Skip to content

GitLab

Commit c2461875 authored by Peter Eastman's avatar Peter Eastman
Browse files

Continuing to convert AmoebaGeneralizedKirkwoodForce to new CUDA platform

parent 6c0082a5
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 234 additions and 155 deletions
plugins/amoeba/platforms/cuda2/src/AmoebaCudaKernels.cpp
View file @ c2461875
......@@ -1321,6 +1321,7 @@ double CudaCalcAmoebaMultipoleForceKernel::execute(ContextImpl& context, bool in
int numForceThreadBlocks = nb.getNumForceThreadBlocks();
int forceThreadBlockSize = nb.getForceThreadBlockSize();
int elementSize = (cu.getUseDoublePrecision() ? sizeof(double) : sizeof(float));
void* npt = NULL;
if (pmeGrid == NULL) {
// Compute induced dipoles.
......@@ -1355,11 +1356,40 @@ double CudaCalcAmoebaMultipoleForceKernel::execute(ContextImpl& context, bool in
for (int i = 0; i < maxInducedIterations; i++) {
cu.clearBuffer(*inducedField);
cu.clearBuffer(*inducedFieldPolar);
void* computeInducedFieldArgs[] = {&inducedField->getDevicePointer(), &inducedFieldPolar->getDevicePointer(), &cu.getPosq().getDevicePointer(),
&inducedDipole->getDevicePointer(), &inducedDipolePolar->getDevicePointer(), &startTileIndex, &numTileIndices,
&dampingAndThole->getDevicePointer()};
cu.executeKernel(computeInducedFieldKernel, computeInducedFieldArgs, numForceThreadBlocks*forceThreadBlockSize, forceThreadBlockSize);
void* updateInducedFieldArgs[] = {&field->getDevicePointer(), &fieldPolar->getDevicePointer(), &inducedField->getDevicePointer(),
if (gkKernel == NULL) {
void* computeInducedFieldArgs[] = {&inducedField->getDevicePointer(), &inducedFieldPolar->getDevicePointer(), &cu.getPosq().getDevicePointer(),
&inducedDipole->getDevicePointer(), &inducedDipolePolar->getDevicePointer(), &startTileIndex, &numTileIndices,
&dampingAndThole->getDevicePointer()};
cu.executeKernel(computeInducedFieldKernel, computeInducedFieldArgs, numForceThreadBlocks*forceThreadBlockSize, forceThreadBlockSize);
}
else {
cu.clearBuffer(*gkKernel->getInducedField());
cu.clearBuffer(*gkKernel->getInducedFieldPolar());
vector<float> d, dp;
gkKernel->getInducedDipoles()->download(d);
gkKernel->getInducedDipolesPolar()->download(dp);
printf("dipole\n");
for (int i = 0; i < cu.getNumAtoms(); i++)
printf("%d %g %g %g, %g %g %g\n", i, d[3*i], d[3*i+1], d[3*i+2], dp[3*i], dp[3*i+1], dp[3*i+2]);
void* computeInducedFieldArgs[] = {&inducedField->getDevicePointer(), &inducedFieldPolar->getDevicePointer(), &cu.getPosq().getDevicePointer(),
&inducedDipole->getDevicePointer(), &inducedDipolePolar->getDevicePointer(), &startTileIndex, &numTileIndices,
&gkKernel->getInducedField()->getDevicePointer(), &gkKernel->getInducedFieldPolar()->getDevicePointer(),
