void StaticMap::SendPosition(Player *player)
{
switch (PosType())
{
case POSITION_SIMPLE :
player->SendSimplePositionalUpdate(
GameID(),
PosInfo());
break;
case POSITION_PLANET :
player->SendPlanetPositionalUpdate(
GameID(),
PosInfo());
break;
case POSITION_CONSTANT :
player->SendConstantPositionalUpdate(
GameID(),
PosX(),
PosY(),
PosZ(),
Orientation());
break;
}
}
const PosType Renderer::getWorldCoord(int x, int y, float z)
{
GLint viewport[4];
GLdouble mvmatrix[16], projmatrix[16];
GLdouble wx, wy, wz; /* returned world x, y, z coords */
glGetIntegerv (GL_VIEWPORT, viewport);
glGetDoublev (GL_MODELVIEW_MATRIX, mvmatrix);
glGetDoublev (GL_PROJECTION_MATRIX, projmatrix);
gluUnProject (x, y, z, mvmatrix, projmatrix, viewport, &wx, &wy, &wz);
std::cout << "SCREEN {" << x << ";" << y << ";" << z << std::endl << std::flush;
std::cout << "WORLD {" << wx << ";" << wy << ";" << wz << std::endl << std::flush;
return PosType(wx, wy, wz);
}
VBM3D_Process_Base::Pos3PairCode VBM3D_Process_Base::BlockMatching(
const std::vector<const FLType *> &ref, PCType j, PCType i) const
{
// Skip block matching if GroupSize is 1 or thMSE is not positive,
// and take the reference block as the only element in the group
if (d.para.GroupSize == 1 || d.para.thMSE <= 0)
{
return Pos3PairCode(1, Pos3Pair(KeyType(0), Pos3Type(cur, j, i)));
}
int f;
Pos3PairCode matchCode;
PosPairCode frameMatch;
PosCode prePosCode;
// Get reference block from the reference plane in current frame
block_type refBlock(ref[cur], ref_stride[0], d.para.BlockSize, d.para.BlockSize, PosType(j, i));
// Block Matching in current frame
f = cur;
frameMatch = refBlock.BlockMatchingMulti(ref[f],
ref_height[0], ref_width[0], ref_stride[0], FLType(1),
d.para.BMrange, d.para.BMstep, d.para.thMSE, 1, d.para.GroupSize, true);
matchCode.resize(matchCode.size() + frameMatch.size());
std::transform(frameMatch.begin(), frameMatch.end(),
matchCode.end() - frameMatch.size(), [&](const PosPair &x)
{
return Pos3Pair(x.first, Pos3Type(x.second, f));
});
PCType nextPosNum = Min(d.para.PSnum, static_cast<PCType>(frameMatch.size()));
PosCode curPosCode(nextPosNum);
std::transform(frameMatch.begin(), frameMatch.begin() + nextPosNum,
curPosCode.begin(), [](const PosPair &x)
{
return x.second;
});
PosCode curSearchPos = refBlock.GenSearchPos(curPosCode,
ref_height[0], ref_width[0], d.para.PSrange, d.para.PSstep);
// Predictive Search Block Matching in backward frames
f = cur - 1;
for (; f >= 0; --f)
{
if (f == cur - 1)
{
frameMatch = refBlock.BlockMatchingMulti(ref[f], ref_stride[0], FLType(1),
curSearchPos, d.para.thMSE, d.para.GroupSize, true);
}
else
{
PCType nextPosNum = Min(d.para.PSnum, static_cast<PCType>(frameMatch.size()));
prePosCode.resize(nextPosNum);
std::transform(frameMatch.begin(), frameMatch.begin() + nextPosNum,
prePosCode.begin(), [](const PosPair &x)
{
return x.second;
});
PosCode searchPos = refBlock.GenSearchPos(prePosCode,
ref_height[0], ref_width[0], d.para.PSrange, d.para.PSstep);
frameMatch = refBlock.BlockMatchingMulti(ref[f], ref_stride[0], FLType(1),
searchPos, d.para.thMSE, d.para.GroupSize, true);
}
matchCode.resize(matchCode.size() + frameMatch.size());
std::transform(frameMatch.begin(), frameMatch.end(),
matchCode.end() - frameMatch.size(), [&](const PosPair &x)
{
return Pos3Pair(x.first, Pos3Type(x.second, f));
});
}
// Predictive Search Block Matching in forward frames
f = cur + 1;
for (; f < frames; ++f)
{
if (f == cur + 1)
{
frameMatch = refBlock.BlockMatchingMulti(ref[f], ref_stride[0], FLType(1),
curSearchPos, d.para.thMSE, d.para.GroupSize, true);
}
else
{
PCType nextPosNum = Min(d.para.PSnum, static_cast<PCType>(frameMatch.size()));
prePosCode.resize(nextPosNum);
std::transform(frameMatch.begin(), frameMatch.begin() + nextPosNum,
prePosCode.begin(), [](const PosPair &x)
{
return x.second;
});
PosCode searchPos = refBlock.GenSearchPos(prePosCode,
ref_height[0], ref_width[0], d.para.PSrange, d.para.PSstep);
frameMatch = refBlock.BlockMatchingMulti(ref[f], ref_stride[0], FLType(1),
searchPos, d.para.thMSE, d.para.GroupSize, true);
}
matchCode.resize(matchCode.size() + frameMatch.size());
std::transform(frameMatch.begin(), frameMatch.end(),
matchCode.end() - frameMatch.size(), [&](const PosPair &x)
{
return Pos3Pair(x.first, Pos3Type(x.second, f));
});
}
// Limit the number of matched code to GroupSize
if (d.para.GroupSize > 0 && static_cast<PCType>(matchCode.size()) > d.para.GroupSize)
{
