mirror of
https://github.com/TREX-CoE/qmckl.git
synced 2024-08-24 22:21:46 +02:00
Fixed efence compilation
This commit is contained in:
commit
d1e88ad475
40
configure.ac
40
configure.ac
@ -330,6 +330,26 @@ AS_IF([test "$HAVE_CUBLAS_OFFLOAD" = "yes"], [
|
||||
esac
|
||||
])
|
||||
|
||||
AC_ARG_ENABLE(malloc-trace, [AS_HELP_STRING([--enable-malloc-trace],[use debug malloc/free])], ok=$enableval, ok=no)
|
||||
if test "$ok" = "yes"; then
|
||||
AC_DEFINE(MALLOC_TRACE,"malloc_trace.dat",[Define to use debugging malloc/free])
|
||||
ARGS="${ARGS} malloc-trace"
|
||||
fi
|
||||
|
||||
AC_ARG_ENABLE(prof, [AS_HELP_STRING([--enable-prof],[compile for profiling])], ok=$enableval, ok=no)
|
||||
if test "$ok" = "yes"; then
|
||||
CFLAGS="${CFLAGS} -pg"
|
||||
AC_DEFINE(ENABLE_PROF,1,[Define when using the profiler tool])
|
||||
ARGS="${ARGS} prof"
|
||||
fi
|
||||
|
||||
AC_ARG_WITH(efence, [AS_HELP_STRING([--with-efence],[use ElectricFence library])], ok=$withval, ok=no)
|
||||
if test "$ok" = "yes"; then
|
||||
AC_CHECK_LIB([efence], [malloc])
|
||||
ARGS="${ARGS} efence"
|
||||
fi
|
||||
|
||||
|
||||
|
||||
##
|
||||
|
||||
@ -356,26 +376,6 @@ if test "$ok" = "yes"; then
|
||||
ARGS="${ARGS} debug"
|
||||
fi
|
||||
|
||||
AC_ARG_ENABLE(malloc-trace, [AS_HELP_STRING([--enable-malloc-trace],[use debug malloc/free])], ok=$enableval, ok=no)
|
||||
if test "$ok" = "yes"; then
|
||||
AC_DEFINE(MALLOC_TRACE,"malloc_trace.dat",[Define to use debugging malloc/free])
|
||||
ARGS="${ARGS} malloc-trace"
|
||||
fi
|
||||
|
||||
AC_ARG_ENABLE(prof, [AS_HELP_STRING([--enable-prof],[compile for profiling])], ok=$enableval, ok=no)
|
||||
if test "$ok" = "yes"; then
|
||||
CFLAGS="${CFLAGS} -pg"
|
||||
AC_DEFINE(ENABLE_PROF,1,[Define when using the profiler tool])
|
||||
ARGS="${ARGS} prof"
|
||||
fi
|
||||
|
||||
AC_ARG_WITH(efence, [AS_HELP_STRING([--with-efence],[use ElectricFence library])], ok=$withval, ok=no)
|
||||
if test "$ok" = "yes"; then
|
||||
AC_CHECK_LIB(efence, malloc)
|
||||
ARGS="${ARGS} efence"
