only one hopping body for most of the ifdefs

Hopping_Matrix.c

@@ -84,7 +84,6 @@
 #endif
 #include "global.h"
 #include "su3.h"
-#include "sse.h"
 #ifdef MPI
 #  include "xchange_field.h"
 #  ifdef _USE_TSPLITPAR
@@ -101,6 +100,7 @@
 
 #if defined _USE_HALFSPINOR
 #  if ((defined SSE2)||(defined SSE3))
+#    include "sse.h"
 #    include "operator/halfspinor_sse_dbl.c"
 
 #  elif (defined BGL && defined XLC)
@@ -118,19 +118,32 @@
 
 #else /* thats _USE_HALFSPINOR */
 
-#  if ((defined SSE2)||(defined SSE3))
+#  if (((defined SSE2)||(defined SSE3)) && defined _USE_TSPLITPAR)
+#    include "sse.h"
 #    include "operator/hopping_sse_dbl.c"
 
-#  elif (defined BGQ && defined XLC)
-#    include "bgq.h"
-#    include "operator/hopping_bgq_dbl.c"
-
 #  elif (defined BGL && defined XLC)
 #    include "bgl.h"
 #    include "operator/hopping_bg_dbl.c"
 
 #  else
-#    include "operator/hopping_dbl.c"
+#    if ((defined SSE2)||(defined SSE3))
+#      include "sse.h"
+
+#    elif (defined BGQ && defined XLC)
+#      include "bgq.h"
+#      include "xlc_prefetch.h"
+
+#    elif defined XLC
+#      include"xlc_prefetch.h"
+
+#    endif
+void Hopping_Matrix(const int ieo, spinor * const l, spinor * const k) {
+
+#    include "operator/hopping_body_dbl.c"
+
+  return;
+}
 #  endif
 
 #endif /* thats _USE_HALFSPINOR */

X_psi.c

@@ -22,6 +22,7 @@
 # include<config.h>
 #endif
 #include <stdlib.h>
+#include <math.h>
 #ifdef HAVE_CONFIG_H
 # include<config.h>
 #endif

bgq2.h

@@ -917,6 +917,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   return;
 }
 
+#ifdef _declare_regs
+#  undef _declare_regs
+#endif
 /* We have 32 registers available */
 #define _declare_regs()							\
   vector4double ALIGN r0, r1, r2, r3, r4, r5, r6, r7, r8, r9, r10, r11; \
@@ -926,6 +929,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   __alignx(16,l);							\
   __alignx(16,k);
 
+#ifdef _hop_t_p
+#  undef _hop_t_p
+#endif
 #define _hop_t_p()							\
   _vec_load_spinor(r4, r5, r6, r7, r8, r9, sp->s0);			\
   _vec_add_ul_spinor(r0, r1, r2, r4, r5, r6, r7, r8, r9);		\
@@ -936,6 +942,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   rs6 = rs0; rs7 = rs1; rs8 = rs2;					\
   rs9 = rs3; rs10= rs4; rs11= rs5;
 
+#ifdef _hop_t_m
+#  undef _hop_t_m
+#endif
 #define _hop_t_m()							\
   _vec_load_spinor(r4, r5, r6, r7, r8, r9, sm->s0);			\
   _vec_sub_ul_spinor(r0, r1, r2, r4, r5, r6, r7, r8, r9);		\
@@ -945,6 +954,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   _vec_add_double2(rs0, rs1, rs2, rs3, rs4, rs5, r0, r1, r2, r3, r4, r5); \
   _vec_sub_double2(rs6, rs7, rs8, rs9, rs10, rs11, r0, r1, r2, r3, r4, r5);
 
+#ifdef _hop_x_p
+#  undef _hop_x_p
+#endif
 #define _hop_x_p()							\
   _vec_load(r4, r5, sp->s0);						\
   _vec_load16(r6, r7, sp->s1, U0);					\
@@ -960,6 +972,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   _vec_i_mul_sub2(rs6, rs7, rs8, r3, r4, r5, U0);			\
   _vec_i_mul_sub2(rs9, rs10, rs11, r0, r1, r2, U1);
 