&gkKernel->getInducedDipoles()->getDevicePointer(), &gkKernel->getInducedDipolesPolar()->getDevicePointer(),
&gkKernel->getBornRadii()->getDevicePointer(), &dampingAndThole->getDevicePointer()};
cu.executeKernel(computeInducedFieldKernel, computeInducedFieldArgs, numForceThreadBlocks*forceThreadBlockSize, forceThreadBlockSize);
vector<long long> f, fp;
gkKernel->getInducedField()->download(f);
gkKernel->getInducedFieldPolar()->download(fp);
printf("field\n");
for (int i = 0; i < cu.getNumAtoms(); i++)
printf("%d %g %g %g, %g %g %g\n", i, f[i]/(double) 0xFFFFFFFF, f[i+cu.getPaddedNumAtoms()]/(double) 0xFFFFFFFF, f[i+2*cu.getPaddedNumAtoms()]/(double) 0xFFFFFFFF, fp[i]/(double) 0xFFFFFFFF, fp[i+cu.getPaddedNumAtoms()]/(double) 0xFFFFFFFF, fp[i+2*cu.getPaddedNumAtoms()]/(double) 0xFFFFFFFF);
void* updateInducedGkFieldArgs[] = {&field->getDevicePointer(), &fieldPolar->getDevicePointer(),
&gkKernel->getField()->getDevicePointer(), &gkKernel->getFieldPolar()->getDevicePointer(), &gkKernel->getInducedField()->getDevicePointer(),
&gkKernel->getInducedFieldPolar()->getDevicePointer(), &gkKernel->getInducedDipoles()->getDevicePointer(), &gkKernel->getInducedDipolesPolar()->getDevicePointer(),
&polarizability->getDevicePointer(), &inducedDipoleErrors->getDevicePointer()};
cu.executeKernel(updateInducedFieldKernel, updateInducedGkFieldArgs, cu.getNumThreadBlocks()*cu.ThreadBlockSize, cu.ThreadBlockSize, cu.ThreadBlockSize*elementSize*2);
}
void* updateInducedFieldArgs[] = {&field->getDevicePointer(), &fieldPolar->getDevicePointer(), &npt, &npt, &inducedField->getDevicePointer(),
&inducedFieldPolar->getDevicePointer(), &inducedDipole->getDevicePointer(), &inducedDipolePolar->getDevicePointer(),
&polarizability->getDevicePointer(), &inducedDipoleErrors->getDevicePointer()};
cu.executeKernel(updateInducedFieldKernel, updateInducedFieldArgs, cu.getNumThreadBlocks()*cu.ThreadBlockSize, cu.ThreadBlockSize, cu.ThreadBlockSize*elementSize*2);
......@@ -1487,7 +1517,7 @@ double CudaCalcAmoebaMultipoleForceKernel::execute(ContextImpl& context, bool in
void* pmeRecordInducedFieldDipolesArgs[] = {&pmePhid->getDevicePointer(), &pmePhip->getDevicePointer(),
&inducedField->getDevicePointer(), &inducedFieldPolar->getDevicePointer(), cu.getInvPeriodicBoxSizePointer()};
cu.executeKernel(pmeRecordInducedFieldDipolesKernel, pmeRecordInducedFieldDipolesArgs, cu.getNumAtoms());
void* updateInducedFieldArgs[] = {&field->getDevicePointer(), &fieldPolar->getDevicePointer(), &inducedField->getDevicePointer(),
void* updateInducedFieldArgs[] = {&field->getDevicePointer(), &fieldPolar->getDevicePointer(), &npt, &npt, &inducedField->getDevicePointer(),
&inducedFieldPolar->getDevicePointer(), &inducedDipole->getDevicePointer(), &inducedDipolePolar->getDevicePointer(),
