\":%.2lf",
MPI_Wtime());
len = strlen(ptr);
ptr += len;
remain -= len;
snprintf(ptr, remain, ", \"mpisize\":%d ", size);
len = strlen(ptr);
ptr += len;
remain -= len;
/* */
agg = 0.0;
for (j = 0; j < size; j++) {
agg += rbuf[j];
snprintf(ptr, remain, ", \"dgemm%d\":%.3lf ",
j, rbuf[j]);
len = strlen(ptr);
ptr += len;
remain -= len;
}
snprintf(ptr, remain, ", \"dgemm_agg\":%.3lf ", agg);
len = strlen(ptr);
ptr += len;
remain -= len;
snprintf(ptr, remain, "}");
fprintf(fp, "%s\n", strbuf);
fflush(fp);
if (0) {
FILE *fp;
fp = fopen("dgemmlast.log", "w");
if (fp) {
fprintf(fp, "%s\n", strbuf);
fflush(fp);
fclose(fp);
}
}
}
}
if (rank == 0) {
if (fp != stdout)
fclose(fp);
}
MPI_Barrier(MPI_COMM_WORLD);
MPI_Finalize();
return 0;
}
void
test (int n, int o, int p, int q, int r, int s, int *pp)
{
int a[o], i, j;
#pragma omp target data device (n + 1) if (n != 6) map (tofrom: n, r)
{
#pragma omp target device (n + 1) if (n != 6) map (from: n) map (alloc: a[2:o-2])
dosomething (a, n, 0);
#pragma omp target teams device (n + 1) num_teams (n + 4) thread_limit (n * 2) \
if (n != 6)map (from: n) map (alloc: a[2:o-2]) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp target teams distribute device (n + 1) num_teams (n + 4) collapse (2) \
if (n != 6)map (from: n) map (alloc: a[2:o-2]) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp target teams distribute device (n + 1) num_teams (n + 4) \
if (n != 6)map (from: n) map (alloc: a[2:o-2]) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp target teams distribute parallel for device (n + 1) num_teams (n + 4) \
if (n != 6)map (from: n) map (alloc: a[2:o-2]) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) collapse (2) \
num_threads (n + 4) proc_bind (spread) lastprivate (s) \
ordered schedule (static, 8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
#pragma omp ordered
p = q;
s = i * 10 + j;
}
#pragma omp target teams distribute parallel for device (n + 1) num_teams (n + 4) \
if (n != 6)map (from: n) map (alloc: a[2:o-2]) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) num_threads (n + 4) \
proc_bind (master) lastprivate (s) ordered schedule (static, 8)
for (i = 0; i < 10; i++)
{
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp ordered
p = q;
s = i * 10;
}
#pragma omp target teams distribute parallel for simd device (n + 1) \
if (n != 6)map (from: n) map (alloc: a[2:o-2]) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) collapse (2) \
num_threads (n + 4) proc_bind (spread) lastprivate (s) \
schedule (static, 8) num_teams (n + 4) safelen(8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
a[2+i*10+j] = p + q;
s = i * 10 + j;
}
#pragma omp target teams distribute parallel for simd device (n + 1) \
if (n != 6)map (from: n) map (alloc: a[2:o-2]) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) num_threads (n + 4) \
proc_bind (master) lastprivate (s) schedule (static, 8) \
num_teams (n + 4) safelen(16) linear(i:1) aligned (pp:4)
for (i = 0; i < 10; i++)
{
r = r + 1;
p = q;
a[2+i] = p + q;
s = i * 10;
}
#pragma omp target teams distribute simd device (n + 1) \
if (n != 6)map (from: n) map (alloc: a[2:o-2]) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) collapse (2) \
lastprivate (s) num_teams (n + 4) safelen(8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
a[2+i*10+j] = p + q;
s = i * 10 + j;
}
#pragma omp target teams distribute simd device (n + 1) \
if (n != 6)map (from: n) map (alloc: a[2:o-2]) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) lastprivate (s) \
num_teams (n + 4) safelen(16) linear(i:1) aligned (pp:4)
for (i = 0; i < 10; i++)
{
r = r + 1;
p = q;
a[2+i] = p + q;
s = i * 10;
}
#pragma omp target device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2])
#pragma omp teams num_teams (n + 4) thread_limit (n * 2) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp target device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2])
#pragma omp teams distribute num_teams (n + 4) collapse (2) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp target device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2])
