File: malloc.c

package info (click to toggle)
swiftlang 6.1.3-4
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid
  • size: 2,791,644 kB
  • sloc: cpp: 9,901,738; ansic: 2,201,433; asm: 1,091,827; python: 308,252; objc: 82,166; f90: 80,126; lisp: 38,358; pascal: 25,559; sh: 20,429; ml: 5,058; perl: 4,745; makefile: 4,484; awk: 3,535; javascript: 3,018; xml: 918; fortran: 664; cs: 573; ruby: 396
file content (37 lines) | stat: -rw-r--r-- 991 bytes parent folder | download | duplicates (11) 
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
 
// RUN: %libomptarget-compile-generic && %libomptarget-run-generic
// RUN: %libomptarget-compileopt-generic && %libomptarget-run-generic

#include <stdio.h>
#include <stdlib.h>

int main() {
  long unsigned *DP = 0;
  int N = 32;
  int Threads = 64;
  int Teams = 10;

  // Allocate ~55MB on the device.
#pragma omp target map(from : DP)
  DP = (long unsigned *)malloc(sizeof(long unsigned) * N * Threads * Teams);

#pragma omp target teams distribute parallel for num_teams(Teams)              \
    thread_limit(Threads) is_device_ptr(DP)
  for (int i = 0; i < Threads * Teams; ++i) {
    for (int j = 0; j < N; ++j) {
      DP[i * N + j] = i + j;
    }
  }

  long unsigned s = 0;
#pragma omp target teams distribute parallel for num_teams(Teams)              \
    thread_limit(Threads) reduction(+ : s)
  for (int i = 0; i < Threads * Teams; ++i) {
    for (int j = 0; j < N; ++j) {
      s += DP[i * N + j];
    }
  }

  // CHECK: Sum: 6860800
  printf("Sum: %li\n", s);
  return 0;
}