void IceModel_Decode::run_timestep(double time_s,
blitz::Array<int,1> const &indices,
std::map<IceField, blitz::Array<double,1>> const &vals2)
{
printf("BEGIN IceModel_Decode::run_timestep(%f) size=%ld\n", time_s, indices.size());
std::map<IceField, blitz::Array<double,1>> vals2d; /// Decoded fields
// Loop through the fields we require
std::set<IceField> fields;
get_required_fields(fields);
for (auto field = fields.begin(); field != fields.end(); ++field) {
printf("Looking for required field %s\n", field->str());
// Look up the field we require
auto ii = vals2.find(*field);
if (ii == vals2.end()) {
fprintf(stderr, "Cannot find required ice field = %s\n", field->str());
throw std::exception();
}
blitz::Array<double,1> vals(ii->second);
// Decode the field!
blitz::Array<double,1> valsd(ndata());
valsd = nan;
int n = indices.size();
for (int i=0; i < n; ++i) {
int ix = indices(i);
// Do our own bounds checking!
if (ix < 0 || ix >= ndata()) {
fprintf(stderr, "IceModel: index %d out of range [0, %d)\n", ix, ndata());
throw std::exception();
}
#if 0
// Sanity check for NaN coming through
if (std::isnan(vals(i))) {
fprintf(stderr, "IceModel::decode: vals[%d] (index=%d) is NaN!\n", i, ix);
throw std::exception();
}
#endif
// Add this value to existing field
double &oval = valsd(ix);
if (std::isnan(oval)) oval = vals(i);
else oval += vals(i);
}
// Store decoded field in our output
vals2d.insert(std::make_pair(*field, valsd));
printf("Done decoding required field, %s\n", field->str());
}
// Pass decoded fields on to subclass
run_decoded(time_s, vals2d);
printf("END IceModel_Decode::run_timestep(%ld)\n", time_s);
}
double VectorInnerProduct(const blitz::Array<double, Rank> &u, const blitz::Array<double, Rank> &v)
{
if (u.size() != v.size())
{
cout << "Vector u and v is of different size: " << u.size() << " != " << v.size() << endl;
throw std::runtime_error("invalid vector sizes for inner product");
}
if (!u.isStorageContiguous())
{
throw std::runtime_error("Vector u is not contiguous");
}
if (!v.isStorageContiguous())
{
throw std::runtime_error("Vector v is not contiguous");
}
int N = u.size();
int uStride = 1; //u.stride(0);
int vStride = 1; //v.stride(0);
return BLAS_NAME(ddot)(N, (double*)u.data(), uStride, (double*)v.data(), vStride);
}
cplx VectorInnerProduct(const blitz::Array<cplx, Rank> &u, const blitz::Array<cplx, Rank> &v)
{
if (u.size() != v.size())
{
cout << "Vector u and v is of different size: " << u.size() << " != " << v.size() << endl;
throw std::runtime_error("invalid vector sizes for inner product");
}
if (!u.isStorageContiguous())
{
throw std::runtime_error("Vector u is not contiguous");
}
if (!v.isStorageContiguous())
{
throw std::runtime_error("Vector v is not contiguous");
}
int N = u.size();
int uStride = 1; //u.stride(0);
int vStride = 1; //v.stride(0);
cplx result;
acml::doublecomplex ret = BLAS_NAME(zdotc)(N, (acml::doublecomplex*)u.data(), uStride, (acml::doublecomplex*)v.data(), vStride);
result = cplx(ret.real, ret.imag);
return result;
}
/*
* Sets the radial Coulomb wave F_l(k*r, eta), with eta = Z/k into data for all radial
* grid points specified by r
*/
void SetRadialCoulombWave(int Z, int l, double k, blitz::Array<double, 1> r, blitz::Array<double, 1> data)
{
double eta = Z / k;
for (int i=0; i<r.size(); i++)
{
double x = k * r(i);
gsl_sf_result F, Fp, G, Gp;
double exp_F, exp_G;
int error = gsl_sf_coulomb_wave_FG_e(eta, x, (double)l, 0., &F, &Fp, &G, &Gp, &exp_F, &exp_G);
if (error == GSL_EOVRFLW)
{
cout << "WARNING: Overflow in SetCoulombWave(" << Z << ", " << l << ", " << k << ", r=" << r(i) << ");" << endl;
cout << " exp_F = " << exp_F << ", exp_G = " << exp_G << endl;
}
data(i) = F.val;
}
}
void CopyTensorPotentialToEpetraMatrix(Epetra_FECrsMatrix_Ptr epetraMatrix, blitz::Array<cplx, Rank> potentialData, list pyLocalBasisPairs, blitz::TinyVector<int, Rank> globalStrides, double cutoff)
{
blitz::Array<int, 2> indexArray;
indexArray.resize(4 * potentialData.size(), 2);
indexArray = -100;
double sqrCutoff = sqr(cutoff);
//Setup structures for calculating matrix row/col indices from the
//basis pairs in the tensor potential
blitz::TinyVector< blitz::Array<int, 2>, Rank > localBasisPairs;
for (int rank=0; rank<Rank; rank++)
{
localBasisPairs(rank).reference( boost::python::extract< blitz::Array<int, 2> >(pyLocalBasisPairs[rank]) );
}
//Iterate over all items in potentialData
typename blitz::Array<cplx, Rank>::iterator it = potentialData.begin();
for (int linearCount=0; linearCount<potentialData.size(); linearCount++)
{
int globalRow = 0;
int globalCol = 0;
for (int rank=0; rank<Rank; rank++)
{
int rankPos = it.position()(rank);
globalRow += globalStrides(rank) * localBasisPairs(rank)(rankPos, 0);
globalCol += globalStrides(rank) * localBasisPairs(rank)(rankPos, 1);
}
double realVal = real(*it);
double imagVal = imag(*it);
//Skip padded elements (they have negative row/col index)
if ((globalRow < 0) || (globalCol < 0))
{
it++;
continue;
}
/*
* Because epetra does not support complex natively,
* each matrix element is a 2x2 block
*
* (A_r -A_i ) (c_r) = (A_r + i A_i) * (c_r + i c_i) = A * c
* (A_i A_r ) (c_i)
*
* Detect if A_i or A_r is zero to avoid redundant elements
*/
//Insert values into matrix
if (sqr(realVal) > sqrCutoff)
{
int r,c;
r = 2*globalRow; c = 2*globalCol;
indexArray(4*linearCount, 0) = r;
indexArray(4*linearCount, 1) = c;
epetraMatrix->InsertGlobalValues(r, 1, &realVal, &c);
r++; c++;
epetraMatrix->InsertGlobalValues(r, 1, &realVal, &c);
indexArray(4*linearCount+1, 0) = r;
indexArray(4*linearCount+1, 1) = c;
}
if (sqr(imagVal) > sqrCutoff)
{
int r,c;
//Upper row, A_i with minus sign
r = 2*globalRow; c = 2*globalCol+1;
indexArray(4*linearCount+2, 0) = r;
indexArray(4*linearCount+2, 1) = c;
imagVal = -imagVal;
epetraMatrix->InsertGlobalValues(r, 1, &imagVal, &c);
//Lower row, A_i without minus sign
imagVal = -imagVal;
epetraMatrix->InsertGlobalValues(c, 1, &imagVal, &r);
indexArray(4*linearCount+3, 0) = c;
indexArray(4*linearCount+3, 1) = r;
}
++it;
}
}