#pragma omp teams distribute num_teams (n + 4) default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp target device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2])
#pragma omp teams distribute parallel for num_teams (n + 4) if (n != 6) \
default(shared) private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) collapse (2) \
num_threads (n + 4) proc_bind (spread) lastprivate (s) \
ordered schedule (static, 8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
#pragma omp ordered
p = q;
s = i * 10 + j;
}
#pragma omp target device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2])
#pragma omp teams distribute parallel for num_teams (n + 4) if (n != 6) \
default(shared) private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) num_threads (n + 4) \
proc_bind (master) lastprivate (s) ordered schedule (static, 8)
for (i = 0; i < 10; i++)
{
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp ordered
p = q;
s = i * 10;
}
#pragma omp target device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2])
#pragma omp teams distribute parallel for simd if (n != 6)default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) collapse (2) \
num_threads (n + 4) proc_bind (spread) lastprivate (s) \
schedule (static, 8) num_teams (n + 4) safelen(8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
a[2+i*10+j] = p + q;
s = i * 10 + j;
}
#pragma omp target device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2])
#pragma omp teams distribute parallel for simd if (n != 6)default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) num_threads (n + 4) \
proc_bind (master) lastprivate (s) schedule (static, 8) \
num_teams (n + 4) safelen(16) linear(i:1) aligned (pp:4)
for (i = 0; i < 10; i++)
{
r = r + 1;
p = q;
a[2+i] = p + q;
s = i * 10;
}
#pragma omp target device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2])
#pragma omp teams distribute simd default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) collapse (2) \
lastprivate (s) num_teams (n + 4) safelen(8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
a[2+i*10+j] = p + q;
s = i * 10 + j;
}
#pragma omp target device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2])
#pragma omp teams distribute simd default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
thread_limit (n * 2) dist_schedule (static, 4) lastprivate (s) \
num_teams (n + 4) safelen(16) linear(i:1) aligned (pp:4)
for (i = 0; i < 10; i++)
{
r = r + 1;
p = q;
a[2+i] = p + q;
s = i * 10;
}
#pragma omp target teams device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2]) \
num_teams (n + 4) thread_limit (n * 2)default(shared) shared(n) \
private (p) reduction (+: r)
#pragma omp distribute collapse (2) dist_schedule (static, 4) firstprivate (q)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp target teams device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2]) \
num_teams (n + 4) thread_limit (n * 2) shared(n) private(p) reduction (+ : r) \
default(shared)
#pragma omp distribute dist_schedule (static, 4) firstprivate (q)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp target teams device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2]) \
num_teams (n + 4) thread_limit (n * 2)
#pragma omp distribute parallel for if (n != 6) \
default(shared) private (p) firstprivate (q) shared (n) reduction (+: r) \
collapse (2) dist_schedule (static, 4) \
num_threads (n + 4) proc_bind (spread) lastprivate (s) \
ordered schedule (static, 8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
#pragma omp ordered
p = q;
s = i * 10 + j;
}
#pragma omp target teams device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2]) \
num_teams (n + 4) thread_limit (n * 2)
#pragma omp distribute parallel for if (n != 6) \
default(shared) private (p) firstprivate (q) shared (n) reduction (+: r) \
num_threads (n + 4) dist_schedule (static, 4) \
proc_bind (master) lastprivate (s) ordered schedule (static, 8)
for (i = 0; i < 10; i++)
{
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp ordered
p = q;
s = i * 10;
}
#pragma omp target teams device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2]) \
