--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/src/parallel_execution/examples/twice/createArray.cbc	Wed Sep 06 22:01:27 2017 +0900
@@ -0,0 +1,23 @@
+#include <stdio.h>
+
+#include "../../../context.h"
+
+extern int length;
+extern int split;
+extern int* array_ptr;
+
+__code createArray(__code next(struct Array* output, ...)) {
+    struct Array* output = *O_output;
+    output->prefix = length/split;
+    output->array  = array_ptr;
+    output->size   = length;
+    *O_output = output;
+    goto meta(context, context->next);
+}
+
+__code createArray_stub(struct Context* context) {
+    Array** O_output = (struct Array **)&context->data[context->odg];
+    goto createArray(context,
+                     context->next,
+                     O_output);
+}

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/src/parallel_execution/examples/twice/main.cbc	Wed Sep 06 22:01:27 2017 +0900
@@ -0,0 +1,134 @@
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+#include <unistd.h>
+
+#include "../../../context.h"
+
+int cpu_num = 1;
+int length = 102400;
+int split = 8;
+int* array_ptr;
+int gpu_num = 0;
+int CPU_ANY = -1;
+int CPU_CUDA = -1;
+
+void *start_taskManager(struct Context *context) {
+    goto initDataGears(context, Gearef(context, LoopCounter), Gearef(context, TaskManager));
+    return 0;
+}
+
+#ifdef USE_CUDAWorker
+#ifdef USE_CUDA_MAIN_THREAD
+extern volatile int cuda_initialized;
+#endif
+#endif
+
+__code initDataGears(struct LoopCounter* loopCounter, struct TaskManager* taskManager) {
+    // loopCounter->tree = createRedBlackTree(context);
+    loopCounter->i = 0;
+    taskManager->taskManager = (union Data*)createTaskManagerImpl(context, cpu_num, gpu_num, 0);
+#ifdef USE_CUDAWorker
+#ifdef USE_CUDA_MAIN_THREAD
+    while(! cuda_initialized) {};
+#endif
+#endif
+    goto meta(context, C_code1);
+}
+
+__code initDataGears_stub(struct Context* context) {
+    struct TaskManager* taskManager =  Gearef(context, TaskManager);
+    taskManager->taskManager = 0;
+#if (! defined(USE_CUDAWorker) || ! defined(USE_CUDA_MAIN_THREAD))
+    struct LoopCounter* loopCounter = Gearef(context, LoopCounter);
+    goto initDataGears(context, loopCounter, taskManager);
+#else
+    cuda_initialized = 0;
+    pthread_t thread;
+    pthread_create(&thread, NULL, (void*)&start_taskManager, context);
+    while (taskManager->taskManager == 0);
+    TaskManager *t = (TaskManager*)taskManager->taskManager;
+    TaskManagerImpl *im = (TaskManagerImpl*)t->taskManager;
+    struct Queue *q = (Queue *)im->workers[0];
+    createCUDAWorker(context,0,q, im);
+    pthread_join(thread,0);
+    exit(0);
+#endif
+}
+
+__code code1(struct Time* time) {
+    printf("cpus:\t\t%d\n", cpu_num);
+    printf("gpus:\t\t%d\n", gpu_num);
+    printf("length:\t\t%d\n", length);
+    printf("length/task:\t%d\n", length/split);
+    /* puts("queue"); */
+    /* print_queue(context->data[ActiveQueue]->queue.first); */
+    /* puts("tree"); */
+    /* print_tree(context->data[Tree]->tree.root); */
+    /* puts("result"); */
+    time->time = (union Data*)createTimeImpl(context);
+    time->next = C_createTask1;
+    goto meta(context, time->time->Time.start);
+}
+
+__code code2(struct Time* time, struct TaskManager* taskManager) {
+    time->next = C_code3;
+    taskManager->next = time->time->Time.end;
+    goto meta(context, taskManager->taskManager->TaskManager.shutdown);
+}
+
+__code code3(struct LoopCounter* loopCounter) {
+    int i = loopCounter->i;
+
+    if (i < length) {
+        //printf("%d\n", array_ptr[i]);
+        if (array_ptr[i] == (i*2)) {
+            loopCounter->i++;
+            goto meta(context, C_code3);
+        } else
+            puts("wrong result");
+
+    }
+
+    goto meta(context, C_exit_code);
+}
+
+__code createTask1(struct LoopCounter* loopCounter, struct TaskManager* taskManager) {
+    Array* array = &ALLOCATE_DATA_GEAR(context, Array)->Array;
+
+    par goto createArray(array, __exit);
+
+    par goto twice(array, iterate(split), __exit);
+    goto code2();
+}
+
+void init(int argc, char** argv) {
+    for (int i = 1; argv[i]; ++i) {
+        if (strcmp(argv[i], "-cpu") == 0)
+            cpu_num = (int)atoi(argv[i+1]);
+        else if (strcmp(argv[i], "-l") == 0)
+            length = (int)atoi(argv[i+1]);
+        else if (strcmp(argv[i], "-s") == 0)
+            split = (int)atoi(argv[i+1]);
+        else if (strcmp(argv[i], "-cuda") == 0) {
+            gpu_num = 1;
+            CPU_CUDA = 0;
+        }
+    }
+}
+
+
+int main(int argc, char** argv) {
+    init(argc, argv);
+
+    array_ptr = NEWN(length, int);
+
+    for(int i=0; i<length; i++)
+        array_ptr[i]=i;
+
+    struct Context* main_context = NEW(struct Context);
+    initContext(main_context);
+    main_context->next = C_initDataGears;
+
+    goto start_code(main_context);
+}