&polarizability->getDevicePointer(), &inducedDipoleErrors->getDevicePointer()};
cu.executeKernel(updateInducedFieldKernel, updateInducedFieldArgs, cu.getNumThreadBlocks()*cu.ThreadBlockSize, cu.ThreadBlockSize, cu.ThreadBlockSize*elementSize*2);
......@@ -1704,8 +1734,8 @@ private:
};
CudaCalcAmoebaGeneralizedKirkwoodForceKernel::CudaCalcAmoebaGeneralizedKirkwoodForceKernel(std::string name, const Platform& platform, CudaContext& cu, System& system) :
CalcAmoebaGeneralizedKirkwoodForceKernel(name, platform), cu(cu), system(system), params(NULL), bornRadii(NULL), field(NULL), inducedDipoleS(NULL),
inducedDipolePolarS(NULL), bornSum(NULL), bornForce(NULL) {
CalcAmoebaGeneralizedKirkwoodForceKernel(name, platform), cu(cu), system(system), params(NULL), bornRadii(NULL), field(NULL), fieldPolar(NULL),
inducedField(NULL), inducedFieldPolar(NULL), inducedDipoleS(NULL), inducedDipolePolarS(NULL), bornSum(NULL), bornForce(NULL) {
}
CudaCalcAmoebaGeneralizedKirkwoodForceKernel::~CudaCalcAmoebaGeneralizedKirkwoodForceKernel() {
......@@ -1716,6 +1746,12 @@ CudaCalcAmoebaGeneralizedKirkwoodForceKernel::~CudaCalcAmoebaGeneralizedKirkwood
delete bornRadii;
if (field != NULL)
delete field;
if (fieldPolar != NULL)
delete fieldPolar;
if (inducedField != NULL)
delete inducedField;
if (inducedFieldPolar != NULL)
delete inducedFieldPolar;
if (inducedDipoleS != NULL)
delete inducedDipoleS;
if (inducedDipolePolarS != NULL)
......@@ -1741,11 +1777,17 @@ void CudaCalcAmoebaGeneralizedKirkwoodForceKernel::initialize(const System& syst
params = CudaArray::create<float2>(cu, paddedNumAtoms, "amoebaGkParams");
bornRadii = new CudaArray(cu, paddedNumAtoms, elementSize, "bornRadii");
field = new CudaArray(cu, 3*paddedNumAtoms, sizeof(long long), "gkField");
fieldPolar = new CudaArray(cu, 3*paddedNumAtoms, sizeof(long long), "gkFieldPolar");
bornSum = CudaArray::create<long long>(cu, paddedNumAtoms, "bornSum");
bornForce = CudaArray::create<long long>(cu, paddedNumAtoms, "bornForce");
inducedDipoleS = new CudaArray(cu, 3*paddedNumAtoms, elementSize, "inducedDipoleS");
inducedDipolePolarS = new CudaArray(cu, 3*paddedNumAtoms, elementSize, "inducedDipolePolarS");
if (multipoles->getPolarizationType() == AmoebaMultipoleForce::Mutual) {
inducedField = new CudaArray(cu, 3*paddedNumAtoms, sizeof(long long), "gkInducedField");
inducedFieldPolar = new CudaArray(cu, 3*paddedNumAtoms, sizeof(long long), "gkInducedFieldPolar");
}
cu.addAutoclearBuffer(*field);
cu.addAutoclearBuffer(*fieldPolar);
cu.addAutoclearBuffer(*bornSum);
cu.addAutoclearBuffer(*bornForce);
vector<float2> paramsVector(paddedNumAtoms);
......
plugins/amoeba/platforms/cuda2/src/AmoebaCudaKernels.h
View file @ c2461875
......@@ -470,6 +470,15 @@ public:
CudaArray* getField() {
return field;
}
CudaArray* getFieldPolar() {
return fieldPolar;
}
CudaArray* getInducedField() {
return inducedField;
}
CudaArray* getInducedFieldPolar() {
return inducedFieldPolar;
}
CudaArray* getInducedDipoles() {
return inducedDipoleS;
}
......@@ -485,6 +494,9 @@ private:
CudaArray* bornRadii;
CudaArray* bornForce;
CudaArray* field;
CudaArray* fieldPolar;
CudaArray* inducedField;
CudaArray* inducedFieldPolar;
CudaArray* inducedDipoleS;
CudaArray* inducedDipolePolarS;
CUfunction computeBornSumKernel, reduceBornSumKernel, gkForceKernel, chainRuleKernel, ediffKernel;
......
plugins/amoeba/platforms/cuda2/src/kernels/gkEDiffPairForce.cu
View file @ c2461875
#define APPLY_SCALE
#if defined F1
__device__ void computeOneEDiffInteractionF1(AtomData4& atom1, volatile AtomData4& atom2, float dScale, float pScale, real& outputEnergy, real3& outputForce) {
#elif defined T1
......@@ -77,7 +75,6 @@ __device__ void computeOneEDiffInteractionT3(AtomData4& atom1, volatile AtomData
real scale3i = 3*scale3*uscale*rr3*rr2;
real scale5i = 3*scale5*uscale*rr3*rr2;
#ifdef APPLY_SCALE
real dsc3 = scale3*dScale*rr3;
real dsc5 = 3*scale5*dScale*rr3*rr2;
real dsc7 = 15*scale7*dScale*rr3*rr3*rr1;
......@@ -85,11 +82,6 @@ __device__ void computeOneEDiffInteractionT3(AtomData4& atom1, volatile AtomData
real psc3 = scale3*pScale*rr3;
real psc5 = 3*scale5*pScale*rr3*rr2;
real psc7 = 15*scale7*pScale*rr3*rr3*rr1;
#else
real psc3 = scale3*rr3;
real psc5 = 3*scale5*rr3*rr2;
real psc7 = 15*scale7*rr3*rr3*rr1;
#endif
#ifdef T1
real dixr1 = atom1.dipole.y*zr - atom1.dipole.z*yr;
......@@ -167,15 +159,9 @@ __device__ void computeOneEDiffInteractionT3(AtomData4& atom1, volatile AtomData
qiukp2 -= atom1.quadrupoleXY*atom2.inducedDipolePolar.x + atom1.quadrupoleYY*atom2.inducedDipolePolar.y + atom1.quadrupoleYZ*atom2.inducedDipolePolar.z;
qiukp3 -= atom1.quadrupoleXZ*atom2.inducedDipolePolar.x + atom1.quadrupoleYZ*atom2.inducedDipolePolar.y + atom1.quadrupoleZZ*atom2.inducedDipolePolar.z;
#ifdef APPLY_SCALE
ftm2i1 -= psc5*qiuk1 + dsc5*qiukp1;
ftm2i2 -= psc5*qiuk2 + dsc5*qiukp2;
ftm2i3 -= psc5*qiuk3 + dsc5*qiukp3;
#else
ftm2i1 -= psc5*(qiuk1 + qiukp1);
ftm2i2 -= psc5*(qiuk2 + qiukp2);
ftm2i3 -= psc5*(qiuk3 + qiukp3);
#endif
#endif
#endif
......@@ -197,15 +183,9 @@ __device__ void computeOneEDiffInteractionT3(AtomData4& atom1, volatile AtomData
qkuip2 -= atom2.quadrupoleXY*atom1.inducedDipolePolar.x + atom2.quadrupoleYY*atom1.inducedDipolePolar.y + atom2.quadrupoleYZ*atom1.inducedDipolePolar.z;
qkuip3 -= atom2.quadrupoleXZ*atom1.inducedDipolePolar.x + atom2.quadrupoleYZ*atom1.inducedDipolePolar.y + atom2.quadrupoleZZ*atom1.inducedDipolePolar.z;
#ifdef APPLY_SCALE
ftm2i1 += psc5*qkui1 + dsc5*qkuip1;