|
||||
fi
|
||||
|
||||
|
||||
# Checks for header files.
|
||||
|
||||
## qmckl
|
||||
|
@ -1478,11 +1478,11 @@ end function qmckl_compute_asymp_jasb_f
|
||||
|
||||
#+begin_src c :comments org :tangle (eval c) :noweb yes
|
||||
qmckl_exit_code qmckl_compute_asymp_jasb (
|
||||
const qmckl_context context,
|
||||
const int64_t bord_num,
|
||||
const double* bord_vector,
|
||||
const double rescale_factor_kappa_ee,
|
||||
double* const asymp_jasb ) {
|
||||
const qmckl_context context,
|
||||
const int64_t bord_num,
|
||||
const double* bord_vector,
|
||||
const double rescale_factor_kappa_ee,
|
||||
double* const asymp_jasb ) {
|
||||
|
||||
if (context == QMCKL_NULL_CONTEXT){
|
||||
return QMCKL_INVALID_CONTEXT;
|
||||
|
@ -256,10 +256,8 @@ bool qmckl_mo_basis_provided(const qmckl_context context) {
|
||||
|
||||
#+end_src
|
||||
|
||||
|
||||
*** Fortran interfaces
|
||||
|
||||
#+begin_src f90 :tangle (eval fh_func) :comments org
|
||||
interface
|
||||
integer(c_int32_t) function qmckl_get_mo_basis_mo_num (context, &
|
||||
mo_num) bind(C)
|
||||
@ -665,26 +663,41 @@ integer function qmckl_compute_mo_basis_mo_vgl_doc_f(context, &
|
||||
integer*8 :: i,j,k
|
||||
double precision :: c1, c2, c3, c4, c5
|
||||
|
||||
do j=1,point_num
|
||||
mo_vgl(:,:,j) = 0.d0
|
||||
do k=1,ao_num
|
||||
if (ao_vgl(k,1,j) /= 0.d0) then
|
||||
c1 = ao_vgl(k,1,j)
|
||||
c2 = ao_vgl(k,2,j)
|
||||
c3 = ao_vgl(k,3,j)
|
||||
c4 = ao_vgl(k,4,j)
|
||||
c5 = ao_vgl(k,5,j)
|
||||
do i=1,mo_num
|
||||
mo_vgl(i,1,j) = mo_vgl(i,1,j) + coef_normalized_t(i,k) * c1
|
||||
mo_vgl(i,2,j) = mo_vgl(i,2,j) + coef_normalized_t(i,k) * c2
|
||||
mo_vgl(i,3,j) = mo_vgl(i,3,j) + coef_normalized_t(i,k) * c3
|
||||
mo_vgl(i,4,j) = mo_vgl(i,4,j) + coef_normalized_t(i,k) * c4
|
||||
mo_vgl(i,5,j) = mo_vgl(i,5,j) + coef_normalized_t(i,k) * c5
|
||||
end do
|
||||
end if
|
||||
end do
|
||||
end do
|
||||
integer*8 :: LDA, LDB, LDC
|
||||
|
||||
info = QMCKL_SUCCESS
|
||||
if (.True.) then ! fast algorithm
|
||||
do j=1,point_num
|
||||
mo_vgl(:,:,j) = 0.d0
|
||||
do k=1,ao_num
|
||||
if (ao_vgl(k,1,j) /= 0.d0) then
|
||||
c1 = ao_vgl(k,1,j)
|
||||
c2 = ao_vgl(k,2,j)
|
||||
c3 = ao_vgl(k,3,j)
|
||||
c4 = ao_vgl(k,4,j)
|
||||
c5 = ao_vgl(k,5,j)
|
||||
do i=1,mo_num
|
||||
mo_vgl(i,1,j) = mo_vgl(i,1,j) + coef_normalized_t(i,k) * c1
|
||||
mo_vgl(i,2,j) = mo_vgl(i,2,j) + coef_normalized_t(i,k) * c2
|
||||
mo_vgl(i,3,j) = mo_vgl(i,3,j) + coef_normalized_t(i,k) * c3
|
||||
mo_vgl(i,4,j) = mo_vgl(i,4,j) + coef_normalized_t(i,k) * c4
|
||||
mo_vgl(i,5,j) = mo_vgl(i,5,j) + coef_normalized_t(i,k) * c5
|
||||
end do
|
||||
end if
|
||||
end do
|
||||
end do
|
||||
|
||||
else ! dgemm
|
||||
|
||||
LDA = size(coef_normalized_t,1)
|
||||
LDB = size(ao_vgl,1)
|
||||
LDC = size(mo_vgl,1)
|
||||
|
||||
info = qmckl_dgemm(context,'N', 'N', mo_num, point_num*5_8, ao_num*1_8, 1.d0, &
|
||||
coef_normalized_t, LDA, ao_vgl, LDB, &
