int main(int argc, char** argv){
int retval = 0;
int i, Navg;
param_obj param;
kw_obj kw;
ts_obj ts;
vsip_mview_f *dtaIn, *gramOut;
if(retval += param_read(argv[1],¶m)){
printf("Failed to read parameter file\n");
exit(-1);
}
param_log(¶m);
kw_init(&kw, ¶m);
ts_init(&ts, ¶m);
/* simulate time series and beamform */
Navg = (int) param.Navg;
dtaIn = ts_instance(&ts);
gramOut = kw_instance(&kw);
kw_zero(&kw);
for(i=0; i<Navg; i++){
ts_zero(&ts);
ts_sim_noise(&ts);
ts_sim_nb(&ts);
komega(&kw,dtaIn);
}
/* beamform done. Gram should be in gramOut */
/*************/
/* Massage gram data and save to file for plotting */
for(i=0; i<vsip_mgetrowlength_f(gramOut); i++)
{/* move zero to middle */
vsip_vview_f *v = vsip_mcolview_f(gramOut, i);
vsip_vfreqswap_f(v);
vsip_vdestroy_f(v);
}
{/* massage the data for plot*/
vsip_scalar_f max = vsip_mmaxval_f(gramOut, NULL),min;
vsip_scalar_f avg = vsip_mmeanval_f(gramOut);
vsip_mclip_f(gramOut,0.0, max, avg/100000.0, max, gramOut);
vsip_mlog10_f(gramOut,gramOut);
min = -vsip_mminval_f(gramOut, NULL);
vsip_smadd_f(min, gramOut, gramOut);
max = vsip_mmaxval_f(gramOut, NULL);
vsip_smmul_f(1.0/max, gramOut, gramOut);
}
{ /* output data and plot with octave */
FILE *fptr = fopen("gramOut","w");
size_t size = vsip_mgetrowlength_f(gramOut) * vsip_mgetcollength_f(gramOut);
vsip_scalar_f *out = (vsip_scalar_f*)malloc(size * sizeof(vsip_scalar_f));
vsip_mcopyto_user_f(gramOut, VSIP_COL, out);
fwrite(out,size,sizeof(vsip_scalar_f),fptr);
fclose(fptr);
free(out);
}
/* cleanup */
kw_fin(&kw);
ts_fin(&ts);
param_free(¶m);
return retval;
}
void (vsip_gemp_f)(vsip_scalar_f alpha,
const vsip_mview_f *AA,
vsip_mat_op OpA,
const vsip_mview_f *BB,
vsip_mat_op OpB,
vsip_scalar_f beta,
const vsip_mview_f *r){
register vsip_scalar_f a1 = alpha;
register vsip_scalar_f b1 = beta;
vsip_mview_f At = *AA,
Bt = *BB;
vsip_mview_f *a = &At,
*b = &Bt;
if(OpA == VSIP_MAT_TRANS){
a->row_length = AA->col_length;
a->col_length = AA->row_length;
a->row_stride = AA->col_stride;
a->col_stride = AA->row_stride;
}
if(OpB == VSIP_MAT_TRANS){
b->row_length = BB->col_length;
b->col_length = BB->row_length;
b->row_stride = BB->col_stride;
b->col_stride = BB->row_stride;
}
{
/* c => column major, r => row major */
/* decide if ccc => 0 ccr=> 1, crc => 2, etc to rrr => 7 */
unsigned method = (unsigned)(r->row_stride <= r->col_stride) +
