#include <moc.h>
#include <stdlib.h>
#include <string.h>
#include <malloc_align.h>

extern spe_program_handle_t test_jacobi_spu;

int main(int argc, char **argv)
    {
    int t, rank, size, rl, mspu, aspu, err, stringl, nothr;
    char proc[4];

    int n __attribute__((__aligned__(16)));
    int iterations __attribute__((__aligned__(16)));
    float omega __attribute__((__aligned__(16)));
    char string[255];

    if (argc < 5)
	{
	printf("Usage: \"%s Matrix size Iterations Omega nothr\"\n", argv[0]);
	return -1;
	}

    n = atoi(argv[1]);
    iterations = atoi(argv[2]);
    omega = atof(argv[3]);
    nothr = atoi(argv[4]);

    if (MOC_Get_max_spu_count(&mspu))
	{
	printf("An error occured during MOC_Get_max_spu_count\n");
	return 0;
	}
    if (MOC_Get_avail_spu_count(&aspu))
	{
	printf("An error occured during MOC_Get_avail_spu_count\n");
	return 0;
	}

    // only use as many spus as specified
    aspu = nothr;
    for (t = 0; t < aspu; t++)
	MOC_Add_spu_thread(&test_jacobi_spu, NULL);

    MOC_Init(NULL, NULL);

    MOC_Comm_rank(MOC_COMM_WORLD, &rank);
    MOC_Comm_size(MOC_COMM_WORLD, &size);
    MOC_Get_processor_name(proc, &rl);

    // sending input and other runtime specific data to spes
    for (t = 0; t < aspu; t++)
	{
	err = MOC_Bsend(&n, sizeof(int), MOC_INT, t + 1, 123, MOC_COMM_WORLD);
	if (err)
	    {
	    MOC_Error_string(err, string, &stringl);
	    printf("%i: ERROR-bsend1: %s (%i)\n", rank, string, err);
	    return 0;
	    }

	err = MOC_Bsend(&iterations, sizeof(int), MOC_INT, t + 1, 123,
		MOC_COMM_WORLD);
	if (err)
	    {
	    MOC_Error_string(err, string, &stringl);
	    printf("%i: ERROR-bsend2: %s (%i)\n", rank, string, err);
	    return 0;
	    }

	err = MOC_Bsend(&omega, sizeof(float), MOC_FLOAT, t + 1, 123,
		MOC_COMM_WORLD);
	if (err)
	    {
	    MOC_Error_string(err, string, &stringl);
	    printf("%i: ERROR-bsend3: %s (%i)\n", rank, string, err);
	    return 0;
	    }
	}

    MOC_Finalize();
    return 0;
    }