ftm2i2 += psc5*qkui2 + dsc5*qkuip2;
ftm2i3 += psc5*qkui3 + dsc5*qkuip3;
#else
ftm2i1 += psc5*(qkui1 + qkuip1);
ftm2i2 += psc5*(qkui2 + qkuip2);
ftm2i3 += psc5*(qkui3 + qkuip3);
#endif
#endif
#endif
......@@ -326,44 +306,24 @@ __device__ void computeOneEDiffInteractionT3(AtomData4& atom1, volatile AtomData
real glip3 = scip3Y*sc6 - scip4Y*sc5;
#ifdef F1
#ifdef APPLY_SCALE
ftm2i1 -= 0.5f*((gli1*pScale + glip1*dScale)*ddsc3_1 + (gli2*pScale + glip2*dScale)*ddsc5_1 + (gli3*pScale+glip3*dScale)*ddsc7_1);
ftm2i2 -= 0.5f*((gli1*pScale + glip1*dScale)*ddsc3_2 + (gli2*pScale + glip2*dScale)*ddsc5_2 + (gli3*pScale+glip3*dScale)*ddsc7_2);
ftm2i3 -= 0.5f*((gli1*pScale + glip1*dScale)*ddsc3_3 + (gli2*pScale + glip2*dScale)*ddsc5_3 + (gli3*pScale+glip3*dScale)*ddsc7_3);
#else
ftm2i1 -= 0.5f*((gli1 + glip1)*ddsc3_1 + (gli2 + glip2)*ddsc5_1 + (gli3 + glip3)*ddsc7_1);
ftm2i2 -= 0.5f*((gli1 + glip1)*ddsc3_2 + (gli2 + glip2)*ddsc5_2 + (gli3 + glip3)*ddsc7_2);
ftm2i3 -= 0.5f*((gli1 + glip1)*ddsc3_3 + (gli2 + glip2)*ddsc5_3 + (gli3 + glip3)*ddsc7_3);
#endif
outputEnergy = gli1*psc3 + gli2*psc5 + gli3*psc7;
#endif
#ifdef APPLY_SCALE
gfi1 += 1.5f*(gli1*psc3 + glip1*dsc3);
gfi1 += 2.5f*(gli2*psc5 + glip2*dsc5);
gfi1 += 3.5f*(gli3*psc7 + glip3*dsc7);
#else
gfi1 += 1.5f*psc3*(gli1 + glip1);
gfi1 += 2.5f*psc5*(gli2 + glip2);
gfi1 += 3.5f*psc7*(gli3 + glip3);
#endif
gfi1 *= rr2;
gfi1 += 0.5f*scip2*scale3i;
#if defined F1 || defined T1
#ifdef APPLY_SCALE
real gfi5 = (sci4Y*psc7+scip4Y*dsc7);
#else
real gfi5 = psc7*(sci4Y + scip4Y);
#endif
#endif
#if defined F1 || defined T3
#ifdef APPLY_SCALE
real gfi6 = -(sci3Y*psc7+scip3Y*dsc7);
#else
real gfi6 = -psc7*(sci3Y + scip3Y);
#endif
#endif
#ifdef F1
......@@ -371,66 +331,35 @@ __device__ void computeOneEDiffInteractionT3(AtomData4& atom1, volatile AtomData
real diff0 = atom1.inducedDipoleS.x - atom1.inducedDipole.x;
real diff1 = atom1.inducedDipolePolarS.x - atom1.inducedDipolePolar.x;
#ifdef APPLY_SCALE
ftm2i1 += 0.5f*(-atom2.q*(diff0*psc3 + diff1*dsc3) + sc4*(diff0*psc5 + diff1*dsc5) - sc6*(diff0*psc7 + diff1*dsc7));
#else
ftm2i1 += 0.5f*(-atom2.q*psc3*(diff0 + diff1) + sc4*psc5*(diff0 + diff1) - sc6*psc7*(diff0 + diff1));
#endif
diff0 = atom2.inducedDipoleS.x - atom2.inducedDipole.x;
diff1 = atom2.inducedDipolePolarS.x - atom2.inducedDipolePolar.x;
#ifdef APPLY_SCALE
ftm2i1 += 0.5f*(atom1.q*(diff0*psc3 + diff1*dsc3) + sc3*(diff0*psc5 + diff1*dsc5) + sc5*(diff0*psc7 + diff1*dsc7));
ftm2i1 += 0.5f*(sci4Y*psc5+scip4Y*dsc5)*atom1.dipole.x + 0.5f*(sci3Y*psc5+scip3Y*dsc5)*atom2.dipole.x + gfi5*qir1 + gfi6*qkr1;
#else
ftm2i1 += 0.5f*(atom1.q*psc3*(diff0 + diff1) + sc3*psc5*(diff0 + diff1) + sc5*psc7*(diff0 + diff1));
ftm2i1 += 0.5f*psc5*(sci4Y + scip4Y)*atom1.dipole.x + 0.5f*psc5*(sci3Y + scip3Y)*atom2.dipole.x + gfi5*qir1 + gfi6*qkr1;
