#include <stdio.h>

#include "../../../context.h"

#ifdef USE_CUDAWorker
extern void cudaLoadFunction(struct Context* context, char* filename, char* function);
#endif

__code twice(struct Array* array, struct MultiDim* multiDim, __code next(struct Array* output, ...), struct LoopCounter* loopCounter) {
    int i = loopCounter->i;
    int index = multiDim->x;
    if (i < array->prefix) {
        array->array[i+index*array->prefix].value = array->array[i+index*array->prefix].value*2;
        loopCounter->i++;

        goto meta(context, C_twice);
    }

    loopCounter->i = 0;
    output->array = array->array;
    goto next(output, ...);
}

__code twice_stub(struct Context* context) {
#ifdef USE_CUDAWorker
    if (context->gpu) {
        Array* inputArray  = &context->data[context->idg]->Array;
        Array* outputArray = &context->data[context->odg]->Array;
        CUDABuffer* buffer = &ALLOCATE(context, CUDABuffer)->CUDABuffer;
        buffer->inputData = (union Data**)ALLOCATE_PTR_ARRAY(context, Array, 2);
        buffer->inputData[0] = (union Data*)inputArray->array;
        buffer->inputData[1] = (union Data*)inputArray;
        buffer->outputData = NULL;
        buffer->inputLen = 2;
        buffer->outputLen = 0;
        Executor* executor = context->worker->worker->CUDAWorker.executor;
        executor->executor->CUDAExecutor.buffer = buffer;
        cudaLoadFunction(context, "c/examples/twice/CUDAtwice.ptx", "twice");
        outputArray->array = inputArray->array;
        Gearef(context, Executor)->executor = (union Data*)executor;
        Gearef(context, Executor)->task = context;
        Gearef(context, Executor)->next = context->next;
        goto meta(context, executor->read);
    }
#endif
    Array** O_output = (struct Array **)&context->data[context->odg];
    goto twice(context,
               &context->data[context->idg]->Array,
               &context->data[context->idg+1]->MultiDim,
               context->next,
               O_output,
               Gearef(context, LoopCounter));
}