su3 slow_expon(su3 in, int nterm) {
su3 out , out_new , aa , aa_old , aa_scale ;
int i ;
double fact = 1 ;
_su3_one(out) ;
aa = in ;
aa_old = aa ;
_su3_plus_su3(out_new,out,aa) ;
out = out_new ;
for(i=2 ; i < nterm ; ++i) {
_su3_times_su3(aa,in,aa_old);
aa_old = aa ;
aa_scale = aa ;
fact *= i ;
scale_su3(&aa_scale, 1.0/fact ) ;
_su3_plus_su3(out_new,out,aa_scale) ;
out = out_new ;
}
return(out);
}
void sw_spinor(const int ieo, spinor * const kk, spinor * const ll) {
int ioff;
int icx;
int x;
spinor *r,*s;
static su3 v0,v1,v2,v3;
static su3 u0,u1,u2,u3;
static su3 lswp[4],lswm[4];
if(ieo == 0) {
ioff=0;
}
else {
ioff=(VOLUME+RAND)/2;
}
/************************ loop over half of the lattice sites ***********/
for(icx = ioff; icx < (VOLUME/2+ioff); icx++) {
x = g_eo2lexic[icx];
r = kk + icx - ioff;
s = ll + icx - ioff;
_vector_tensor_vector(v0,(*r).s0,(*s).s0);
_vector_tensor_vector(v1,(*r).s0,(*s).s1);
_vector_tensor_vector(v2,(*r).s1,(*s).s1);
_vector_tensor_vector(v3,(*r).s1,(*s).s0);
_vector_tensor_vector(u0,(*r).s2,(*s).s2);
_vector_tensor_vector(u1,(*r).s2,(*s).s3);
_vector_tensor_vector(u2,(*r).s3,(*s).s3);
_vector_tensor_vector(u3,(*r).s3,(*s).s2);
/* compute the insertion matrix */
_su3_plus_su3(lswp[0],u0,v0);
_su3_plus_su3(lswp[1],u1,v1);
_su3_plus_su3(lswp[2],u2,v2);
_su3_plus_su3(lswp[3],u3,v3);
_su3_minus_su3(lswm[0],u0,v0);
_su3_minus_su3(lswm[1],u1,v1);
_su3_minus_su3(lswm[2],u2,v2);
_su3_minus_su3(lswm[3],u3,v3);
/* add up the swm[0] and swp[0] */
_su3_acc(swm[x][0], lswm[0]);
_su3_acc(swm[x][1], lswm[1]);
_su3_acc(swm[x][2], lswm[2]);
_su3_acc(swm[x][3], lswm[3]);
_su3_acc(swp[x][0], lswp[0]);
_su3_acc(swp[x][1], lswp[1]);
_su3_acc(swp[x][2], lswp[2]);
_su3_acc(swp[x][3], lswp[3]);
}
return;
}
void sw_term(su3 ** const gf, const double kappa, const double c_sw) {
int k,l;
int x,xpk,xpl,xmk,xml,xpkml,xplmk,xmkml;
su3 *w1,*w2,*w3,*w4;
double ka_csw_8 = kappa*c_sw/8.;
static su3 v1,v2,plaq;
static su3 fkl[4][4];
static su3 magnetic[4],electric[4];
static su3 aux;
/* compute the clover-leave */
/* l __ __
| | | |
|__| |__|
__ __
| | | |
|__| |__| k */
for(x = 0; x < VOLUME; x++) {
for(k = 0; k < 4; k++) {
for(l = k+1; l < 4; l++) {
xpk=g_iup[x][k];
xpl=g_iup[x][l];
xmk=g_idn[x][k];
xml=g_idn[x][l];
xpkml=g_idn[xpk][l];
xplmk=g_idn[xpl][k];
xmkml=g_idn[xml][k];
w1=&gf[x][k];
w2=&gf[xpk][l];
w3=&gf[xpl][k];
w4=&gf[x][l];
_su3_times_su3(v1,*w1,*w2);
_su3_times_su3(v2,*w4,*w3);
_su3_times_su3d(plaq,v1,v2);
w1=&gf[x][l];
w2=&gf[xplmk][k];
w3=&gf[xmk][l];
w4=&gf[xmk][k];
_su3_times_su3d(v1,*w1,*w2);
_su3d_times_su3(v2,*w3,*w4);
_su3_times_su3_acc(plaq,v1,v2);
w1=&gf[xmk][k];
w2=&gf[xmkml][l];
w3=&gf[xmkml][k];
w4=&gf[xml][l];
_su3_times_su3(v1,*w2,*w1);
_su3_times_su3(v2,*w3,*w4);
_su3d_times_su3_acc(plaq,v1,v2);
w1=&gf[xml][l];
w2=&gf[xml][k];
w3=&gf[xpkml][l];
w4=&gf[x][k];
_su3d_times_su3(v1,*w1,*w2);
_su3_times_su3d(v2,*w3,*w4);
_su3_times_su3_acc(plaq,v1,v2);
_su3_dagger(v2,plaq);
_su3_minus_su3(fkl[k][l],plaq,v2);
}
}
// this is the one in flavour and colour space
// twisted mass term is treated in clover, sw_inv and
// clover_gamma5
_su3_one(sw[x][0][0]);
_su3_one(sw[x][2][0]);
_su3_one(sw[x][0][1]);
_su3_one(sw[x][2][1]);
for(k = 1; k < 4; k++)
