portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-06-11 03:46:55 +01:00
Code Releases Activity

Merge branch 'develop' into feature/dwf-multirhs

Browse Source
This commit is contained in:
paboyle
2017-10-02 11:41:49 +01:00
parent ac740f73ce fddeb29d6b
commit 4f8b6f26b4
160 changed files with 23407 additions and 9004 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
lib/algorithms/Algorithms.h
View File

@ -1,6 +1,6 @@
/*************************************************************************************
Grid physics library, www.github.com/paboyle/Grid
Grid physics library, www.github.com/paboyle/Grid
Source file: ./lib/Algorithms.h
@ -37,6 +37,7 @@ Author: Peter Boyle <paboyle@ph.ed.ac.uk>
#include <Grid/algorithms/approx/Chebyshev.h>
#include <Grid/algorithms/approx/Remez.h>
#include <Grid/algorithms/approx/MultiShiftFunction.h>
#include <Grid/algorithms/approx/Forecast.h>
#include <Grid/algorithms/iterative/ConjugateGradient.h>
#include <Grid/algorithms/iterative/ConjugateResidual.h>
@ -45,6 +46,7 @@ Author: Peter Boyle <paboyle@ph.ed.ac.uk>
#include <Grid/algorithms/iterative/ConjugateGradientMultiShift.h>
#include <Grid/algorithms/iterative/ConjugateGradientMixedPrec.h>
#include <Grid/algorithms/iterative/BlockConjugateGradient.h>
#include <Grid/algorithms/iterative/ConjugateGradientReliableUpdate.h>
#include <Grid/algorithms/iterative/ImplicitlyRestartedLanczos.h>
#include <Grid/algorithms/CoarsenedMatrix.h>
#include <Grid/algorithms/FFT.h>

5
lib/algorithms/FFT.h
View File

@ -230,6 +230,7 @@ namespace Grid {
// Barrel shift and collect global pencil
std::vector<int> lcoor(Nd), gcoor(Nd);
result = source;
int pc = processor_coor[dim];
for(int p=0;p<processors[dim];p++) {
PARALLEL_REGION
{
@ -240,7 +241,8 @@ namespace Grid {
for(int idx=0;idx<sgrid->lSites();idx++) {
sgrid->LocalIndexToLocalCoor(idx,cbuf);
peekLocalSite(s,result,cbuf);
cbuf[dim]+=p*L;
cbuf[dim]+=((pc+p) % processors[dim])*L;
// cbuf[dim]+=p*L;
pokeLocalSite(s,pgbuf,cbuf);
}
}
@ -278,7 +280,6 @@ namespace Grid {
flops+= flops_call*NN;
// writing out result
int pc = processor_coor[dim];
PARALLEL_REGION
{
std::vector<int> clbuf(Nd), cgbuf(Nd);

