Skip to content

Commit

Permalink
...EJB
Browse files Browse the repository at this point in the history
  • Loading branch information
@ebylaska
ebylaska committed Dec 30, 2023
1 parent 4655b3f commit 56ef97b
Show file tree
Hide file tree
Showing 5 changed files with 158 additions and 145 deletions.
27 changes: 14 additions & 13 deletions Nwpw/nwpwlib/C3dB/Cneb.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -844,6 +844,7 @@ void Cneb::hr_aSumSqr(const double alpha, double *psir, double *dn)
}
}
c3db::parall->Vector_SumAll(2, ispin*nfft3d, dn);
c3db::parall->Vector_SumAll(3, ispin*nfft3d, dn);
}

/*************************************
Expand Down Expand Up @@ -953,7 +954,7 @@ void Cneb::ggm_sym_Multiply(double *psi1, double *psi2, double *hml)
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi1+shift0,psi2+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand Down Expand Up @@ -1042,7 +1043,7 @@ void Cneb::ggw_sym_Multiply(double *psi1, double *psi2, double *hml)
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi1+shift0,psi2+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand Down Expand Up @@ -1136,7 +1137,7 @@ void Cneb::ggw_Multiply(double *psi1, double *psi2, double *hml)
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi1+shift0,psi2+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand Down Expand Up @@ -1214,7 +1215,7 @@ void Cneb::ffm_sym_Multiply(const int mb, double *psi1, double *psi2, double *hm
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi1+shift0,psi1+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand Down Expand Up @@ -1324,7 +1325,7 @@ void Cneb::ffm_Multiply(const int mb, double *psi1, double *psi2, double *hml)
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi1+shift0,psi1+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand Down Expand Up @@ -1440,7 +1441,7 @@ void Cneb::ffw3_sym_Multiply(const int mb, double *psi1, double *psi2,
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi1+shift0,psi1+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand All @@ -1458,7 +1459,7 @@ void Cneb::ffw3_sym_Multiply(const int mb, double *psi1, double *psi2,
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi2+shift0,psi1+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand All @@ -1476,7 +1477,7 @@ void Cneb::ffw3_sym_Multiply(const int mb, double *psi1, double *psi2,
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi2+shift0,psi2+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand Down Expand Up @@ -1615,7 +1616,7 @@ void Cneb::ffm4_sym_Multiply(const int mb, double *psi1, double *psi2,
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi1+shift0,psi1+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand All @@ -1633,7 +1634,7 @@ void Cneb::ffm4_sym_Multiply(const int mb, double *psi1, double *psi2,
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi2+shift0,psi1+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand All @@ -1651,7 +1652,7 @@ void Cneb::ffm4_sym_Multiply(const int mb, double *psi1, double *psi2,
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi1+shift0,psi2+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand All @@ -1669,7 +1670,7 @@ void Cneb::ffm4_sym_Multiply(const int mb, double *psi1, double *psi2,
{
auto shift0 = ms*neq[0]*npack1;
auto shift2 = ms*ishift2;
c1db::CMatrix_dgemm2c(c1db::parall, &mygdevice,
c1db::CMatrix_zgemm2c(c1db::parall, &mygdevice,
ne[ms],ne[ms],npack1_all,128,
psi2+shift0,psi2+shift0, ma[ms][taskid_i],ma[ms],ma1[ms],na[ms],
mat_tmp+shift2,mc[ms][taskid_i],mc[ms],nc[ms],
Expand Down Expand Up @@ -1842,7 +1843,7 @@ void Cneb::fwf_Multiply(const int mb, double *psi1, double *hml, double *alpha,
for (ms=ms1; ms<ms2; ++ms)
{
n = ne[ms];
c3db::mygdevice.NN_zgemm(n,n,npack1,alpha, psi1+2*shift1, npack1, hml+2*mshift1, n, beta, psi2+2*shift1,npack1);
c3db::mygdevice.NN_zgemm(npack1,n,n,alpha, psi1+2*shift1, npack1, hml+2*mshift1, n, beta, psi2+2*shift1,npack1);

shift1 += ne[0]*npack1;
mshift1 += ishift2;
Expand Down
Loading

0 comments on commit 56ef97b

Please sign in to comment.