{
_su3_assign(electric[k], fkl[0][k]);
}
_su3_assign(magnetic[1], fkl[2][3]);
_su3_minus_assign(magnetic[2], fkl[1][3]);
_su3_assign(magnetic[3], fkl[1][2]);
/* upper left block 6x6 matrix */
_itimes_su3_minus_su3(aux,electric[3],magnetic[3]);
_su3_refac_acc(sw[x][0][0],ka_csw_8,aux);
_itimes_su3_minus_su3(aux,electric[1],magnetic[1]);
_su3_minus_su3(v2,electric[2],magnetic[2]);
_su3_acc(aux,v2);
_real_times_su3(sw[x][1][0],ka_csw_8,aux);
_itimes_su3_minus_su3(aux,magnetic[3],electric[3]);
_su3_refac_acc(sw[x][2][0],ka_csw_8,aux);
/* lower right block 6x6 matrix */
_itimes_su3_plus_su3(aux,electric[3],magnetic[3]);
_su3_refac_acc(sw[x][0][1],(-ka_csw_8),aux);
_itimes_su3_plus_su3(aux,electric[1],magnetic[1]);
_su3_plus_su3(v2,electric[2],magnetic[2]);
_su3_acc(aux,v2);
_real_times_su3(sw[x][1][1],(-ka_csw_8),aux);
_itimes_su3_plus_su3(aux,magnetic[3],electric[3]);
_su3_refac_acc(sw[x][2][1],ka_csw_8,aux);
}
return;
}
void sw_deriv(const int ieo, const double mu) {
int ioff;
int x;
double fac = 1.0000;
static su3 lswp[4],lswm[4];
/* convention: Tr clover-leaf times insertion */
if(ieo == 0) {
ioff=0;
}
else {
ioff = (VOLUME+RAND)/2;
}
if(fabs(mu) > 0.) fac = 0.5;
for(int icx = ioff, icy=0; icx < (VOLUME/2+ioff); icx++, icy++) {
x = g_eo2lexic[icx];
/* compute the insertion matrix */
_su3_plus_su3(lswp[0],sw_inv[icy][0][1],sw_inv[icy][0][0]);
_su3_plus_su3(lswp[1],sw_inv[icy][1][1],sw_inv[icy][1][0]);
_su3_plus_su3(lswp[2],sw_inv[icy][2][1],sw_inv[icy][2][0]);
_su3_plus_su3(lswp[3],sw_inv[icy][3][1],sw_inv[icy][3][0]);
_su3_minus_su3(lswm[0],sw_inv[icy][0][1],sw_inv[icy][0][0]);
_su3_minus_su3(lswm[1],sw_inv[icy][1][1],sw_inv[icy][1][0]);
_su3_minus_su3(lswm[2],sw_inv[icy][2][1],sw_inv[icy][2][0]);
_su3_minus_su3(lswm[3],sw_inv[icy][3][1],sw_inv[icy][3][0]);
/* add up to swm[] and swp[] */
_su3_refac_acc(swm[x][0], fac, lswm[0]);
_su3_refac_acc(swm[x][1], fac, lswm[1]);
_su3_refac_acc(swm[x][2], fac, lswm[2]);
_su3_refac_acc(swm[x][3], fac, lswm[3]);
_su3_refac_acc(swp[x][0], fac, lswp[0]);
_su3_refac_acc(swp[x][1], fac, lswp[1]);
_su3_refac_acc(swp[x][2], fac, lswp[2]);
_su3_refac_acc(swp[x][3], fac, lswp[3]);
if(fabs(mu) > 0.) {
/* compute the insertion matrix */
_su3_plus_su3(lswp[0],sw_inv[icy+VOLUME/2][0][1],sw_inv[icy+VOLUME/2][0][0]);
_su3_plus_su3(lswp[1],sw_inv[icy+VOLUME/2][1][1],sw_inv[icy+VOLUME/2][1][0]);
_su3_plus_su3(lswp[2],sw_inv[icy+VOLUME/2][2][1],sw_inv[icy+VOLUME/2][2][0]);
_su3_plus_su3(lswp[3],sw_inv[icy+VOLUME/2][3][1],sw_inv[icy+VOLUME/2][3][0]);
_su3_minus_su3(lswm[0],sw_inv[icy+VOLUME/2][0][1],sw_inv[icy+VOLUME/2][0][0]);
_su3_minus_su3(lswm[1],sw_inv[icy+VOLUME/2][1][1],sw_inv[icy+VOLUME/2][1][0]);
_su3_minus_su3(lswm[2],sw_inv[icy+VOLUME/2][2][1],sw_inv[icy+VOLUME/2][2][0]);
_su3_minus_su3(lswm[3],sw_inv[icy+VOLUME/2][3][1],sw_inv[icy+VOLUME/2][3][0]);
/* add up to swm[] and swp[] */
_su3_refac_acc(swm[x][0], fac, lswm[0]);
_su3_refac_acc(swm[x][1], fac, lswm[1]);
_su3_refac_acc(swm[x][2], fac, lswm[2]);
_su3_refac_acc(swm[x][3], fac, lswm[3]);
_su3_refac_acc(swp[x][0], fac, lswp[0]);
_su3_refac_acc(swp[x][1], fac, lswp[1]);
_su3_refac_acc(swp[x][2], fac, lswp[2]);
_su3_refac_acc(swp[x][3], fac, lswp[3]);
}
}
return;
}