+#ifdef _hop_x_m
+#  undef _hop_x_m
+#endif
 #define _hop_x_m()							\
   _vec_load(r4, r5, sm->s0);						\
   _vec_load16(r6, r7, sm->s1, U0);					\
@@ -974,6 +989,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   _vec_i_mul_add2(rs6, rs7, rs8, r3, r4, r5, U0);			\
   _vec_i_mul_add2(rs9, rs10, rs11, r0, r1, r2, U1);
 
+#ifdef _hop_y_p
+#  undef _hop_y_p
+#endif
 #define _hop_y_p()							\
   _vec_load(r4, r5, sp->s0);						\
   _vec_load16(r6, r7, sp->s1, U0);					\
@@ -989,6 +1007,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   _vec_sub2(rs6, rs7, rs8, r3, r4, r5);					\
   _vec_add2(rs9, rs10, rs11, r0, r1, r2);
 
+#ifdef _hop_y_m
+#  undef _hop_y_m
+#endif
 #define _hop_y_m()							\
   _vec_load(r4, r5, sm->s0);						\
   _vec_load16(r6, r7, sm->s1, U0);					\
@@ -1003,6 +1024,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   _vec_add2(rs6, rs7, rs8, r3, r4, r5);					\
   _vec_sub2(rs9, rs10, rs11, r0, r1, r2);
 
+#ifdef _hop_z_p
+#  undef _hop_z_p
+#endif
 #define _hop_z_p()							\
   _vec_load(r4, r5, sp->s0);						\
   _vec_load16(r6, r7, sp->s1, U0);					\
@@ -1018,6 +1042,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   _vec_i_mul_sub2(rs6, rs7, rs8, r0, r1, r2, U0);			\
   _vec_i_mul_add2(rs9, rs10, rs11, r3, r4, r5, U1);
 
+#ifdef _hop_z_m
+#  undef _hop_z_m
+#endif
 #define _hop_z_m()							\
   _vec_load(r4, r5, sm->s0);						\
   _vec_load16(r6, r7, sm->s1, U0);					\
@@ -1032,6 +1059,9 @@ inline void vec_su3_multiply_double2b(su3 * const u, vector4double * U, vector4d
   _vec_i_mul_add2(rs6, rs7, rs8, r0, r1, r2, U0);			\
   _vec_i_mul_sub2(rs9, rs10, rs11, r3, r4, r5, U1);
 
+#ifdef _store_res
+#  undef _store_res
+#endif
 #define _store_res()				\
   _vec_store2(rn->s0, rs0, rs1, rs2);		\
   _vec_store2(rn->s1, rs3, rs4, rs5);		\

linalg/assign_add_mul_r_add_mul.c

@@ -29,9 +29,9 @@
 #ifdef OMP
 # include <omp.h>
 #endif
-#include "sse.h"
 #include "su3.h"
 #include "su3adj.h"
+#include "sse.h"
 #include "assign_add_mul_r_add_mul.h"
 
 #if ( defined SSE2 || defined SSE3 )

linalg/assign_add_mul_r_bi.c

@@ -15,9 +15,6 @@
  * 
  * You should have received a copy of the GNU General Public License
  * along with tmLQCD.  If not, see <http://www.gnu.org/licenses/>.
- ***********************************************************************/
-
-/************************************************************************
  *
  *      Adpated routine evaluating the P=P+c*Q where P,Q are bispinors
  *
@@ -38,8 +35,8 @@
 #ifdef OMP
 # include <omp.h>
 #endif
-#include "sse.h"
 #include "su3.h"
+#include "sse.h"
 #include "assign_add_mul_r_bi.h"
 