std::partial_sort(matchCode.begin() + 1, matchCode.begin() + d.para.GroupSize, matchCode.end());
matchCode.resize(d.para.GroupSize);
}
return matchCode;
}
void
EinsplineOrb<complex<double>,3>::read (hid_t h5file, string groupPath)
{
uMin = PosType(0.0, 0.0, 0.0);
uMax = PosType(1.0, 1.0, 1.0);
Center = PosType(0.5, 0.5, 0.5);
string centerName = groupPath + "center";
string vectorName = groupPath + "eigenvector";
string valueName = groupPath + "eigenvalue";
string radiusName = groupPath + "radius";
string uminName = groupPath + "umin";
string umaxName = groupPath + "umax";
HDFAttribIO<PosType> h_Center(Center);
HDFAttribIO<PosType> h_uMin(uMin);
HDFAttribIO<PosType> h_uMax(uMax);
HDFAttribIO<double> h_Radius(Radius);
HDFAttribIO<double> h_Energy(Energy);
h_Center.read(h5file, centerName.c_str());
uCenter = Lattice.toUnit (Center);
h_Radius.read(h5file, radiusName.c_str());
h_Energy.read(h5file, valueName.c_str());
h_uMax.read(h5file, umaxName.c_str());
h_Radius.read(h5file, radiusName.c_str());
Localized = Radius > 0.0;
Array<complex<double>,3> rawData;
HDFAttribIO<Array<complex<double>,3> > h_rawData(rawData);
h_rawData.read(h5file, vectorName.c_str());
int nx, ny, nz;
nx = rawData.size(0);
ny=rawData.size(1);
nz=rawData.size(2);
Ugrid x_grid, y_grid, z_grid;
BCtype_z xBC, yBC, zBC;
if (Localized)
{
xBC.lCode = NATURAL;
xBC.rCode = NATURAL;
yBC.lCode = NATURAL;
yBC.rCode = NATURAL;
zBC.lCode = NATURAL;
zBC.rCode = NATURAL;
x_grid.start = uMin[0];
x_grid.end = uMax[0];
x_grid.num = nx;
y_grid.start = uMin[1];
y_grid.end = uMax[1];
y_grid.num = ny;
z_grid.start = uMin[2];
z_grid.end = uMax[2];
z_grid.num = nz;
Spline = create_UBspline_3d_z (x_grid, y_grid, z_grid, xBC, yBC, zBC, rawData.data());
}
else
{
Array<complex<double>,3> splineData;
splineData.resize(nx-1,ny-1,nz-1);
for (int ix=0; ix<nx-1; ix++)
for (int iy=0; iy<ny-1; iy++)
for (int iz=0; iz<nz-1; iz++)
splineData(ix,iy,iz) = rawData(ix,iy,iz);
xBC.lCode = PERIODIC;
xBC.rCode = PERIODIC;
yBC.lCode = PERIODIC;
yBC.rCode = PERIODIC;
zBC.lCode = PERIODIC;
zBC.rCode = PERIODIC;
x_grid.start = 0.0;
x_grid.end = 1.0;
x_grid.num = nx-1;
y_grid.start = 0.0;
y_grid.end = 1.0;
y_grid.num = ny-1;
z_grid.start = 0.0;
z_grid.end = 1.0;
z_grid.num = nz-1;
Spline = create_UBspline_3d_z (x_grid, y_grid, z_grid, xBC, yBC, zBC, splineData.data());
}
}
bool DMCcuda::run()
{
bool NLmove = NonLocalMove == "yes";
bool scaleweight = ScaleWeight == "yes";
if (NLmove)
app_log() << " Using Casula nonlocal moves in DMCcuda.\n";
if (scaleweight)
app_log() << " Scaling weight per Umrigar/Nightengale.\n";
resetRun();
Mover->MaxAge = 1;
IndexType block = 0;
IndexType nAcceptTot = 0;
IndexType nRejectTot = 0;
int nat = W.getTotalNum();
int nw = W.getActiveWalkers();
vector<RealType> LocalEnergy(nw), LocalEnergyOld(nw),
oldScale(nw), newScale(nw);
vector<PosType> delpos(nw);
vector<PosType> dr(nw);
vector<PosType> newpos(nw);
vector<ValueType> ratios(nw), rplus(nw), rminus(nw), R2prop(nw), R2acc(nw);
vector<PosType> oldG(nw), newG(nw);
vector<ValueType> oldL(nw), newL(nw);
vector<Walker_t*> accepted(nw);
Matrix<ValueType> lapl(nw, nat);
Matrix<GradType> grad(nw, nat);
vector<ValueType> V2(nw), V2bar(nw);
vector<vector<NonLocalData> > Txy(nw);
for (int iw=0; iw<nw; iw++)
W[iw]->Weight = 1.0;
do {
IndexType step = 0;
nAccept = nReject = 0;
Estimators->startBlock(nSteps);
do {
step++;
CurrentStep++;
nw = W.getActiveWalkers();
ResizeTimer.start();
LocalEnergy.resize(nw); oldScale.resize(nw);
newScale.resize(nw); delpos.resize(nw);
dr.resize(nw); newpos.resize(nw);
ratios.resize(nw); rplus.resize(nw);
rminus.resize(nw); oldG.resize(nw);
newG.resize(nw); oldL.resize(nw);
newL.resize(nw); accepted.resize(nw);
lapl.resize(nw, nat); grad.resize(nw, nat);
R2prop.resize(nw,0.0); R2acc.resize(nw,0.0);
V2.resize(nw,0.0); V2bar.resize(nw,0.0);
W.updateLists_GPU();
ResizeTimer.stop();
if (NLmove) {
Txy.resize(nw);
for (int iw=0; iw<nw; iw++) {
Txy[iw].clear();
Txy[iw].push_back(NonLocalData(-1, 1.0, PosType()));
}
}
for (int iw=0; iw<nw; iw++)
W[iw]->Age++;
DriftDiffuseTimer.start();
for(int iat=0; iat<nat; iat++) {
Psi.getGradient (W, iat, oldG);
//create a 3N-Dimensional Gaussian with variance=1
makeGaussRandomWithEngine(delpos,Random);
for(int iw=0; iw<nw; iw++) {
delpos[iw] *= m_sqrttau;
oldScale[iw] = getDriftScale(m_tauovermass,oldG[iw]);
dr[iw] = delpos[iw] + (oldScale[iw]*oldG[iw]);
newpos[iw]=W[iw]->R[iat] + dr[iw];
ratios[iw] = 1.0;
R2prop[iw] += dot(delpos[iw], delpos[iw]);
}