(unsigned)(b->row_stride <= b->col_stride) * 2 +
(unsigned)(a->row_stride <= a->col_stride) * 4;
vsip_stride a_st_r = a->row_stride * a->block->rstride,
a_st_c = a->col_stride * a->block->rstride,
b_st_r = b->row_stride * b->block->rstride,
b_st_c = b->col_stride * b->block->rstride,
r_st_r = r->row_stride * r->block->rstride,
r_st_c = r->col_stride * r->block->rstride;
vsip_length a_c_l = a->col_length; /* i_size */
vsip_length r_r_l = r->row_length; /* j_size */
vsip_length a_r_l = a->row_length; /* k_size */
register vsip_scalar_f a_scalar, b_scalar, temp;
vsip_scalar_f *ap = (a->block->array) + a->offset * a->block->rstride,
*bp = (b->block->array) + b->offset * b->block->rstride,
*rp = (r->block->array) + r->offset * r->block->rstride;
vsip_scalar_f *ap0 = ap,
*bp0 = bp,
*rp0 = rp;
vsip_scalar_f *bp1 = bp;
switch(method){
case 0 : { /* ccc */
vsip_scalar_f *ap_ik = ap0, *ap_0k = ap0;
vsip_scalar_f *bp_kj = bp0, *bp_0j = bp0;
vsip_scalar_f *rp_ij = rp0, *rp_0j = rp0;
if(b1 != 1.0) vsip_smmul_f(b1,r,r);
while(r_r_l-- > 0){ /* j */
rp_ij = rp_0j;
bp_kj = bp_0j;
while(a_r_l-- > 0){ /* k */
b_scalar = a1 * *bp_kj;
bp_kj += b_st_c;
ap_ik = ap_0k;
while(a_c_l-- > 0){ /* i */
*rp_ij = *ap_ik * b_scalar + *rp_ij;
rp_ij += r_st_c;
ap_ik += a_st_c;
} a_c_l = a->col_length;
ap_0k += a_st_r;
rp_ij = rp_0j;
} a_r_l = a->row_length;
rp_0j += r_st_r;
bp_0j += b_st_r;
ap_0k = ap0;
}
} break;
case 1 : { /* ccr */
vsip_scalar_f *ap_ik = ap0, *ap_i0 = ap0;
vsip_scalar_f *bp_kj = bp0, *bp_0j = bp0;
vsip_scalar_f *rp_ij = rp0, *rp_i0 = rp0;
while(a_c_l-- > 0){ /* i */
rp_ij = rp_i0;
bp_0j = bp0;
while(r_r_l-- > 0){ /* j */
ap_ik = ap_i0;
temp = 0;
bp_kj = bp_0j;
while(a_r_l-- > 0){ /* k */
temp += *ap_ik * *bp_kj;
ap_ik += a_st_r;
bp_kj += b_st_c;
} a_r_l = a->row_length;
*rp_ij = b1 * *rp_ij + a1 * temp;
rp_ij += r_st_r;
bp_0j += b_st_r;
} r_r_l = r->row_length;
rp_i0 += r_st_c;
ap_i0 += a_st_c;
}
} break;
case 2 : { /* crc */
vsip_scalar_f *ap_ik = ap0, *ap_0k = ap0;
vsip_scalar_f *bp_kj = bp0, *bp_0j = bp0;
vsip_scalar_f *rp_ij = rp0, *rp_0j = rp0;
if(b1 != 1.0) vsip_smmul_f(b1,r,r);
while(r_r_l-- > 0){ /* j */
bp_kj = bp_0j;
while(a_r_l-- > 0){ /* k */
b_scalar = a1 * *bp_kj;
ap_ik = ap_0k;
rp_ij = rp_0j;
while(a_c_l-- > 0){ /* i */
*rp_ij += *ap_ik * b_scalar;
rp_ij += r_st_c;
ap_ik += a_st_c;
} a_c_l = a->col_length;
ap_0k += a_st_r;
bp_kj += b_st_c;
} a_r_l = a->row_length;