 

linalg/assign_mul_bra_add_mul_ket_add_bi.c

@@ -15,8 +15,6 @@
  * 
  * You should have received a copy of the GNU General Public License
  * along with tmLQCD.  If not, see <http://www.gnu.org/licenses/>.
- ***********************************************************************/
-/*******************************************************************************
  *
  * File assign_mul_bra_add_mul_ket_add.c
  *
@@ -38,8 +36,8 @@
 #include <stdlib.h>
 #include <stdio.h>
 #include <math.h>
-#include "sse.h"
 #include "su3.h"
+#include "sse.h"
 #include "assign_mul_bra_add_mul_ket_add_bi.h"
 
 /* R inoutput, S input, U input, c1 input, c2 input */

measure_gauge_action.c

@@ -32,9 +32,9 @@
 #include <stdlib.h>
 #include <stdio.h>
 #include <math.h>
-#include "sse.h"
 #include "su3.h"
 #include "su3adj.h"
+#include "sse.h"
 #include "geometry_eo.h"
 #include "global.h"
 #include <io/params.h>

operator/hopping_bgq_dbl.c → operator/hopping_body_dbl.c

@@ -24,11 +24,9 @@
  *
  **********************************************************************/
 
-#include "bgq.h"
-#include "xlc_prefetch.h"
-
-void Hopping_Matrix(const int ieo, spinor * const l, spinor * const k) {
-#pragma disjoint(*l, *k)
+#ifdef XLC
+#  pragma disjoint(*l, *k)
+#endif
 #ifdef _GAUGE_COPY
   if(g_update_gauge_copy) {
     update_backward_gauge(g_gauge_field);
@@ -51,7 +49,9 @@ void Hopping_Matrix(const int ieo, spinor * const l, spinor * const k) {
   spinor * restrict ALIGN sm;
   spinor * restrict ALIGN rn;
 
-#pragma disjoint(*sp, *sm, *rn, *up, *um, *l)
+#ifdef XLC
+#  pragma disjoint(*sp, *sm, *rn, *up, *um, *l)
+#endif
   _declare_regs();
 
   if(ieo == 0){
@@ -195,5 +195,3 @@ void Hopping_Matrix(const int ieo, spinor * const l, spinor * const k) {
 #ifdef OMP
   } /* OpenMP closing brace */
 #endif
-  return;
-}

operator/hopping_dbl.c

@@ -26,114 +26,6 @@
  *
  **********************************************************************/
 