#endif
ftm2i2 += gfi1*yr;
diff0 = atom1.inducedDipoleS.y - atom1.inducedDipole.y;
diff1 = atom1.inducedDipolePolarS.y - atom1.inducedDipolePolar.y;
#ifdef APPLY_SCALE
ftm2i2 += 0.5f*(-atom2.q*(diff0*psc3 + diff1*dsc3) + sc4*(diff0*psc5 + diff1*dsc5) - sc6*(diff0*psc7 + diff1*dsc7));
#else
ftm2i2 += 0.5f*(-atom2.q*psc3*(diff0 + diff1) + sc4*psc5*(diff0 + diff1) - sc6*psc7*(diff0 + diff1));
#endif
diff0 = atom2.inducedDipoleS.y - atom2.inducedDipole.y;
diff1 = atom2.inducedDipolePolarS.y - atom2.inducedDipolePolar.y;
#ifdef APPLY_SCALE
ftm2i2 += 0.5f*(atom1.q*(diff0*psc3 + diff1*dsc3) + sc3*(diff0*psc5 + diff1*dsc5) + sc5*(diff0*psc7 + diff1*dsc7));
ftm2i2 += 0.5f*(sci4Y*psc5+scip4Y*dsc5)*atom1.dipole.y + 0.5f*(sci3Y*psc5+scip3Y*dsc5)*atom2.dipole.y + gfi5*qir2 + gfi6*qkr2;
#else
ftm2i2 += 0.5f*(atom1.q*psc3*(diff0 + diff1) + sc3*psc5*(diff0 + diff1) + sc5*psc7*(diff0 + diff1));
ftm2i2 += 0.5f*psc5*(sci4Y +scip4Y)*atom1.dipole.y + 0.5f*psc5*(sci3Y +scip3Y)*atom2.dipole.y + gfi5*qir2 + gfi6*qkr2;
#endif
ftm2i3 += gfi1*zr;
diff0 = atom1.inducedDipoleS.z - atom1.inducedDipole.z;
diff1 = atom1.inducedDipolePolarS.z - atom1.inducedDipolePolar.z;
#ifdef APPLY_SCALE
ftm2i3 += 0.5f*(-atom2.q*(diff0*psc3 + diff1*dsc3) + sc4*(diff0*psc5 + diff1*dsc5) - sc6*(diff0*psc7 + diff1*dsc7));
#else
ftm2i3 += 0.5f*(-atom2.q*psc3*(diff0 + diff1) + sc4*psc5*(diff0 + diff1) - sc6*psc7*(diff0 + diff1));
#endif
diff0 = atom2.inducedDipoleS.z - atom2.inducedDipole.z;
diff1 = atom2.inducedDipolePolarS.z - atom2.inducedDipolePolar.z;
#ifdef APPLY_SCALE
ftm2i3 += 0.5f*(atom1.q*(diff0*psc3 + diff1*dsc3) + sc3*(diff0*psc5 + diff1*dsc5) + sc5*(diff0*psc7 + diff1*dsc7));
ftm2i3 += 0.5f*(sci4Y*psc5+scip4Y*dsc5)*atom1.dipole.z + 0.5f*(sci3Y*psc5+scip3Y*dsc5)*atom2.dipole.z + gfi5*qir3 + gfi6*qkr3;
#else
ftm2i3 += 0.5f*(atom1.q*psc3*(diff0 + diff1) + sc3*psc5*(diff0 + diff1) + sc5*psc7*(diff0 + diff1));
ftm2i3 += 0.5f*psc5*(sci4Y + scip4Y)*atom1.dipole.z + 0.5f*psc5*(sci3Y+scip3Y)*atom2.dipole.z + gfi5*qir3 + gfi6*qkr3;
#endif
// intermediate values needed for partially excluded interactions
......@@ -475,31 +404,17 @@ __device__ void computeOneEDiffInteractionT3(AtomData4& atom1, volatile AtomData
#endif
#ifdef T1
#ifdef APPLY_SCALE
real gti2 = 0.5f*(sci4Y*psc5 + scip4Y*dsc5);
real ttm2i1 = -(dixuk1*psc3+dixukp1*dsc3)*0.5f + gti2*dixr1 + ((ukxqir1+rxqiuk1)*psc5 +(ukxqirp1+rxqiukp1)*dsc5) - gfi5*rxqir1;
real ttm2i2 = -(dixuk2*psc3+dixukp2*dsc3)*0.5f + gti2*dixr2 + ((ukxqir2+rxqiuk2)*psc5 +(ukxqirp2+rxqiukp2)*dsc5) - gfi5*rxqir2;
real ttm2i3 = -(dixuk3*psc3+dixukp3*dsc3)*0.5f + gti2*dixr3 + ((ukxqir3+rxqiuk3)*psc5 +(ukxqirp3+rxqiukp3)*dsc5) - gfi5*rxqir3;
#else