152
lib/algorithms/approx/Forecast.h Normal file
View File

@ -0,0 +1,152 @@
/*************************************************************************************
Grid physics library, www.github.com/paboyle/Grid
Source file: ./lib/algorithms/approx/Forecast.h
Copyright (C) 2015
Author: Peter Boyle <paboyle@ph.ed.ac.uk>
Author: paboyle <paboyle@ph.ed.ac.uk>
Author: David Murphy <dmurphy@phys.columbia.edu>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
See the full license in the file "LICENSE" in the top level distribution directory
*************************************************************************************/
/* END LEGAL */
#ifndef INCLUDED_FORECAST_H
#define INCLUDED_FORECAST_H
namespace Grid {
// Abstract base class.
// Takes a matrix (Mat), a source (phi), and a vector of Fields (chi)
// and returns a forecasted solution to the system D*psi = phi (psi).
template<class Matrix, class Field>
class Forecast
{
public:
virtual Field operator()(Matrix &Mat, const Field& phi, const std::vector<Field>& chi) = 0;
};
// Implementation of Brower et al.'s chronological inverter (arXiv:hep-lat/9509012),
// used to forecast solutions across poles of the EOFA heatbath.
//
// Modified from CPS (cps_pp/src/util/dirac_op/d_op_base/comsrc/minresext.C)
template<class Matrix, class Field>
class ChronoForecast : public Forecast<Matrix,Field>
{
public:
Field operator()(Matrix &Mat, const Field& phi, const std::vector<Field>& prev_solns)
{
int degree = prev_solns.size();
Field chi(phi); // forecasted solution
// Trivial cases
if(degree == 0){ chi = zero; return chi; }
else if(degree == 1){ return prev_solns[0]; }
RealD dot;
ComplexD xp;
Field r(phi); // residual
Field Mv(phi);
std::vector<Field> v(prev_solns); // orthonormalized previous solutions
std::vector<Field> MdagMv(degree,phi);
// Array to hold the matrix elements
std::vector<std::vector<ComplexD>> G(degree, std::vector<ComplexD>(degree));
// Solution and source vectors
std::vector<ComplexD> a(degree);
std::vector<ComplexD> b(degree);
// Orthonormalize the vector basis
for(int i=0; i<degree; i++){
v[i] *= 1.0/std::sqrt(norm2(v[i]));
for(int j=i+1; j<degree; j++){ v[j] -= innerProduct(v[i],v[j]) * v[i]; }
}
// Perform sparse matrix multiplication and construct rhs
for(int i=0; i<degree; i++){
b[i] = innerProduct(v[i],phi);
Mat.M(v[i],Mv);
Mat.Mdag(Mv,MdagMv[i]);
G[i][i] = innerProduct(v[i],MdagMv[i]);
}
// Construct the matrix
for(int j=0; j<degree; j++){
for(int k=j+1; k<degree; k++){
G[j][k] = innerProduct(v[j],MdagMv[k]);
G[k][j] = std::conj(G[j][k]);
}}
// Gauss-Jordan elimination with partial pivoting
for(int i=0; i<degree; i++){
// Perform partial pivoting
int k = i;
for(int j=i+1; j<degree; j++){ if(std::abs(G[j][j]) > std::abs(G[k][k])){ k = j; } }
if(k != i){
xp = b[k];
b[k] = b[i];
b[i] = xp;
for(int j=0; j<degree; j++){
xp = G[k][j];
G[k][j] = G[i][j];
G[i][j] = xp;
}
}
// Convert matrix to upper triangular form
for(int j=i+1; j<degree; j++){
xp = G[j][i]/G[i][i];
b[j] -= xp * b[i];
for(int k=0; k<degree; k++){ G[j][k] -= xp*G[i][k]; }
}
}
// Use Gaussian elimination to solve equations and calculate initial guess
chi = zero;
r = phi;
for(int i=degree-1; i>=0; i--){
a[i] = 0.0;
for(int j=i+1; j<degree; j++){ a[i] += G[i][j] * a[j]; }
a[i] = (b[i]-a[i])/G[i][i];
chi += a[i]*v[i];
r -= a[i]*MdagMv[i];
}
RealD true_r(0.0);
ComplexD tmp;
for(int i=0; i<degree; i++){
tmp = -b[i];
for(int j=0; j<degree; j++){ tmp += G[i][j]*a[j]; }
tmp = std::conj(tmp)*tmp;
true_r += std::sqrt(tmp.real());
}
RealD error = std::sqrt(norm2(r)/norm2(phi));
std::cout << GridLogMessage << "ChronoForecast: |res|/|src| = " << error << std::endl;
return chi;
};
};
}
#endif