W.proposeMove_GPU(newpos, iat);
Psi.ratio(W,iat,ratios,newG, newL);
accepted.clear();
vector<bool> acc(nw, false);
for(int iw=0; iw<nw; ++iw) {
PosType drOld =
newpos[iw] - (W[iw]->R[iat] + oldScale[iw]*oldG[iw]);
RealType logGf = -m_oneover2tau * dot(drOld, drOld);
newScale[iw] = getDriftScale(m_tauovermass,newG[iw]);
PosType drNew =
(newpos[iw] + newScale[iw]*newG[iw]) - W[iw]->R[iat];
RealType logGb = -m_oneover2tau * dot(drNew, drNew);
RealType x = logGb - logGf;
RealType prob = ratios[iw]*ratios[iw]*std::exp(x);
if(Random() < prob && ratios[iw] > 0.0) {
accepted.push_back(W[iw]);
nAccept++;
W[iw]->R[iat] = newpos[iw];
W[iw]->Age = 0;
acc[iw] = true;
R2acc[iw] += dot(delpos[iw], delpos[iw]);
V2[iw] += m_tauovermass * m_tauovermass * dot(newG[iw],newG[iw]);
V2bar[iw] += newScale[iw] * newScale[iw] * dot(newG[iw],newG[iw]);
}
else {
nReject++;
V2[iw] += m_tauovermass * m_tauovermass * dot(oldG[iw],oldG[iw]);
V2bar[iw] += oldScale[iw] * oldScale[iw] * dot(oldG[iw],oldG[iw]);
}
}
W.acceptMove_GPU(acc);
if (accepted.size())
Psi.update(accepted,iat);
}
DriftDiffuseTimer.stop();
// Psi.recompute(W, false);
Psi.gradLapl(W, grad, lapl);
HTimer.start();
if (NLmove) H.evaluate (W, LocalEnergy, Txy);
else H.evaluate (W, LocalEnergy);
HTimer.stop();
// for (int iw=0; iw<nw; iw++) {
// branchEngine->clampEnergy(LocalEnergy[iw]);
// W[iw]->getPropertyBase()[LOCALENERGY] = LocalEnergy[iw];
// }
if (CurrentStep == 1)
LocalEnergyOld = LocalEnergy;
if (NLmove) {
// Now, attempt nonlocal move
accepted.clear();
vector<int> iatList;
vector<PosType> accPos;
for (int iw=0; iw<nw; iw++) {
/// HACK HACK HACK
// if (LocalEnergy[iw] < -2300.0) {
// cerr << "Walker " << iw << " has energy "
// << LocalEnergy[iw] << endl;;
// double maxWeight = 0.0;
// int elMax = -1;
// PosType posMax;
// for (int j=1; j<Txy[iw].size(); j++)
// if (std::fabs(Txy[iw][j].Weight) > std::fabs(maxWeight)) {
// maxWeight = Txy[iw][j].Weight;
// elMax = Txy[iw][j].PID;
// posMax = W[iw]->R[elMax] + Txy[iw][j].Delta;
// }
// cerr << "Maximum weight is " << maxWeight << " for electron "
// << elMax << " at position " << posMax << endl;
// PosType unit = W.Lattice.toUnit(posMax);
// unit[0] -= round(unit[0]);
// unit[1] -= round(unit[1]);
// unit[2] -= round(unit[2]);
// cerr << "Reduced position = " << unit << endl;
// }
int ibar = NLop.selectMove(Random(), Txy[iw]);
if (ibar) {
accepted.push_back(W[iw]);
int iat = Txy[iw][ibar].PID;
iatList.push_back(iat);
accPos.push_back(W[iw]->R[iat] + Txy[iw][ibar].Delta);
}
}
if (accepted.size()) {
Psi.ratio(accepted,iatList, accPos, ratios, newG, newL);
Psi.update(accepted,iatList);
for (int i=0; i<accepted.size(); i++)
accepted[i]->R[iatList[i]] = accPos[i];
W.NLMove_GPU (accepted, accPos, iatList);
// HACK HACK HACK
// Recompute the kinetic energy
// Psi.gradLapl(W, grad, lapl);
// H.evaluate (W, LocalEnergy);
//W.copyWalkersToGPU();
}
}
// Now branch
BranchTimer.start();
for (int iw=0; iw<nw; iw++) {
RealType v2=0.0, v2bar=0.0;
for(int iat=0; iat<nat; iat++) {
v2 += dot(W.G[iat],W.G[iat]);
RealType newscale = getDriftScale(m_tauovermass,newG[iw]);
v2 += m_tauovermass * m_tauovermass * dot(newG[iw],newG[iw]);
v2bar += newscale * newscale * dot(newG[iw],newG[iw]);
}
//RealType scNew = std::sqrt(V2bar[iw] / V2[iw]);
RealType scNew = std::sqrt(v2bar/v2);
RealType scOld = (CurrentStep == 1) ? scNew : W[iw]->getPropertyBase()[DRIFTSCALE];
W[iw]->getPropertyBase()[DRIFTSCALE] = scNew;
// fprintf (stderr, "iw = %d scNew = %1.8f scOld = %1.8f\n", iw, scNew, scOld);
RealType tauRatio = R2acc[iw] / R2prop[iw];
if (tauRatio < 0.5)
cerr << " tauRatio = " << tauRatio << endl;
RealType taueff = m_tauovermass * tauRatio;
if (scaleweight)
W[iw]->Weight *= branchEngine->branchWeightTau
(LocalEnergy[iw], LocalEnergyOld[iw], scNew, scOld, taueff);
else
W[iw]->Weight *= branchEngine->branchWeight
(LocalEnergy[iw], LocalEnergyOld[iw]);
W[iw]->getPropertyBase()[R2ACCEPTED] = R2acc[iw];
W[iw]->getPropertyBase()[R2PROPOSED] = R2prop[iw];
}
Mover->setMultiplicity(W.begin(), W.end());
branchEngine->branch(CurrentStep,W);
nw = W.getActiveWalkers();
LocalEnergyOld.resize(nw);
for (int iw=0; iw<nw; iw++)
LocalEnergyOld[iw] = W[iw]->getPropertyBase()[LOCALENERGY];
BranchTimer.stop();
} while(step<nSteps);
Psi.recompute(W, true);
double accept_ratio = (double)nAccept/(double)(nAccept+nReject);
Estimators->stopBlock(accept_ratio);
nAcceptTot += nAccept;
nRejectTot += nReject;
++block;
recordBlock(block);
} while(block<nBlocks);