rp_0j += r_st_r;
bp_0j += b_st_r;
ap_0k = ap0;
}
} break;
case 3 : { /* crr */
vsip_scalar_f *ap_ik = ap0, *ap_i0 = ap0;
vsip_scalar_f *bp_kj = bp0, *bp_k0 = bp0;
vsip_scalar_f *rp_ij = rp0, *rp_i0 = rp0;
if(b1 != 1.0) vsip_smmul_f(b1,r,r);
while(a_c_l-- > 0){ /* i */
ap_ik = ap_i0;
while(a_r_l-- > 0){ /* k */
a_scalar = a1 * *ap_ik;
ap_ik += a_st_r;
bp_kj = bp_k0;
rp_ij = rp_i0;
while(r_r_l-- > 0){ /* j */
*rp_ij = *bp_kj * a_scalar + *rp_ij;
rp_ij += r_st_r;
bp_kj += b_st_r;
} r_r_l = r->row_length;
bp_k0 += b_st_c;
}a_r_l = a->row_length;
bp_k0 = bp0;
rp_i0 += r_st_c;
ap_i0 += a_st_c;
}
} break;
case 4 : { /* rcc */
while(a_c_l-- > 0){
while(r_r_l-- > 0){
temp = 0;
while(a_r_l-- > 0){
temp += *ap * *bp;
ap += a_st_r; bp += b_st_c;
}
*rp = b1 * *rp + a1 * temp;
rp += r_st_r;
ap = ap0;
bp0 += b_st_r; bp = bp0;
a_r_l = a->row_length;
}
ap0 += a_st_c; ap = ap0;
bp0 = bp1; bp = bp0;
rp0 += r_st_c; rp = rp0;
r_r_l = r->row_length;
}
} break;
case 5 : { /* rcr */
while(a_c_l-- > 0){
while(r_r_l-- > 0){
temp = 0;
while(a_r_l-- > 0){
temp += *ap * *bp;
ap += a_st_r; bp += b_st_c;
}
*rp = b1 * *rp + a1 * temp;
rp += r_st_r;
ap = ap0;
bp0 += b_st_r; bp = bp0;
a_r_l = a->row_length;
}
ap0 += a_st_c; ap = ap0;
bp0 = bp1; bp = bp0;
rp0 += r_st_c; rp = rp0;
r_r_l = r->row_length;
}
} break;
case 6 : { /* rrc */
vsip_scalar_f *ap_ik = ap0, *ap_i0 = ap0;
vsip_scalar_f *bp_kj = bp0, *bp_0j = bp0;
vsip_scalar_f *rp_ij = rp0, *rp_0j = rp0;
/* jik */
while(r_r_l-- > 0){ /* j */
rp_ij = rp_0j;
ap_i0 = ap0;
while(a_c_l-- > 0){ /* i */
temp = 0;
ap_ik = ap_i0;
bp_kj = bp_0j;
while(a_r_l-- > 0){ /* k */
temp += *ap_ik * *bp_kj;
ap_ik += a_st_r;
bp_kj += b_st_c;
} a_r_l = a->row_length;
*rp_ij = a1 * temp + b1 * *rp_ij;
rp_ij +=r_st_c;
ap_i0 += a_st_c;
} a_c_l = a->col_length;
rp_0j += r_st_r;
bp_0j += b_st_r;
}
} break;
case 7 : { /* rrr */
vsip_scalar_f *ap_ik = ap0, *ap_i0 = ap0;
vsip_scalar_f *bp_kj = bp0, *bp_k0 = bp0;
vsip_scalar_f *rp_ij = rp0, *rp_i0 = rp0;
if(b1 != 1.0) vsip_smmul_f(b1,r,r);
while(a_c_l-- > 0){ /* i */
ap_ik = ap_i0;
while(a_r_l-- > 0){ /* k */
a_scalar = a1 * *ap_ik;
ap_ik += a_st_r;
bp_kj = bp_k0;
rp_ij = rp_i0;
while(r_r_l-- > 0){ /* j */
*rp_ij = *bp_kj * a_scalar + *rp_ij;
rp_ij += r_st_r;
bp_kj += b_st_r;
} r_r_l = r->row_length;
bp_k0 += b_st_c;
}a_r_l = a->row_length;
bp_k0 = bp0;
rp_i0 += r_st_c;
ap_i0 += a_st_c;
}
}
}
}
}