29
lib/algorithms/iterative/BlockConjugateGradient.h
View File

@ -87,15 +87,22 @@ void ThinQRfact (Eigen::MatrixXcd &m_rr,
////////////////////////////////////////////////////////////////////////////////////////////////////
sliceInnerProductMatrix(m_rr,R,R,Orthog);
////////////////////////////////////////////////////////////////////////////////////////////////////
// Cholesky from Eigen
// There exists a ldlt that is documented as more stable
////////////////////////////////////////////////////////////////////////////////////////////////////
Eigen::MatrixXcd L = m_rr.llt().matrixL();
// Force manifest hermitian to avoid rounding related
m_rr = 0.5*(m_rr+m_rr.adjoint());
#if 0
std::cout << " Calling Cholesky ldlt on m_rr " << m_rr <<std::endl;
Eigen::MatrixXcd L_ldlt = m_rr.ldlt().matrixL();
std::cout << " Called Cholesky ldlt on m_rr " << L_ldlt <<std::endl;
auto D_ldlt = m_rr.ldlt().vectorD();
std::cout << " Called Cholesky ldlt on m_rr " << D_ldlt <<std::endl;
#endif
// std::cout << " Calling Cholesky llt on m_rr " <<std::endl;
Eigen::MatrixXcd L = m_rr.llt().matrixL();
// std::cout << " Called Cholesky llt on m_rr " << L <<std::endl;
C = L.adjoint();
Cinv = C.inverse();
////////////////////////////////////////////////////////////////////////////////////////////////////
// Q = R C^{-1}
//
@ -103,7 +110,6 @@ void ThinQRfact (Eigen::MatrixXcd &m_rr,
//
// NB maddMatrix conventions are Right multiplication X[j] a[j,i] already
////////////////////////////////////////////////////////////////////////////////////////////////////
// FIXME:: make a sliceMulMatrix to avoid zero vector
sliceMulMatrix(Q,Cinv,R,Orthog);
}
////////////////////////////////////////////////////////////////////////////////////////////////////
@ -199,7 +205,12 @@ void BlockCGrQsolve(LinearOperatorBase<Field> &Linop, const Field &B, Field &X)
Linop.HermOp(X, AD);
tmp = B - AD;
//std::cout << GridLogMessage << " initial tmp " << norm2(tmp)<< std::endl;
ThinQRfact (m_rr, m_C, m_Cinv, Q, tmp);
//std::cout << GridLogMessage << " initial Q " << norm2(Q)<< std::endl;
//std::cout << GridLogMessage << " m_rr " << m_rr<<std::endl;
//std::cout << GridLogMessage << " m_C " << m_C<<std::endl;
//std::cout << GridLogMessage << " m_Cinv " << m_Cinv<<std::endl;
D=Q;
std::cout << GridLogMessage<<"BlockCGrQ computed initial residual and QR fact " <<std::endl;
@ -221,13 +232,15 @@ void BlockCGrQsolve(LinearOperatorBase<Field> &Linop, const Field &B, Field &X)
MatrixTimer.Start();
Linop.HermOp(D, Z);
MatrixTimer.Stop();
//std::cout << GridLogMessage << " norm2 Z " <<norm2(Z)<<std::endl;
//4. M = [D^dag Z]^{-1}
sliceInnerTimer.Start();
sliceInnerProductMatrix(m_DZ,D,Z,Orthog);
sliceInnerTimer.Stop();
m_M = m_DZ.inverse();
//std::cout << GridLogMessage << " m_DZ " <<m_DZ<<std::endl;
//5. X = X + D MC
m_tmp = m_M * m_C;
sliceMaddTimer.Start();

256
lib/algorithms/iterative/ConjugateGradientReliableUpdate.h Normal file
View File

@ -0,0 +1,256 @@
/*************************************************************************************
Grid physics library, www.github.com/paboyle/Grid
Source file: ./lib/algorithms/iterative/ConjugateGradientReliableUpdate.h
Copyright (C) 2015
Author: Christopher Kelly <ckelly@phys.columbia.edu>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
See the full license in the file "LICENSE" in the top level distribution directory
*************************************************************************************/
/* END LEGAL */
#ifndef GRID_CONJUGATE_GRADIENT_RELIABLE_UPDATE_H
#define GRID_CONJUGATE_GRADIENT_RELIABLE_UPDATE_H
namespace Grid {
template<class FieldD,class FieldF, typename std::enable_if< getPrecision<FieldD>::value == 2, int>::type = 0,typename std::enable_if< getPrecision<FieldF>::value == 1, int>::type = 0>
class ConjugateGradientReliableUpdate : public LinearFunction<FieldD> {
public:
bool ErrorOnNoConverge; // throw an assert when the CG fails to converge.
// Defaults true.
RealD Tolerance;
Integer MaxIterations;
Integer IterationsToComplete; //Number of iterations the CG took to finish. Filled in upon completion
Integer ReliableUpdatesPerformed;
bool DoFinalCleanup; //Final DP cleanup, defaults to true
Integer IterationsToCleanup; //Final DP cleanup step iterations
LinearOperatorBase<FieldF> &Linop_f;
LinearOperatorBase<FieldD> &Linop_d;
GridBase* SinglePrecGrid;
RealD Delta; //reliable update parameter
//Optional ability to switch to a different linear operator once the tolerance reaches a certain point. Useful for single/half -> single/single
LinearOperatorBase<FieldF> *Linop_fallback;
RealD fallback_transition_tol;
ConjugateGradientReliableUpdate(RealD tol, Integer maxit, RealD _delta, GridBase* _sp_grid, LinearOperatorBase<FieldF> &_Linop_f, LinearOperatorBase<FieldD> &_Linop_d, bool err_on_no_conv = true)
: Tolerance(tol),
MaxIterations(maxit),
Delta(_delta),
Linop_f(_Linop_f),
Linop_d(_Linop_d),
SinglePrecGrid(_sp_grid),
ErrorOnNoConverge(err_on_no_conv),
DoFinalCleanup(true),
Linop_fallback(NULL)
{};
void setFallbackLinop(LinearOperatorBase<FieldF> &_Linop_fallback, const RealD _fallback_transition_tol){
Linop_fallback = &_Linop_fallback;
fallback_transition_tol = _fallback_transition_tol;
}
void operator()(const FieldD &src, FieldD &psi) {
LinearOperatorBase<FieldF> *Linop_f_use = &Linop_f;
bool using_fallback = false;
psi.checkerboard = src.checkerboard;
conformable(psi, src);
RealD cp, c, a, d, b, ssq, qq, b_pred;
FieldD p(src);
FieldD mmp(src);
FieldD r(src);
// Initial residual computation & set up
RealD guess = norm2(psi);
assert(std::isnan(guess) == 0);
Linop_d.HermOpAndNorm(psi, mmp, d, b);
r = src - mmp;
p = r;
a = norm2(p);
cp = a;
ssq = norm2(src);
std::cout << GridLogIterative << std::setprecision(4) << "ConjugateGradientReliableUpdate: guess " << guess << std::endl;
std::cout << GridLogIterative << std::setprecision(4) << "ConjugateGradientReliableUpdate: src " << ssq << std::endl;
std::cout << GridLogIterative << std::setprecision(4) << "ConjugateGradientReliableUpdate: mp " << d << std::endl;
std::cout << GridLogIterative << std::setprecision(4) << "ConjugateGradientReliableUpdate: mmp " << b << std::endl;
std::cout << GridLogIterative << std::setprecision(4) << "ConjugateGradientReliableUpdate: cp,r " << cp << std::endl;
std::cout << GridLogIterative << std::setprecision(4) << "ConjugateGradientReliableUpdate: p " << a << std::endl;
RealD rsq = Tolerance * Tolerance * ssq;
// Check if guess is really REALLY good :)
if (cp <= rsq) {
std::cout << GridLogMessage << "ConjugateGradientReliableUpdate guess was REALLY good\n";
std::cout << GridLogMessage << "\tComputed residual " << sqrt(cp / ssq)<<std::endl;
return;
}
//Single prec initialization
FieldF r_f(SinglePrecGrid);
r_f.checkerboard = r.checkerboard;
precisionChange(r_f, r);
FieldF psi_f(r_f);
psi_f = zero;
FieldF p_f(r_f);
FieldF mmp_f(r_f);
RealD MaxResidSinceLastRelUp = cp; //initial residual