num_teams (n + 4) thread_limit (n * 2)
#pragma omp distribute parallel for simd if (n != 6)default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
collapse (2) dist_schedule (static, 4) \
num_threads (n + 4) proc_bind (spread) lastprivate (s) \
schedule (static, 8) safelen(8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
a[2+i*10+j] = p + q;
s = i * 10 + j;
}
#pragma omp target teams device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2]) \
num_teams (n + 4) thread_limit (n * 2)
#pragma omp distribute parallel for simd if (n != 6)default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
num_threads (n + 4) dist_schedule (static, 4) \
proc_bind (master) lastprivate (s) schedule (static, 8) \
safelen(16) linear(i:1) aligned (pp:4)
for (i = 0; i < 10; i++)
{
r = r + 1;
p = q;
a[2+i] = p + q;
s = i * 10;
}
#pragma omp target teams device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2]) \
num_teams (n + 4) thread_limit (n * 2) default(shared) shared(n) private(p) \
reduction(+:r)
#pragma omp distribute simd private (p) firstprivate (q) reduction (+: r) \
collapse (2) dist_schedule (static, 4) lastprivate (s) safelen(8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
a[2+i*10+j] = p + q;
s = i * 10 + j;
}
#pragma omp target teams device (n + 1) if (n != 6)map(from:n) map(alloc:a[2:o-2]) \
num_teams (n + 4) thread_limit (n * 2) default(shared) shared(n) private(p) \
reduction(+:r)
#pragma omp distribute simd private (p) firstprivate (q) reduction (+: r) \
lastprivate (s) dist_schedule (static, 4) safelen(16) linear(i:1) aligned (pp:4)
for (i = 0; i < 10; i++)
{
r = r + 1;
p = q;
a[2+i] = p + q;
s = i * 10;
}
}
}
void
test2 (int n, int o, int p, int r, int s, int *pp)
{
int a[o];
#pragma omp distribute collapse (2) dist_schedule (static, 4) firstprivate (q)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp distribute dist_schedule (static, 4) firstprivate (q)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp distribute parallel for if (n != 6) \
default(shared) private (p) firstprivate (q) shared (n) reduction (+: r) \
collapse (2) dist_schedule (static, 4) \
num_threads (n + 4) proc_bind (spread) lastprivate (s) \
ordered schedule (static, 8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
#pragma omp ordered
p = q;
s = i * 10 + j;
}
#pragma omp distribute parallel for if (n != 6) \
default(shared) private (p) firstprivate (q) shared (n) reduction (+: r) \
num_threads (n + 4) dist_schedule (static, 4) \
proc_bind (master) lastprivate (s) ordered schedule (static, 8)
for (i = 0; i < 10; i++)
{
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
dosomething (a, n, p + q);
}
#pragma omp ordered
p = q;
s = i * 10;
}
#pragma omp distribute parallel for simd if (n != 6)default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
collapse (2) dist_schedule (static, 4) \
num_threads (n + 4) proc_bind (spread) lastprivate (s) \
schedule (static, 8) safelen(8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
a[2+i*10+j] = p + q;
s = i * 10 + j;
}
#pragma omp distribute parallel for simd if (n != 6)default(shared) \
private (p) firstprivate (q) shared (n) reduction (+: r) \
num_threads (n + 4) dist_schedule (static, 4) \
proc_bind (master) lastprivate (s) schedule (static, 8) \
safelen(16) linear(i:1) aligned (pp:4)
for (i = 0; i < 10; i++)
{
r = r + 1;
p = q;
a[2+i] = p + q;
s = i * 10;
}
#pragma omp distribute simd private (p) firstprivate (q) reduction (+: r) \
collapse (2) dist_schedule (static, 4) lastprivate (s) safelen(8)
for (i = 0; i < 10; i++)
for (j = 0; j < 10; j++)
{
r = r + 1;
p = q;
a[2+i*10+j] = p + q;
s = i * 10 + j;
}
#pragma omp distribute simd private (p) firstprivate (q) reduction (+: r) \
lastprivate (s) dist_schedule (static, 4) safelen(16) linear(i:1) aligned (pp:4)
for (i = 0; i < 10; i++)
{
r = r + 1;
p = q;
a[2+i] = p + q;
s = i * 10;
}
}
void operator() (){
for(int j = 0; j < 100; j++) {
dosomething(i_);
}
}
// waits for a key press and returns it
char getkey(){
while( !kbhit()){ // wait for key press
dosomething();
}
return(NewKey);
}