//finalize a qmc section
return finalize(block);
}
bool DMCcuda::runWithNonlocal()
{
resetRun();
Mover->MaxAge = 1;
IndexType block = 0;
IndexType nAcceptTot = 0;
IndexType nRejectTot = 0;
int nat = W.getTotalNum();
int nw = W.getActiveWalkers();
vector<RealType> LocalEnergy(nw), LocalEnergyOld(nw),
oldScale(nw), newScale(nw);
vector<PosType> delpos(nw);
vector<PosType> dr(nw);
vector<PosType> newpos(nw);
vector<ValueType> ratios(nw), rplus(nw), rminus(nw), R2prop(nw), R2acc(nw);
vector<PosType> oldG(nw), newG(nw);
vector<ValueType> oldL(nw), newL(nw);
vector<Walker_t*> accepted(nw);
Matrix<ValueType> lapl(nw, nat);
Matrix<GradType> grad(nw, nat);
vector<vector<NonLocalData> > Txy(nw);
for (int iw=0; iw<nw; iw++)
W[iw]->Weight = 1.0;
do {
IndexType step = 0;
nAccept = nReject = 0;
Estimators->startBlock(nSteps);
do {
step++;
CurrentStep++;
nw = W.getActiveWalkers();
LocalEnergy.resize(nw); oldScale.resize(nw);
newScale.resize(nw); delpos.resize(nw);
dr.resize(nw); newpos.resize(nw);
ratios.resize(nw); rplus.resize(nw);
rminus.resize(nw); oldG.resize(nw);
newG.resize(nw); oldL.resize(nw);
newL.resize(nw); accepted.resize(nw);
lapl.resize(nw, nat); grad.resize(nw, nat);
R2prop.resize(nw,0.0); R2acc.resize(nw,0.0);
W.updateLists_GPU();
Txy.resize(nw);
for (int iw=0; iw<nw; iw++) {
Txy[iw].clear();
Txy[iw].push_back(NonLocalData(-1, 1.0, PosType()));
W[iw]->Age++;
}
for(int iat=0; iat<nat; iat++) {
Psi.getGradient (W, iat, oldG);
//create a 3N-Dimensional Gaussian with variance=1
makeGaussRandomWithEngine(delpos,Random);
for(int iw=0; iw<nw; iw++) {
delpos[iw] *= m_sqrttau;
oldScale[iw] = getDriftScale(m_tauovermass,oldG[iw]);
dr[iw] = delpos[iw] + (oldScale[iw]*oldG[iw]);
newpos[iw]=W[iw]->R[iat] + dr[iw];
ratios[iw] = 1.0;
R2prop[iw] += dot(delpos[iw], delpos[iw]);
}
W.proposeMove_GPU(newpos, iat);
Psi.ratio(W,iat,ratios,newG, newL);
accepted.clear();
vector<bool> acc(nw, false);
for(int iw=0; iw<nw; ++iw) {
PosType drOld =
newpos[iw] - (W[iw]->R[iat] + oldScale[iw]*oldG[iw]);
RealType logGf = -m_oneover2tau * dot(drOld, drOld);
newScale[iw] = getDriftScale(m_tauovermass,newG[iw]);
PosType drNew =
(newpos[iw] + newScale[iw]*newG[iw]) - W[iw]->R[iat];
RealType logGb = -m_oneover2tau * dot(drNew, drNew);
RealType x = logGb - logGf;
RealType prob = ratios[iw]*ratios[iw]*std::exp(x);
if(Random() < prob && ratios[iw] > 0.0) {
accepted.push_back(W[iw]);
nAccept++;
W[iw]->R[iat] = newpos[iw];
W[iw]->Age = 0;
acc[iw] = true;
R2acc[iw] += dot(delpos[iw], delpos[iw]);
}
else
nReject++;
}
W.acceptMove_GPU(acc);
if (accepted.size())
Psi.update(accepted,iat);
}
for (int iw=0; iw < nw; iw++)
if (W[iw]->Age)
cerr << "Encountered stuck walker with iw=" << iw << endl;
// Psi.recompute(W, false);
Psi.gradLapl(W, grad, lapl);
H.evaluate (W, LocalEnergy, Txy);
if (CurrentStep == 1)
LocalEnergyOld = LocalEnergy;
// Now, attempt nonlocal move
accepted.clear();
vector<int> iatList;
vector<PosType> accPos;
for (int iw=0; iw<nw; iw++) {
int ibar = NLop.selectMove(Random(), Txy[iw]);
// cerr << "Txy[iw].size() = " << Txy[iw].size() << endl;
if (ibar) {
accepted.push_back(W[iw]);
int iat = Txy[iw][ibar].PID;
iatList.push_back(iat);
accPos.push_back(W[iw]->R[iat] + Txy[iw][ibar].Delta);
}
}
if (accepted.size()) {
// W.proposeMove_GPU(newpos, iatList);
Psi.ratio(accepted,iatList, accPos, ratios, newG, newL);
Psi.update(accepted,iatList);
for (int i=0; i<accepted.size(); i++)
accepted[i]->R[iatList[i]] = accPos[i];
W.copyWalkersToGPU();
}
// Now branch
for (int iw=0; iw<nw; iw++) {
W[iw]->Weight *= branchEngine->branchWeight(LocalEnergy[iw], LocalEnergyOld[iw]);
W[iw]->getPropertyBase()[R2ACCEPTED] = R2acc[iw];
W[iw]->getPropertyBase()[R2PROPOSED] = R2prop[iw];
}
Mover->setMultiplicity(W.begin(), W.end());
branchEngine->branch(CurrentStep,W);
nw = W.getActiveWalkers();
LocalEnergyOld.resize(nw);
for (int iw=0; iw<nw; iw++)
LocalEnergyOld[iw] = W[iw]->getPropertyBase()[LOCALENERGY];
} while(step<nSteps);
Psi.recompute(W, true);
double accept_ratio = (double)nAccept/(double)(nAccept+nReject);
Estimators->stopBlock(accept_ratio);
nAcceptTot += nAccept;
nRejectTot += nReject;
++block;
recordBlock(block);
} while(block<nBlocks);
//finalize a qmc section
return finalize(block);
}
void PWOrbitalBuilder::transform2GridData(PWBasis::GIndex_t& nG, int spinIndex, PWOrbitalSet& pwFunc)
{
ostringstream splineTag;
splineTag << "eigenstates_"<<nG[0]<<"_"<<nG[1]<<"_"<<nG[2];
herr_t status = H5Eset_auto(NULL, NULL);