-#include"xlc_prefetch.h"
-
-#define _declare_regs()	     \
-  su3_vector ALIGN psi, chi; \
-  spinor ALIGN temp;
-
-#define _hop_t_p()				\
-  _vector_add(psi,sp->s0,sp->s2);		\
-  _su3_multiply(chi,(*up),psi);			\
-  _complex_times_vector(psi,ka0,chi);		\
-  _vector_assign(temp.s0,psi);			\
-  _vector_assign(temp.s2,psi);			\
-  _vector_add(psi,sp->s1,sp->s3);		\
-  _su3_multiply(chi,(*up),psi);			\
-  _complex_times_vector(psi,ka0,chi);		\
-  _vector_assign(temp.s1,psi);			\
-  _vector_assign(temp.s3,psi);
-
-#define _hop_t_m()				\
-  _vector_sub(psi,sm->s0,sm->s2);		\
-  _su3_inverse_multiply(chi,(*um),psi);		\
-  _complexcjg_times_vector(psi,ka0,chi);	\
-  _vector_add_assign(temp.s0,psi);		\
-  _vector_sub_assign(temp.s2,psi);		\
-  _vector_sub(psi,sm->s1,sm->s3);		\
-  _su3_inverse_multiply(chi,(*um),psi);		\
-  _complexcjg_times_vector(psi,ka0,chi);	\
-  _vector_add_assign(temp.s1,psi);		\
-  _vector_sub_assign(temp.s3,psi);
-
-#define _hop_x_p()				\
-  _vector_i_add(psi,sp->s0,sp->s3);		\
-  _su3_multiply(chi,(*up),psi);			\
-  _complex_times_vector(psi,ka1,chi);		\
-  _vector_add_assign(temp.s0,psi);		\
-  _vector_i_sub_assign(temp.s3,psi);		\
-  _vector_i_add(psi,sp->s1,sp->s2);		\
-  _su3_multiply(chi,(*up),psi);			\
-  _complex_times_vector(psi,ka1,chi);		\
-  _vector_add_assign(temp.s1,psi);		\
-  _vector_i_sub_assign(temp.s2,psi);
-
-#define _hop_x_m()				\
-  _vector_i_sub(psi,sm->s0,sm->s3);		\
-  _su3_inverse_multiply(chi,(*um),psi);		\
-  _complexcjg_times_vector(psi,ka1,chi);	\
-  _vector_add_assign(temp.s0,psi);		\
-  _vector_i_add_assign(temp.s3,psi);		\
-  _vector_i_sub(psi,sm->s1,sm->s2);		\
-  _su3_inverse_multiply(chi,(*um),psi);		\
-  _complexcjg_times_vector(psi,ka1,chi);	\
-  _vector_add_assign(temp.s1,psi);		\
-  _vector_i_add_assign(temp.s2,psi);
-
-#define _hop_y_p()				\
-  _vector_add(psi,sp->s0,sp->s3);		\
-  _su3_multiply(chi,(*up),psi);			\
-  _complex_times_vector(psi,ka2,chi);		\
-  _vector_add_assign(temp.s0,psi);		\
-  _vector_add_assign(temp.s3,psi);		\
-  _vector_sub(psi,sp->s1,sp->s2);		\
-  _su3_multiply(chi,(*up),psi);			\
-  _complex_times_vector(psi,ka2,chi);		\
-  _vector_add_assign(temp.s1,psi);		\
-  _vector_sub_assign(temp.s2,psi);
-
-#define _hop_y_m()				\
-  _vector_sub(psi,sm->s0,sm->s3);		\
-  _su3_inverse_multiply(chi,(*um),psi);		\
-  _complexcjg_times_vector(psi,ka2,chi);	\
-  _vector_add_assign(temp.s0,psi);		\
-  _vector_sub_assign(temp.s3,psi);		\
-  _vector_add(psi,sm->s1,sm->s2);		\
-  _su3_inverse_multiply(chi,(*um),psi);		\
-  _complexcjg_times_vector(psi,ka2,chi);	\
-  _vector_add_assign(temp.s1,psi);		\
-  _vector_add_assign(temp.s2,psi);
-
-#define _hop_z_p()				\
-  _vector_i_add(psi,sp->s0,sp->s2);		\
-  _su3_multiply(chi,(*up),psi);			\
-  _complex_times_vector(psi,ka3,chi);		\
-  _vector_add_assign(temp.s0,psi);		\
-  _vector_i_sub_assign(temp.s2,psi);		\
-  _vector_i_sub(psi,sp->s1,sp->s3);		\
-  _su3_multiply(chi,(*up),psi);			\
-  _complex_times_vector(psi,ka3,chi);		\
-  _vector_add_assign(temp.s1,psi);		\
-  _vector_i_add_assign(temp.s3,psi);
-
-#define _hop_z_m()					\
-  _vector_i_sub(psi,sm->s0,sm->s2);			\
-  _su3_inverse_multiply(chi,(*um),psi);			\
-  _complexcjg_times_vector(psi,ka3,chi);		\
-  _vector_add_assign(temp.s0, psi);		\
-  _vector_i_add_assign(temp.s2, psi);		\
-  _vector_i_add(psi,sm->s1,sm->s3);			\
-  _su3_inverse_multiply(chi,(*um),psi);			\
-  _complexcjg_times_vector(psi,ka3,chi);		\
-  _vector_add_assign(temp.s1, psi);		\
-  _vector_i_sub_assign(temp.s3, psi);
-
-#define _store_res()				\
-  _vector_assign(rn->s0, temp.s0);		\
-  _vector_assign(rn->s1, temp.s1);		\
-  _vector_assign(rn->s2, temp.s2);		\
-  _vector_assign(rn->s3, temp.s3);		\
-
 /* l output , k input*/
 /* for ieo=0, k resides on  odd sites and l on even sites */
 void Hopping_Matrix(int ieo, spinor * const l, spinor * const k){