Feb2021 GPU Package Update - GPU Package Files
This commit is contained in:
@ -40,22 +40,25 @@ __kernel void k_soft(const __global numtyp4 *restrict x_,
|
||||
atom_info(t_per_atom,ii,tid,offset);
|
||||
|
||||
__local numtyp sp_lj[4];
|
||||
int n_stride;
|
||||
local_allocate_store_pair();
|
||||
|
||||
sp_lj[0]=sp_lj_in[0];
|
||||
sp_lj[1]=sp_lj_in[1];
|
||||
sp_lj[2]=sp_lj_in[2];
|
||||
sp_lj[3]=sp_lj_in[3];
|
||||
|
||||
acctyp energy=(acctyp)0;
|
||||
acctyp4 f;
|
||||
f.x=(acctyp)0; f.y=(acctyp)0; f.z=(acctyp)0;
|
||||
acctyp virial[6];
|
||||
for (int i=0; i<6; i++)
|
||||
virial[i]=(acctyp)0;
|
||||
acctyp energy, virial[6];
|
||||
if (EVFLAG) {
|
||||
energy=(acctyp)0;
|
||||
for (int i=0; i<6; i++) virial[i]=(acctyp)0;
|
||||
}
|
||||
|
||||
if (ii<inum) {
|
||||
int nbor, nbor_end;
|
||||
int i, numj;
|
||||
__local int n_stride;
|
||||
nbor_info(dev_nbor,dev_packed,nbor_pitch,t_per_atom,ii,offset,i,numj,
|
||||
n_stride,nbor_end,nbor);
|
||||
|
||||
@ -91,11 +94,11 @@ __kernel void k_soft(const __global numtyp4 *restrict x_,
|
||||
f.y+=dely*force;
|
||||
f.z+=delz*force;
|
||||
|
||||
if (eflag>0) {
|
||||
if (EVFLAG && eflag) {
|
||||
numtyp e=coeff[mtype].x * ((numtyp)1.0+cos(arg));
|
||||
energy+=factor_lj*e;
|
||||
}
|
||||
if (vflag>0) {
|
||||
if (EVFLAG && vflag) {
|
||||
virial[0] += delx*delx*force;
|
||||
virial[1] += dely*dely*force;
|
||||
virial[2] += delz*delz*force;
|
||||
@ -106,9 +109,9 @@ __kernel void k_soft(const __global numtyp4 *restrict x_,
|
||||
}
|
||||
|
||||
} // for nbor
|
||||
store_answers(f,energy,virial,ii,inum,tid,t_per_atom,offset,eflag,vflag,
|
||||
ans,engv);
|
||||
} // if ii
|
||||
store_answers(f,energy,virial,ii,inum,tid,t_per_atom,offset,eflag,vflag,
|
||||
ans,engv);
|
||||
}
|
||||
|
||||
__kernel void k_soft_fast(const __global numtyp4 *restrict x_,
|
||||
@ -125,25 +128,28 @@ __kernel void k_soft_fast(const __global numtyp4 *restrict x_,
|
||||
|
||||
__local numtyp4 coeff[MAX_SHARED_TYPES*MAX_SHARED_TYPES];
|
||||
__local numtyp sp_lj[4];
|
||||
int n_stride;
|
||||
local_allocate_store_pair();
|
||||
|
||||
if (tid<4)
|
||||
sp_lj[tid]=sp_lj_in[tid];
|
||||
if (tid<MAX_SHARED_TYPES*MAX_SHARED_TYPES) {
|
||||
coeff[tid]=coeff_in[tid];
|
||||
}
|
||||
|
||||
acctyp energy=(acctyp)0;
|
||||
acctyp4 f;
|
||||
f.x=(acctyp)0; f.y=(acctyp)0; f.z=(acctyp)0;
|
||||
acctyp virial[6];
|
||||
for (int i=0; i<6; i++)
|
||||
virial[i]=(acctyp)0;
|
||||
acctyp energy, virial[6];
|
||||
if (EVFLAG) {
|
||||
energy=(acctyp)0;
|
||||
for (int i=0; i<6; i++) virial[i]=(acctyp)0;
|
||||
}
|
||||
|
||||
__syncthreads();
|
||||
|
||||
if (ii<inum) {
|
||||
int nbor, nbor_end;
|
||||
int i, numj;
|
||||
__local int n_stride;
|
||||
nbor_info(dev_nbor,dev_packed,nbor_pitch,t_per_atom,ii,offset,i,numj,
|
||||
n_stride,nbor_end,nbor);
|
||||
|
||||
@ -179,11 +185,11 @@ __kernel void k_soft_fast(const __global numtyp4 *restrict x_,
|
||||
f.y+=dely*force;
|
||||
f.z+=delz*force;
|
||||
|
||||
if (eflag>0) {
|
||||
if (EVFLAG && eflag) {
|
||||
numtyp e=coeff[mtype].x * ((numtyp)1.0+cos(arg));
|
||||
energy+=factor_lj*e;
|
||||
}
|
||||
if (vflag>0) {
|
||||
if (EVFLAG && vflag) {
|
||||
virial[0] += delx*delx*force;
|
||||
virial[1] += dely*dely*force;
|
||||
virial[2] += delz*delz*force;
|
||||
@ -194,8 +200,8 @@ __kernel void k_soft_fast(const __global numtyp4 *restrict x_,
|
||||
}
|
||||
|
||||
} // for nbor
|
||||
store_answers(f,energy,virial,ii,inum,tid,t_per_atom,offset,eflag,vflag,
|
||||
ans,engv);
|
||||
} // if ii
|
||||
store_answers(f,energy,virial,ii,inum,tid,t_per_atom,offset,eflag,vflag,
|
||||
ans,engv);
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user