app_log() << " splineTag " << splineTag.str() << endl;
hid_t es_grp_id;
status = H5Gget_objinfo (hfileID, splineTag.str().c_str(), 0, NULL);
if(status)
{
es_grp_id = H5Gcreate(hfileID,splineTag.str().c_str(),0);
HDFAttribIO<PWBasis::GIndex_t> t(nG);
t.write(es_grp_id,"grid");
}
else
{
es_grp_id = H5Gopen(hfileID,splineTag.str().c_str());
}
string tname=myParam->getTwistName();
hid_t twist_grp_id;
status = H5Gget_objinfo (es_grp_id, tname.c_str(), 0, NULL);
if(status)
twist_grp_id = H5Gcreate(es_grp_id,tname.c_str(),0);
else
twist_grp_id = H5Gopen(es_grp_id,tname.c_str());
HDFAttribIO<PosType> hdfobj_twist(TwistAngle);
hdfobj_twist.write(twist_grp_id,"twist_angle");
ParticleSet::ParticleLayout_t& lattice(targetPtcl.Lattice);
RealType dx=1.0/static_cast<RealType>(nG[0]-1);
RealType dy=1.0/static_cast<RealType>(nG[1]-1);
RealType dz=1.0/static_cast<RealType>(nG[2]-1);
#if defined(VERYTINYMEMORY)
typedef Array<ValueType,3> StorageType;
StorageType inData(nG[0],nG[1],nG[2]);
int ib=0;
while(ib<myParam->numBands)
{
string bname(myParam->getBandName(ib));
status = H5Gget_objinfo (twist_grp_id, bname.c_str(), 0, NULL);
hid_t band_grp_id, spin_grp_id=-1;
if(status)
{
band_grp_id = H5Gcreate(twist_grp_id,bname.c_str(),0);
}
else
{
band_grp_id = H5Gopen(twist_grp_id,bname.c_str());
}
hid_t parent_id=band_grp_id;
if(myParam->hasSpin)
{
bname=myParam->getSpinName(spinIndex);
status = H5Gget_objinfo (band_grp_id, bname.c_str(), 0, NULL);
if(status)
{
spin_grp_id = H5Gcreate(band_grp_id,bname.c_str(),0);
}
else
{
spin_grp_id = H5Gopen(band_grp_id,bname.c_str());
}
parent_id=spin_grp_id;
}
for(int ig=0; ig<nG[0]; ig++)
{
RealType x=ig*dx;
for(int jg=0; jg<nG[1]; jg++)
{
RealType y=jg*dy;
for(int kg=0; kg<nG[2]; kg++)
{
inData(ig,jg,kg)= pwFunc.evaluate(ib,lattice.toCart(PosType(x,y,kg*dz)));
}
}
}
app_log() << " Add spline data " << ib << " h5path=" << tname << "/eigvector" << endl;
HDFAttribIO<StorageType> t(inData);
t.write(parent_id,myParam->eigvecTag.c_str());
if(spin_grp_id>=0) H5Gclose(spin_grp_id);
H5Gclose(band_grp_id);
++ib;
}
#else
typedef Array<ValueType,3> StorageType;
vector<StorageType*> inData;
int nb=myParam->numBands;
for(int ib=0; ib<nb; ib++)
inData.push_back(new StorageType(nG[0],nG[1],nG[2]));
PosType tAngle=targetPtcl.Lattice.k_cart(TwistAngle);
PWOrbitalSet::ValueVector_t phi(nb);
for(int ig=0; ig<nG[0]; ig++)
{
RealType x=ig*dx;
for(int jg=0; jg<nG[1]; jg++)
{
RealType y=jg*dy;
for(int kg=0; kg<nG[2]; kg++)
{
targetPtcl.R[0]=lattice.toCart(PosType(x,y,kg*dz));
pwFunc.evaluate(targetPtcl,0,phi);
RealType x(dot(targetPtcl.R[0],tAngle));
ValueType phase(std::cos(x),-std::sin(x));
for(int ib=0; ib<nb; ib++)
(*inData[ib])(ig,jg,kg)=phase*phi[ib];
}
}
}
for(int ib=0; ib<nb; ib++)
{
string bname(myParam->getBandName(ib));
status = H5Gget_objinfo (twist_grp_id, bname.c_str(), 0, NULL);
hid_t band_grp_id, spin_grp_id=-1;
if(status)
{
band_grp_id = H5Gcreate(twist_grp_id,bname.c_str(),0);
}
else
{
band_grp_id = H5Gopen(twist_grp_id,bname.c_str());
}
hid_t parent_id=band_grp_id;
if(myParam->hasSpin)
{
bname=myParam->getSpinName(spinIndex);
status = H5Gget_objinfo (band_grp_id, bname.c_str(), 0, NULL);
if(status)
{
spin_grp_id = H5Gcreate(band_grp_id,bname.c_str(),0);
}
else
{
spin_grp_id = H5Gopen(band_grp_id,bname.c_str());
}
parent_id=spin_grp_id;
}
app_log() << " Add spline data " << ib << " h5path=" << tname << "/eigvector" << endl;
HDFAttribIO<StorageType> t(*(inData[ib]));
t.write(parent_id,myParam->eigvecTag.c_str());
if(spin_grp_id>=0) H5Gclose(spin_grp_id);
H5Gclose(band_grp_id);
}
for(int ib=0; ib<nb; ib++) delete inData[ib];
#endif
H5Gclose(twist_grp_id);
H5Gclose(es_grp_id);
}
SPOSetBase*
EinsplineSetBuilder::createSPOSet(xmlNodePtr cur)
{
//use 2 bohr as the default when truncated orbitals are used based on the extend of the ions
BufferLayer=2.0;
OhmmsAttributeSet attribs;
int numOrbs = 0;
qafm=0;
int sortBands(1);
string sourceName;
string spo_prec("double");
string truncate("no");
#if defined(QMC_CUDA)
string useGPU="yes";
#else
string useGPU="no";
#endif
attribs.add (H5FileName, "href");
attribs.add (TileFactor, "tile");
attribs.add (sortBands, "sort");
attribs.add (qafm, "afmshift");
attribs.add (TileMatrix, "tilematrix");
attribs.add (TwistNum, "twistnum");
attribs.add (givenTwist, "twist");
attribs.add (sourceName, "source");
attribs.add (MeshFactor, "meshfactor");
attribs.add (useGPU, "gpu");
attribs.add (spo_prec, "precision");
attribs.add (truncate, "truncate");