std::cout << GridLogIterative << std::setprecision(4)
<< "ConjugateGradient: k=0 residual " << cp << " target " << rsq << std::endl;
GridStopWatch LinalgTimer;
GridStopWatch MatrixTimer;
GridStopWatch SolverTimer;
SolverTimer.Start();
int k = 0;
int l = 0;
for (k = 1; k <= MaxIterations; k++) {
c = cp;
MatrixTimer.Start();
Linop_f_use->HermOpAndNorm(p_f, mmp_f, d, qq);
MatrixTimer.Stop();
LinalgTimer.Start();
a = c / d;
b_pred = a * (a * qq - d) / c;
cp = axpy_norm(r_f, -a, mmp_f, r_f);
b = cp / c;
// Fuse these loops ; should be really easy
psi_f = a * p_f + psi_f;
//p_f = p_f * b + r_f;
LinalgTimer.Stop();
std::cout << GridLogIterative << "ConjugateGradientReliableUpdate: Iteration " << k
<< " residual " << cp << " target " << rsq << std::endl;
std::cout << GridLogDebug << "a = "<< a << " b_pred = "<< b_pred << " b = "<< b << std::endl;
std::cout << GridLogDebug << "qq = "<< qq << " d = "<< d << " c = "<< c << std::endl;
if(cp > MaxResidSinceLastRelUp){
std::cout << GridLogIterative << "ConjugateGradientReliableUpdate: updating MaxResidSinceLastRelUp : " << MaxResidSinceLastRelUp << " -> " << cp << std::endl;
MaxResidSinceLastRelUp = cp;
}
// Stopping condition
if (cp <= rsq) {
//Although not written in the paper, I assume that I have to add on the final solution
precisionChange(mmp, psi_f);
psi = psi + mmp;
SolverTimer.Stop();
Linop_d.HermOpAndNorm(psi, mmp, d, qq);
p = mmp - src;
RealD srcnorm = sqrt(norm2(src));
RealD resnorm = sqrt(norm2(p));
RealD true_residual = resnorm / srcnorm;
std::cout << GridLogMessage << "ConjugateGradientReliableUpdate Converged on iteration " << k << " after " << l << " reliable updates" << std::endl;
std::cout << GridLogMessage << "\tComputed residual " << sqrt(cp / ssq)<<std::endl;
std::cout << GridLogMessage << "\tTrue residual " << true_residual<<std::endl;
std::cout << GridLogMessage << "\tTarget " << Tolerance << std::endl;
std::cout << GridLogMessage << "Time breakdown "<<std::endl;
std::cout << GridLogMessage << "\tElapsed " << SolverTimer.Elapsed() <<std::endl;
std::cout << GridLogMessage << "\tMatrix " << MatrixTimer.Elapsed() <<std::endl;
std::cout << GridLogMessage << "\tLinalg " << LinalgTimer.Elapsed() <<std::endl;
IterationsToComplete = k;
ReliableUpdatesPerformed = l;
if(DoFinalCleanup){
//Do a final CG to cleanup
std::cout << GridLogMessage << "ConjugateGradientReliableUpdate performing final cleanup.\n";
ConjugateGradient<FieldD> CG(Tolerance,MaxIterations);
CG.ErrorOnNoConverge = ErrorOnNoConverge;
CG(Linop_d,src,psi);
IterationsToCleanup = CG.IterationsToComplete;
}
else if (ErrorOnNoConverge) assert(true_residual / Tolerance < 10000.0);
std::cout << GridLogMessage << "ConjugateGradientReliableUpdate complete.\n";
return;
}
else if(cp < Delta * MaxResidSinceLastRelUp) { //reliable update
std::cout << GridLogMessage << "ConjugateGradientReliableUpdate "
<< cp << "(residual) < " << Delta << "(Delta) * " << MaxResidSinceLastRelUp << "(MaxResidSinceLastRelUp) on iteration " << k << " : performing reliable update\n";
precisionChange(mmp, psi_f);
psi = psi + mmp;
Linop_d.HermOpAndNorm(psi, mmp, d, qq);
r = src - mmp;
psi_f = zero;
precisionChange(r_f, r);
cp = norm2(r);
MaxResidSinceLastRelUp = cp;
b = cp/c;
std::cout << GridLogMessage << "ConjugateGradientReliableUpdate new residual " << cp << std::endl;
l = l+1;
}
p_f = p_f * b + r_f; //update search vector after reliable update appears to help convergence
if(!using_fallback && Linop_fallback != NULL && cp < fallback_transition_tol){
std::cout << GridLogMessage << "ConjugateGradientReliableUpdate switching to fallback linear operator on iteration " << k << " at residual " << cp << std::endl;
Linop_f_use = Linop_fallback;
using_fallback = true;
}
}
std::cout << GridLogMessage << "ConjugateGradientReliableUpdate did NOT converge"
<< std::endl;
if (ErrorOnNoConverge) assert(0);
IterationsToComplete = k;
ReliableUpdatesPerformed = l;
}
};
};
#endif