|
||||
0.d0, mo_vgl, LDC)
|
||||
|
||||
end if
|
||||
|
||||
end function qmckl_compute_mo_basis_mo_vgl_doc_f
|
||||
#+end_src
|
||||
@ -789,16 +802,19 @@ qmckl_compute_mo_basis_mo_vgl_hpc (const qmckl_context context,
|
||||
const double* restrict ao_vgl,
|
||||
double* restrict const mo_vgl )
|
||||
{
|
||||
assert (context != QMCKL_NULL_CONTEXT);
|
||||
|
||||
#ifdef HAVE_OPENMP
|
||||
#pragma omp parallel for
|
||||
#endif
|
||||
for (int64_t ipoint=0 ; ipoint < point_num ; ++ipoint) {
|
||||
double* restrict const vgl1 = &(mo_vgl[ipoint*5*mo_num]);
|
||||
const double* restrict avgl1 = &(ao_vgl[ipoint*5*ao_num]);
|
||||
double* restrict const vgl2 = vgl1 + mo_num;
|
||||
double* restrict const vgl3 = vgl1 + (mo_num << 1);
|
||||
double* restrict const vgl4 = vgl1 + (mo_num << 1) + mo_num;
|
||||
double* restrict const vgl5 = vgl1 + (mo_num << 2);
|
||||
|
||||
const double* restrict avgl1 = &(ao_vgl[ipoint*5*ao_num]);
|
||||
const double* restrict avgl2 = avgl1 + ao_num;
|
||||
const double* restrict avgl3 = avgl1 + (ao_num << 1);
|
||||
const double* restrict avgl4 = avgl1 + (ao_num << 1) + ao_num;
|
||||
@ -832,6 +848,7 @@ qmckl_compute_mo_basis_mo_vgl_hpc (const qmckl_context context,
|
||||
}
|
||||
|
||||
int64_t n;
|
||||
|
||||
for (n=0 ; n < nidx-4 ; n+=4) {
|
||||
const double* restrict ck1 = coef_normalized_t + idx[n ]*mo_num;
|
||||
const double* restrict ck2 = coef_normalized_t + idx[n+1]*mo_num;
|
||||
@ -875,8 +892,7 @@ qmckl_compute_mo_basis_mo_vgl_hpc (const qmckl_context context,
|
||||
}
|
||||
}
|
||||
|
||||
int64_t n0 = nidx-4;
|
||||
n0 = n0 < 0 ? 0 : n0;
|
||||
const int64_t n0 = n < 0 ? 0 : n;
|
||||
for (int64_t m=n0 ; m < nidx ; m+=1) {
|
||||
const double* restrict ck = coef_normalized_t + idx[m]*mo_num;
|
||||
const double a1 = av1[m];
|
||||
|
@ -250,19 +250,19 @@ qmckl_exit_code qmckl_set_numprec_range(const qmckl_context context, const int r
|
||||
# Fortran interface
|
||||
#+begin_src f90 :tangle (eval fh_func)
|
||||
interface
|
||||
integer (qmckl_exit_code) function qmckl_numprec_set_range(context, range) bind(C)
|
||||
integer (qmckl_exit_code) function qmckl_set_numprec_range(context, range) bind(C)
|
||||
use, intrinsic :: iso_c_binding
|
||||
import
|
||||
integer (qmckl_context), intent(in), value :: context
|
||||
integer (c_int32_t), intent(in), value :: range
|
||||
end function qmckl_numprec_set_range
|
||||
end function qmckl_set_numprec_range
|
||||
end interface
|
||||
#+end_src
|
||||
|
||||
~qmckl_get_numprec_range~ returns the value of the numerical range in the context.
|
||||
|
||||
#+begin_src c :comments org :tangle (eval h_func) :exports none
|
||||
int32_t qmckl_get_numprec_get_range(const qmckl_context context);
|
||||
int32_t qmckl_get_numprec_range(const qmckl_context context);
|
||||
#+end_src
|
||||
|
||||
# Source
|
||||
|
Loading…
Reference in New Issue
Block a user