attribs.add (BufferLayer, "buffer");
attribs.put (XMLRoot);
attribs.add (numOrbs, "size");
attribs.add (numOrbs, "norbs");
attribs.put (cur);
///////////////////////////////////////////////
// Read occupation information from XML file //
///////////////////////////////////////////////
cur = cur->children;
int spinSet = -1;
vector<int> Occ_Old(0,0);
Occ.resize(0,0);
bool NewOcc(false);
while (cur != NULL)
{
string cname((const char*)(cur->name));
if(cname == "occupation")
{
string occ_mode("ground");
occ_format="energy";
particle_hole_pairs=0;
OhmmsAttributeSet oAttrib;
oAttrib.add(occ_mode,"mode");
oAttrib.add(spinSet,"spindataset");
oAttrib.add(occ_format,"format");
oAttrib.add(particle_hole_pairs,"pairs");
oAttrib.put(cur);
if(occ_mode == "excited")
{
putContent(Occ,cur);
}
else
if(occ_mode != "ground")
{
app_error() << "Only ground state occupation currently supported "
<< "in EinsplineSetBuilder.\n";
APP_ABORT("EinsplineSetBuilder::createSPOSet");
}
}
cur = cur->next;
}
if (Occ != Occ_Old)
{
NewOcc=true;
Occ_Old = Occ;
}
else
NewOcc=false;
#if defined(QMC_CUDA)
app_log() << "\t QMC_CUDA=1 Overwriting the precision of the einspline storage on the host.\n";
spo_prec="double"; //overwrite
#endif
H5OrbSet aset(H5FileName, spinSet, numOrbs);
std::map<H5OrbSet,SPOSetBase*,H5OrbSet>::iterator iter;
iter = SPOSetMap.find (aset);
if ((iter != SPOSetMap.end() ) && (!NewOcc) && (qafm==0))
{
qafm=0;
app_log() << "SPOSet parameters match in EinsplineSetBuilder: "
<< "cloning EinsplineSet object.\n";
return iter->second->makeClone();
}
// The tiling can be set by a simple vector, (e.g. 2x2x2), or by a
// full 3x3 matrix of integers. If the tilematrix was not set in
// the input file...
bool matrixNotSet = true;
for (int i=0; i<3; i++)
for (int j=0; j<3; j++)
matrixNotSet = matrixNotSet && (TileMatrix(i,j) == 0);
// then set the matrix to what may have been specified in the
// tiling vector
if (matrixNotSet)
for (int i=0; i<3; i++)
for (int j=0; j<3; j++)
TileMatrix(i,j) = (i==j) ? TileFactor(i) : 0;
if (myComm->rank() == 0)
fprintf (stderr, " [ %2d %2d %2d\n %2d %2d %2d\n %2d %2d %2d ]\n",
TileMatrix(0,0), TileMatrix(0,1), TileMatrix(0,2),
TileMatrix(1,0), TileMatrix(1,1), TileMatrix(1,2),
TileMatrix(2,0), TileMatrix(2,1), TileMatrix(2,2));
if (numOrbs == 0)
{
app_error() << "You must specify the number of orbitals in the input file.\n";
APP_ABORT("EinsplineSetBuilder::createSPOSet");
}
else
app_log() << " Reading " << numOrbs << " orbitals from HDF5 file.\n";
Timer mytimer;
mytimer.restart();
/////////////////////////////////////////////////////////////////
// Read the basic orbital information, without reading all the //
// orbitals themselves. //
/////////////////////////////////////////////////////////////////
if (myComm->rank() == 0)
if (!ReadOrbitalInfo())
{
app_error() << "Error reading orbital info from HDF5 file. Aborting.\n";
APP_ABORT("EinsplineSetBuilder::createSPOSet");
}
app_log() << "TIMER EinsplineSetBuilder::ReadOrbitalInfo " << mytimer.elapsed() << endl;
myComm->barrier();
mytimer.restart();
BroadcastOrbitalInfo();
app_log() << "TIMER EinsplineSetBuilder::BroadcastOrbitalInfo " << mytimer.elapsed() << endl;
app_log().flush();
///////////////////////////////////////////////////////////////////
// Now, analyze the k-point mesh to figure out the what k-points //
// are needed //
///////////////////////////////////////////////////////////////////
PrimCell.set(Lattice);
SuperCell.set(SuperLattice);
for (int iat=0; iat<AtomicOrbitals.size(); iat++)
AtomicOrbitals[iat].Lattice = Lattice;
// Copy supercell into the ParticleSets
// app_log() << "Overwriting XML lattice with that from the ESHDF file.\n";
// PtclPoolType::iterator piter;
// for(piter = ParticleSets.begin(); piter != ParticleSets.end(); piter++)
// piter->second->Lattice.copy(SuperCell);
AnalyzeTwists2();
//////////////////////////////////
// Create the OrbitalSet object
//////////////////////////////////
if (HaveLocalizedOrbs)
OrbitalSet = new EinsplineSetLocal;
#ifdef QMC_CUDA
else
if (AtomicOrbitals.size() > 0)
{
if (UseRealOrbitals)
OrbitalSet = new EinsplineSetHybrid<double>;
else
OrbitalSet = new EinsplineSetHybrid<complex<double> >;
}
#endif
else
{
if (UseRealOrbitals)
OrbitalSet = new EinsplineSetExtended<double>;
else
OrbitalSet = new EinsplineSetExtended<complex<double> >;
}
//set the internal parameters
setTiling(OrbitalSet,numOrbs);
if (HaveLocalizedOrbs)
{
EinsplineSetLocal *restrict orbitalSet =
dynamic_cast<EinsplineSetLocal*>(OrbitalSet);
#pragma omp critical(read_einspline_orbs)
{
if ((spinSet == LastSpinSet) && (numOrbs <= NumOrbitalsRead) && (!NewOcc) )
CopyBands(numOrbs);
else
{
// Now, figure out occupation for the bands and read them
OccupyBands(spinSet, sortBands);
ReadBands (spinSet, orbitalSet);
}
}
// Now, store what we have read
LastOrbitalSet = OrbitalSet;
LastSpinSet = spinSet;
NumOrbitalsRead = numOrbs;
}
else // Otherwise, use EinsplineSetExtended
{
mytimer.restart();
bool use_single= (spo_prec == "single" || spo_prec == "float");
if (UseRealOrbitals)
{
OccupyBands(spinSet, sortBands);
//check if a matching BsplineReaderBase exists
BsplineReaderBase* spline_reader=0;
//if(TargetPtcl.Lattice.SuperCellEnum != SUPERCELL_BULK && truncate=="yes")
if(truncate=="yes")
{
if(use_single)
{
if(TargetPtcl.Lattice.SuperCellEnum == SUPERCELL_OPEN)
spline_reader= new SplineMixedAdoptorReader<SplineOpenAdoptor<float,double,3> >(this);
else
if(TargetPtcl.Lattice.SuperCellEnum == SUPERCELL_SLAB)
spline_reader= new SplineMixedAdoptorReader<SplineMixedAdoptor<float,double,3> >(this);
else
spline_reader= new SplineAdoptorReader<SplineR2RAdoptor<float,double,3> >(this);
}
else
{
if(TargetPtcl.Lattice.SuperCellEnum == SUPERCELL_OPEN)
spline_reader= new SplineMixedAdoptorReader<SplineOpenAdoptor<double,double,3> >(this);
else
if(TargetPtcl.Lattice.SuperCellEnum == SUPERCELL_SLAB)
spline_reader= new SplineMixedAdoptorReader<SplineMixedAdoptor<double,double,3> >(this);
else
spline_reader= new SplineAdoptorReader<SplineR2RAdoptor<double,double,3> >(this);
}
}
else
{
if(use_single)
spline_reader= new SplineAdoptorReader<SplineR2RAdoptor<float,double,3> >(this);
}
if(spline_reader)
{
HasCoreOrbs=bcastSortBands(NumDistinctOrbitals,myComm->rank()==0);
SPOSetBase* bspline_zd=spline_reader->create_spline_set(spinSet,OrbitalSet);
delete spline_reader;
if(bspline_zd)
SPOSetMap[aset] = bspline_zd;
return bspline_zd;
}
else
{
app_log() << ">>>> Creating EinsplineSetExtended<double> <<<< " << endl;
EinsplineSetExtended<double> *restrict orbitalSet =
dynamic_cast<EinsplineSetExtended<double>* > (OrbitalSet);
if (Format == ESHDF)
ReadBands_ESHDF(spinSet,orbitalSet);
else
ReadBands(spinSet, orbitalSet);
}
}
else
{
OccupyBands(spinSet, sortBands);
BsplineReaderBase* spline_reader=0;
if(truncate == "yes")
{
app_log() << " Truncated orbitals with multiple kpoints are not supported yet!" << endl;
}
if(use_single)
{
#if defined(QMC_COMPLEX)
spline_reader= new SplineAdoptorReader<SplineC2CPackedAdoptor<float,double,3> >(this);
#else
spline_reader= new SplineAdoptorReader<SplineC2RPackedAdoptor<float,double,3> >(this);
#endif
}
if(spline_reader)
{
RotateBands_ESHDF(spinSet, dynamic_cast<EinsplineSetExtended<complex<double> >*>(OrbitalSet));
HasCoreOrbs=bcastSortBands(NumDistinctOrbitals,myComm->rank()==0);
SPOSetBase* bspline_zd=spline_reader->create_spline_set(spinSet,OrbitalSet);
delete spline_reader;
if(bspline_zd)
SPOSetMap[aset] = bspline_zd;
return bspline_zd;
}
else
{
EinsplineSetExtended<complex<double> > *restrict orbitalSet =
dynamic_cast<EinsplineSetExtended<complex<double> >*>(OrbitalSet);
if (Format == ESHDF)
ReadBands_ESHDF(spinSet,orbitalSet);
else
ReadBands(spinSet, orbitalSet);
}
}
app_log() << "TIMER EinsplineSetBuilder::ReadBands " << mytimer.elapsed() << endl;
}
#ifndef QMC_COMPLEX
if (myComm->rank()==0 && OrbitalSet->MuffinTins.size() > 0)
{
FILE *fout = fopen ("TestMuffins.dat", "w");
Vector<double> phi(numOrbs), lapl(numOrbs);
Vector<PosType> grad(numOrbs);
ParticleSet P;
P.R.resize(6);
for (int i=0; i<P.R.size(); i++)
P.R[i] = PosType (0.0, 0.0, 0.0);
PosType N = 0.25*PrimCell.a(0) + 0.25*PrimCell.a(1) + 0.25*PrimCell.a(2);
for (double x=-1.0; x<=1.0; x+=0.0000500113412)
{
// for (double x=-0.003; x<=0.003; x+=0.0000011329343481381) {
P.R[0] = x * (PrimCell.a(0) + 0.914*PrimCell.a(1) +
0.781413*PrimCell.a(2));
double r = std::sqrt(dot(P.R[0], P.R[0]));
double rN = std::sqrt(dot(P.R[0]-N, P.R[0]-N));
OrbitalSet->evaluate(P, 0, phi, grad, lapl);
// OrbitalSet->evaluate(P, 0, phi);
fprintf (fout, "%1.12e ", r*x/std::fabs(x));
for (int j=0; j<numOrbs; j++)
{
double gmag = std::sqrt(dot(grad[j],grad[j]));
fprintf (fout, "%16.12e ",
/*phi[j]*phi[j]**/(-5.0/r -0.5*lapl[j]/phi[j]));
// double E = -5.0/r -0.5*lapl[j]/phi[j];
fprintf (fout, "%16.12e ", phi[j]);
fprintf (fout, "%16.12e ", gmag);
}
fprintf (fout, "\n");
}
fclose(fout);
}
#endif
SPOSetMap[aset] = OrbitalSet;
if (sourceName.size() && (ParticleSets.find(sourceName) == ParticleSets.end()))
{
app_log() << " EinsplineSetBuilder creates a ParticleSet " << sourceName << endl;
ParticleSet* ions=new ParticleSet;
ions->Lattice=TargetPtcl.Lattice;
ESHDFIonsParser ap(*ions,H5FileID,myComm);
ap.put(XMLRoot);
ap.expand(TileMatrix);
ions->setName(sourceName);
ParticleSets[sourceName]=ions;
//overwrite the lattice and assign random
if(TargetPtcl.Lattice.SuperCellEnum)
{
TargetPtcl.Lattice=ions->Lattice;
makeUniformRandom(TargetPtcl.R);
TargetPtcl.R.setUnit(PosUnit::LatticeUnit);
TargetPtcl.convert2Cart(TargetPtcl.R);
TargetPtcl.createSK();
}
}
#ifdef QMC_CUDA
if (useGPU == "yes" || useGPU == "1")
{
app_log() << "Initializing GPU data structures.\n";
OrbitalSet->initGPU();
}
#endif
return OrbitalSet;
}
void
DiracDeterminantOpt::evaluateDerivatives(ParticleSet& P,
const opt_variables_type& active,
vector<RealType>& dlogpsi,
vector<RealType>& dhpsioverpsi)
{
// The dlogpsi part is simple -- just ratios
// First, evaluate the basis functions
// cerr << "GEMM 1:\n";
// fprintf (stderr, "FirstIndex = %d LastIndex=%d\n", FirstIndex,
// LastIndex);
resetParameters(active);
Phi->evaluateBasis (P, FirstIndex, LastIndex, BasisVals, BasisGrad, BasisLapl);
BLAS::gemm ('N', 'T', NumBasis, NumOrbitals, NumOrbitals, 1.0,
BasisVals.data(), NumBasis, psiM.data(), NumOrbitals,
0.0, dlogdet_dC.data(), NumBasis);
// for (int i=0; i<NumOrbitals; i++)
// for (int j=0; j<NumBasis; j++) {
// dlogdet_dC(i,j) = 0.0;
// for (int n=0; n<NumOrbitals; n++) {
// dlogdet_dC(i,j) += psiM(n,i) * BasisVals(n,j);
// // fprintf (stderr, "BasisVals(%d,%d) = %12.6e\n", n, j, BasisVals(n,j));
// }
// }
Gamma = dlogdet_dC;
// for (int n=0; n<NumOrbitals; n++)
// for (int j=0; j<NumBasis; j++) {
// Gamma(n,j) = 0.0;
// for (int k=0; k<NumOrbitals; k++) {
// Gamma(n,j) += psiM(k,n) * BasisVals(k,j);
// // fprintf (stderr, "BasisVals(%d,%d) = %12.6e\n", n, j, BasisVals(n,j));
// }
// }
// Now, d_dC should hold d/dC_{ij} log(det).
// Multiply L matrix by gamma matrix, as shown in eq. 17 of
// docs/OrbitalOptimization.tex
L_gamma = BasisLapl;
BLAS::gemm ('N', 'N', NumBasis, NumOrbitals, NumOrbitals, -1.0,
Gamma.data(), NumBasis, d2psiM.data(), NumOrbitals,
1.0, L_gamma.data(), NumBasis);
// for (int l=0; l<NumOrbitals; l++)
// for (int j=0; j<NumBasis; j++) {
// // L_gamma(l,j) = BasisLapl(l,j);
// L_gamma(l,j) = BasisLapl(l,j);
// G_gamma(l,j) = BasisGrad(l,j);
// for (int n=0; n<NumOrbitals; n++) {
// L_gamma(l,j) -= d2psiM(l,n)*Gamma(n,j);
// G_gamma(l,j) -= dpsiM(l,n)*Gamma(n,j);
// }
// }
for (int l=0; l<NumOrbitals; l++)
for (int j=0; j<NumBasis; j++)
{
G_gamma(l,j) = BasisGrad(l,j);
for (int n=0; n<NumOrbitals; n++)
G_gamma(l,j) -= dpsiM(l,n)*Gamma(n,j);
}
// Now, compute d/dC_{ij} lapl(det)/det by multiplying by Ainv
BLAS::gemm('N', 'T', NumBasis, NumOrbitals, NumOrbitals, -0.5,
L_gamma.data(), NumBasis, psiM.data(), NumOrbitals,
0.0, dlapl_dC.data(), NumBasis);
for (int ptcl=0; ptcl<NumOrbitals; ptcl++)
{
MyG[ptcl] = PosType();
for (int orb=0; orb<NumOrbitals; orb++)
MyG[ptcl] += dpsiM(ptcl,orb)*psiM(ptcl,orb);
// fprintf (stderr, "myG = %11.4e %11.4e %11.4e\n",
// myG[ptcl][0], myG[ptcl][1], myG[ptcl][2]);
// fprintf (stderr, "MyG = %11.4e %11.4e %11.4e\n",
// MyG[ptcl][0], MyG[ptcl][1], MyG[ptcl][2]);
// fprintf (stderr, "P.G = %11.4e %11.4e %11.4e\n",
// P.G[ptcl+FirstIndex][0],
// P.G[ptcl+FirstIndex][1],
// P.G[ptcl+FirstIndex][2]);
}
for (int i=0; i<NumOrbitals; i++)
for (int j=0; j<NumBasis; j++)
for (int l=0; l<NumOrbitals; l++)
{
GradType g = P.G[FirstIndex+l];
GradType dg1 = psiM(l,i)*G_gamma(l,j);
GradType dg2 = g*dlogdet_dC(i,j);
GradType dg = dg1;// - dg2;
// fprintf (stderr, "dg = %9.4f %9.4f %9.4f\n",
// dg[0], dg[1], dg[2]);
// fprintf (stderr, "g1 = %11.4e %11.4e %11.4e g2 = %11.4e %11.4e %11.4e\n",
// dg1[0], dg1[1], dg1[2], dg2[0], dg2[1], dg2[2]);
g -= MyG[l];
dlapl_dC(i,j) -= dot(g, dg);
}
// for (int i=0; i<NumOrbitals; i++)
// for (int j=0; j<NumBasis; j++) {
// dlapl_dC(i,j) = 0.0;
// for (int l=0; l<NumOrbitals; l++) {
// // dlapl_dC(i,j) += psiM(n,i)*L_gamma(n,j);
// // dlapl_dC(i,j) += psiM(l,i)*L_gamma(l,j);
// dlapl_dC(i,j) -= 0.5*psiM(l,i)*L_gamma(l,j);
// //dlapl_dC(i,j) += psiM(l,i)*BasisLapl(l,j);
// }
// }
// Pull elements from dense d_dC matrices and put into parameter
// derivatives, dlogpsi and dhpsioverpsi
Phi->copyParamsFromMatrix(active, dlogdet_dC, dlogpsi);
Phi->copyParamsFromMatrix(active, dlapl_dC, dhpsioverpsi);
}