real gti2 = 0.5f*psc5*(sci4Y + scip4Y);
real ttm2i1 = -psc3*(dixuk1 + dixukp1)*0.5f + gti2*dixr1 + psc5*((ukxqir1+rxqiuk1) + (ukxqirp1+rxqiukp1)) - gfi5*rxqir1;
real ttm2i2 = -psc3*(dixuk2 + dixukp2)*0.5f + gti2*dixr2 + psc5*((ukxqir2+rxqiuk2) + (ukxqirp2+rxqiukp2)) - gfi5*rxqir2;
real ttm2i3 = -psc3*(dixuk3 + dixukp3)*0.5f + gti2*dixr3 + psc5*((ukxqir3+rxqiuk3) + (ukxqirp3+rxqiukp3)) - gfi5*rxqir3;
#endif
#endif
#ifdef T3
#ifdef APPLY_SCALE
real gti3 = 0.5f*(sci3Y*psc5 + scip3Y*dsc5);
real ttm3i1 = -(dkxui1*psc3+dkxuip1*dsc3)*0.5f + gti3*dkxr1 - ((uixqkr1+rxqkui1)*psc5 +(uixqkrp1+rxqkuip1)*dsc5) - gfi6*rxqkr1;
real ttm3i2 = -(dkxui2*psc3+dkxuip2*dsc3)*0.5f + gti3*dkxr2 - ((uixqkr2+rxqkui2)*psc5 +(uixqkrp2+rxqkuip2)*dsc5) - gfi6*rxqkr2;
real ttm3i3 = -(dkxui3*psc3+dkxuip3*dsc3)*0.5f + gti3*dkxr3 - ((uixqkr3+rxqkui3)*psc5 +(uixqkrp3+rxqkuip3)*dsc5) - gfi6*rxqkr3;
#else
real gti3 = 0.5f*psc5*(sci3Y + scip3Y);
real ttm3i1 = -psc3*(dkxui1 + dkxuip1)*0.5f + gti3*dkxr1 - psc5*((uixqkr1+rxqkui1) + (uixqkrp1+rxqkuip1)) - gfi6*rxqkr1;
real ttm3i2 = -psc3*(dkxui2 + dkxuip2)*0.5f + gti3*dkxr2 - psc5*((uixqkr2+rxqkui2) + (uixqkrp2+rxqkuip2)) - gfi6*rxqkr2;
real ttm3i3 = -psc3*(dkxui3 + dkxuip3)*0.5f + gti3*dkxr3 - psc5*((uixqkr3+rxqkui3) + (uixqkrp3+rxqkuip3)) - gfi6*rxqkr3;
#endif
#endif
// update force and torque on site k
......@@ -603,27 +518,15 @@ __device__ void computeOneEDiffInteractionT3(AtomData4& atom1, volatile AtomData
#endif
#ifdef T1
#ifdef APPLY_SCALE
ttm2i1 = -(dixuk1*psc3+dixukp1*dsc3)*0.5f + ((ukxqir1+rxqiuk1)*psc5 +(ukxqirp1+rxqiukp1)*dsc5);
ttm2i2 = -(dixuk2*psc3+dixukp2*dsc3)*0.5f + ((ukxqir2+rxqiuk2)*psc5 +(ukxqirp2+rxqiukp2)*dsc5);
ttm2i3 = -(dixuk3*psc3+dixukp3*dsc3)*0.5f + ((ukxqir3+rxqiuk3)*psc5 +(ukxqirp3+rxqiukp3)*dsc5);
#else
ttm2i1 = -psc3*(dixuk1+dixukp1)*0.5f + psc5*((ukxqir1+rxqiuk1) + (ukxqirp1+rxqiukp1));
ttm2i2 = -psc3*(dixuk2+dixukp2)*0.5f + psc5*((ukxqir2+rxqiuk2) + (ukxqirp2+rxqiukp2));
ttm2i3 = -psc3*(dixuk3+dixukp3)*0.5f + psc5*((ukxqir3+rxqiuk3) + (ukxqirp3+rxqiukp3));
#endif
#endif
#ifdef T3
#ifdef APPLY_SCALE
ttm3i1 = -(dkxui1*psc3+dkxuip1*dsc3)*0.5f - ((uixqkr1+rxqkui1)*psc5 +(uixqkrp1+rxqkuip1)*dsc5);
ttm3i2 = -(dkxui2*psc3+dkxuip2*dsc3)*0.5f - ((uixqkr2+rxqkui2)*psc5 +(uixqkrp2+rxqkuip2)*dsc5);
ttm3i3 = -(dkxui3*psc3+dkxuip3*dsc3)*0.5f - ((uixqkr3+rxqkui3)*psc5 +(uixqkrp3+rxqkuip3)*dsc5);
#else
ttm3i1 = -psc3*(dkxui1 + dkxuip1)*0.5f - psc5*((uixqkr1+rxqkui1) + (uixqkrp1+rxqkuip1));
ttm3i2 = -psc3*(dkxui2 + dkxuip2)*0.5f - psc5*((uixqkr2+rxqkui2) + (uixqkrp2+rxqkuip2));
ttm3i3 = -psc3*(dkxui3 + dkxuip3)*0.5f - psc5*((uixqkr3+rxqkui3) + (uixqkrp3+rxqkuip3));
#endif
#endif
// update force and torque on site k;
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment