ICODE-ADC/lammps
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Gradually cleaned up and removed redundancy in amoeba and hippo

Browse Source
This commit is contained in:
Trung Nguyen
2021-10-02 00:09:53 -05:00
parent f126f785a4
commit f4d3d3a2b5
7 changed files with 61 additions and 309 deletions
Download Patch File Download Diff File Expand all files Collapse all files

176
lib/gpu/lal_hippo.cpp
View File

@ -48,22 +48,20 @@ int HippoT::bytes_per_atom(const int max_nbors) const {
template <class numtyp, class acctyp>
int HippoT::init(const int ntypes, const int max_amtype, const int max_amclass,
const double *host_pdamp, const double *host_thole,
const double *host_dirdamp, const int *host_amtype2class,
const double *host_special_hal,
const double *host_special_repel,
const double *host_special_disp,
const double *host_special_mpole,
const double *host_special_polar_wscale,
const double *host_special_polar_piscale,
const double *host_special_polar_pscale,
const double *host_sizpr, const double *host_dmppr, const double *host_elepr,
const double *host_csix, const double *host_adisp,
const double *host_pcore, const double *host_palpha,
const int nlocal, const int nall, const int max_nbors,
const int maxspecial, const int maxspecial15,
const double cell_size, const double gpu_split, FILE *_screen,
const double polar_dscale, const double polar_uscale) {
const double *host_pdamp, const double *host_thole,
const double *host_dirdamp, const int *host_amtype2class,
const double *host_special_repel, const double *host_special_disp,
const double *host_special_mpole,
const double *host_special_polar_wscale,
const double *host_special_polar_piscale,
const double *host_special_polar_pscale,
const double *host_sizpr, const double *host_dmppr, const double *host_elepr,
const double *host_csix, const double *host_adisp,
const double *host_pcore, const double *host_palpha,
const int nlocal, const int nall, const int max_nbors,
const int maxspecial, const int maxspecial15,
const double cell_size, const double gpu_split, FILE *_screen,
const double polar_dscale, const double polar_uscale) {
int success;
success=this->init_atomic(nlocal,nall,max_nbors,maxspecial,maxspecial15,
cell_size,gpu_split,_screen,hippo,
@ -133,9 +131,9 @@ int HippoT::init(const int ntypes, const int max_amtype, const int max_amclass,
sp_nonpolar.alloc(5,*(this->ucl_device),UCL_READ_ONLY);
for (int i=0; i<5; i++) {
dview[i].x=host_special_hal[i];
dview[i].y=host_special_repel[i];
dview[i].z=host_special_disp[i];
dview[i].x=host_special_repel[i];
dview[i].y=host_special_disp[i];
dview[i].z=(numtyp)0;
dview[i].w=(numtyp)0;
}
ucl_copy(sp_nonpolar,dview,5,false);
@ -211,7 +209,7 @@ int** HippoT::compute_repulsion(const int ago, const int inum_full,
// to be able to turn on/off the udirect2b kernel (which comes before this)
// Once all the kernels are ready, precompute() is needed only once
// in the first kernel in a time step.
// We only need to cast uind and uinp from host to device here
// We only need to cast the necessary from host to device here
// if the neighbor lists are rebuilt and other per-atom arrays
// (x, type, amtype, amgroup, rpole) are ready on the device.
@ -240,7 +238,7 @@ int** HippoT::compute_repulsion(const int ago, const int inum_full,
_c3 = c3;
_c4 = c4;
_c5 = c5;
const int red_blocks=repulsion(eflag,vflag);
const int red_blocks=repulsion(this->_eflag,this->_vflag);
// only copy them back if this is the last kernel
// otherwise, commenting out these two lines to leave the answers
@ -316,32 +314,14 @@ int** HippoT::compute_dispersion_real(const int ago, const int inum_full,
const double cpu_time, bool &success,
const double aewald, const double off2_disp,
double *host_q, double *boxlo, double *prd) {
this->acc_timers();
int eflag, vflag;
if (eatom) eflag=2;
else if (eflag_in) eflag=1;
else eflag=0;
if (vatom) vflag=2;
else if (vflag_in) vflag=1;
else vflag=0;
#ifdef LAL_NO_BLOCK_REDUCE
if (eflag) eflag=2;
if (vflag) vflag=2;
#endif
this->set_kernel(eflag,vflag);
// reallocate per-atom arrays, transfer data from the host
// and build the neighbor lists if needed
// NOTE:
// For now we invoke precompute() again here,
// to be able to turn on/off the udirect2b kernel (which comes before this)
// Once all the kernels are ready, precompute() is needed only once
// in the first kernel in a time step.
// We only need to cast uind and uinp from host to device here
// if the neighbor lists are rebuilt and other per-atom arrays
// (x, type, amtype, amgroup, rpole) are ready on the device.
// We only need to cast necesary data arrays from host to device here
// because the neighbor lists are rebuilt and other per-atom arrays
// (x, type) are ready on the device.
int** firstneigh = nullptr;
firstneigh = this->precompute(ago, inum_full, nall, host_x, host_type,
@ -350,11 +330,11 @@ int** HippoT::compute_dispersion_real(const int ago, const int inum_full,
nspecial, special, nspecial15, special15,
eflag_in, vflag_in, eatom, vatom,
host_start, ilist, jnum, cpu_time,
success, host_q, boxlo, prd);
success, host_q, boxlo, prd);
this->_off2_disp = off2_disp;
this->_aewald = aewald;
const int red_blocks=dispersion_real(eflag,vflag);
const int red_blocks=dispersion_real(this->_eflag,this->_vflag);
// only copy them back if this is the last kernel
// otherwise, commenting out these two lines to leave the answers
@ -427,22 +407,6 @@ int** HippoT::compute_multipole_real(const int ago, const int inum_full,
const double aewald, const double felec,
const double off2_mpole, double *host_q,
double *boxlo, double *prd, void **tep_ptr) {
this->acc_timers();
int eflag, vflag;
if (eatom) eflag=2;
else if (eflag_in) eflag=1;
else eflag=0;
if (vatom) vflag=2;
else if (vflag_in) vflag=1;
else vflag=0;
#ifdef LAL_NO_BLOCK_REDUCE
if (eflag) eflag=2;
if (vflag) vflag=2;
#endif
this->set_kernel(eflag,vflag);
// reallocate per-atom arrays, transfer data from the host
// and build the neighbor lists if needed
// NOTE:
@ -474,7 +438,7 @@ int** HippoT::compute_multipole_real(const int ago, const int inum_full,
this->_off2_mpole = off2_mpole;
this->_felec = felec;
this->_aewald = aewald;
const int red_blocks=multipole_real(eflag,vflag);
const int red_blocks=multipole_real(this->_eflag,this->_vflag);
// leave the answers (forces, energies and virial) on the device,
// only copy them back in the last kernel (this one, or polar_real once done)
@ -486,13 +450,7 @@ int** HippoT::compute_multipole_real(const int ago, const int inum_full,
// copy tep from device to host
this->_tep.update_host(this->_max_tep_size*4,false);
/*
printf("GPU lib: tep size = %d: max tep size = %d\n", this->_tep.cols(), _max_tep_size);
for (int i = 0; i < 10; i++) {
numtyp4* p = (numtyp4*)(&this->_tep[4*i]);
printf("i = %d; tep = %f %f %f\n", i, p->x, p->y, p->z);
}
*/
return firstneigh; // nbor->host_jlist.begin()-host_start;
}
@ -558,22 +516,6 @@ int** HippoT::compute_udirect2b(const int ago, const int inum_full,
const double aewald, const double off2_polar,
double *host_q, double *boxlo, double *prd,
void** fieldp_ptr) {
this->acc_timers();
int eflag, vflag;
if (eatom) eflag=2;
else if (eflag_in) eflag=1;
else eflag=0;
if (vatom) vflag=2;
else if (vflag_in) vflag=1;
else vflag=0;
#ifdef LAL_NO_BLOCK_REDUCE
if (eflag) eflag=2;
if (vflag) vflag=2;
#endif
this->set_kernel(eflag,vflag);
// reallocate per-atom arrays, transfer data from the host
// and build the neighbor lists if needed
@ -596,19 +538,12 @@ int** HippoT::compute_udirect2b(const int ago, const int inum_full,
this->_off2_polar = off2_polar;
this->_aewald = aewald;
const int red_blocks=udirect2b(eflag,vflag);
const int red_blocks=udirect2b(this->_eflag,this->_vflag);
// copy field and fieldp from device to host (_fieldp store both arrays, one after another)
this->_fieldp.update_host(this->_max_fieldp_size*8,false);
/*
printf("GPU lib: _fieldp size = %d: max fieldp size = %d\n",
this->_fieldp.cols(), _max_fieldp_size);
for (int i = 0; i < 10; i++) {
numtyp4* p = (numtyp4*)(&this->_fieldp[4*i]);
printf("i = %d; field = %f %f %f\n", i, p->x, p->y, p->z);
}
*/
return firstneigh; //nbor->host_jlist.begin()-host_start;
}
@ -673,22 +608,6 @@ int** HippoT::compute_umutual2b(const int ago, const int inum_full,
const double aewald, const double off2_polar,
double *host_q, double *boxlo, double *prd,
void** fieldp_ptr) {
this->acc_timers();
int eflag, vflag;
if (eatom) eflag=2;
else if (eflag_in) eflag=1;
else eflag=0;
if (vatom) vflag=2;
else if (vflag_in) vflag=1;
else vflag=0;
#ifdef LAL_NO_BLOCK_REDUCE
if (eflag) eflag=2;
if (vflag) vflag=2;
#endif
this->set_kernel(eflag,vflag);
// reallocate per-atom arrays, transfer extra data from the host
// and build the neighbor lists if needed
@ -711,19 +630,12 @@ int** HippoT::compute_umutual2b(const int ago, const int inum_full,
this->_off2_polar = off2_polar;
this->_aewald = aewald;
const int red_blocks=umutual2b(eflag,vflag);
const int red_blocks=umutual2b(this->_eflag,this->_vflag);
// copy field and fieldp from device to host (_fieldp store both arrays, one after another)
this->_fieldp.update_host(this->_max_fieldp_size*8,false);
/*
printf("GPU lib: _fieldp size = %d: max fieldp size = %d\n",
this->_fieldp.cols(), _max_fieldp_size);
for (int i = 0; i < 10; i++) {
numtyp4* p = (numtyp4*)(&this->_fieldp[4*i]);
printf("i = %d; field = %f %f %f\n", i, p->x, p->y, p->z);
}
*/
return firstneigh; //nbor->host_jlist.begin()-host_start;
}
@ -786,29 +698,11 @@ int** HippoT::compute_polar_real(const int ago, const int inum_full,
const double aewald, const double felec,
const double off2_polar, double *host_q,
double *boxlo, double *prd, void **tep_ptr) {
this->acc_timers();
int eflag, vflag;
if (eatom) eflag=2;
else if (eflag_in) eflag=1;
else eflag=0;
if (vatom) vflag=2;
else if (vflag_in) vflag=1;
else vflag=0;
#ifdef LAL_NO_BLOCK_REDUCE
if (eflag) eflag=2;
if (vflag) vflag=2;
#endif
this->set_kernel(eflag,vflag);
// reallocate per-atom arrays, transfer data from the host
// and build the neighbor lists if needed
// NOTE:
// For now we invoke precompute() again here,
// to be able to turn on/off the udirect2b kernel (which comes before this)
// Once all the kernels are ready, precompute() is needed only once
// in the first kernel in a time step.
// We only need to cast uind and uinp from host to device here
// if the neighbor lists are rebuilt and other per-atom arrays
// (x, type, amtype, amgroup, rpole) are ready on the device.
@ -833,7 +727,7 @@ int** HippoT::compute_polar_real(const int ago, const int inum_full,
this->_off2_polar = off2_polar;
this->_felec = felec;
this->_aewald = aewald;
const int red_blocks=polar_real(eflag,vflag);
const int red_blocks=polar_real(this->_eflag,this->_vflag);
// only copy answers (forces, energies and virial) back from the device
// in the last kernel (which is polar_real here)
@ -845,13 +739,7 @@ int** HippoT::compute_polar_real(const int ago, const int inum_full,
// copy tep from device to host
this->_tep.update_host(this->_max_tep_size*4,false);
/*
printf("GPU lib: tep size = %d: max tep size = %d\n", this->_tep.cols(), _max_tep_size);
for (int i = 0; i < 10; i++) {
numtyp4* p = (numtyp4*)(&this->_tep[4*i]);
printf("i = %d; tep = %f %f %f\n", i, p->x, p->y, p->z);
}
*/
return firstneigh; // nbor->host_jlist.begin()-host_start;
}