Page MenuHomeHEPForge
Files F10881184

No OneTemporary
Actions

View Options

Index: trunk/src/CrossSection.cpp
===================================================================
--- trunk/src/CrossSection.cpp (revision 356)
+++ trunk/src/CrossSection.cpp (revision 357)
@@ -1,910 +1,911 @@
//==============================================================================
// CrossSection.cpp
//
// Copyright (C) 2010-2018 Tobias Toll and Thomas Ullrich
//
// This file is part of Sartre.
//
// 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.
// 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, see <http://www.gnu.org/licenses/>.
//
// Author: Thomas Ullrich
// Last update:
// $Date$
// $Author$
//==============================================================================
#include "CrossSection.h"
#include "EventGeneratorSettings.h"
#include "Kinematics.h"
#include "TableCollection.h"
#include "Table.h"
#include "Constants.h"
#include "TH2F.h"
#include "TFile.h"
#include "DglapEvolution.h"
#include <cstdio>
#include <cmath>
#include <limits>
#define PR(x) cout << #x << " = " << (x) << endl;
CrossSection::CrossSection(TableCollection* tc, TableCollection* ptc)
{
mSettings = EventGeneratorSettings::instance();
mRandom = mSettings->randomGenerator();
mS = Kinematics::s(mSettings->eBeam(), mSettings->hBeam());
mPhotonFlux.setS(mS);
mTableCollection = tc;
mProtonTableCollection = ptc;
TParticlePDG *vectorMesonPDG = mSettings->lookupPDG(mSettings->vectorMesonId());
mVmMass = vectorMesonPDG->Mass();
mCheckKinematics = true;
}
CrossSection::~CrossSection() { /* no op */ }
void CrossSection::setTableCollection(TableCollection* tc) {mTableCollection = tc;}
void CrossSection::setProtonTableCollection(TableCollection* ptc) {mProtonTableCollection = ptc;}
void CrossSection::setCheckKinematics(bool val) {mCheckKinematics = val;}
double CrossSection::operator()(double t, double Q2, double W2)
{
return dsigdtdQ2dW2_total_checked(t, Q2, W2);
}
// UPC Version
double CrossSection::operator()(double t, double xpom)
{
return dsigdtdxp_total_checked(t, xpom);
}
double CrossSection::operator()(const double* array)
{
if (mSettings->UPC())
return dsigdtdxp_total_checked(array[0], array[1]);
else
return dsigdtdQ2dW2_total_checked(array[0], array[1], array[2]);
}
//
// PDF passed to UNURAN
// Array holds: t, log(Q2), W2 for e+p/A running
// t, log(xpom) for UPC
//
double CrossSection::unuranPDF(const double* array) // array is t, log(Q2), W2
{ // or t and log(xpom)
double result = 0;
if (mSettings->UPC()) {
double xpom = exp(array[1]);
result = dsigdtdxp_total_checked(array[0], xpom); // t, xpom
result *= xpom; // Jacobian
}
else {
double Q2 = exp(array[1]);
result = dsigdtdQ2dW2_total_checked(array[0], Q2, array[2]); // t, Q2, W2
result *= Q2; // Jacobian
}
return log(result);
}
double CrossSection::dsigdtdQ2dW2_total_checked(double t, double Q2, double W2)
{
double result = 0;
//
// Check if in kinematically allowed region
//
if (mCheckKinematics && !Kinematics::valid(mS, t, Q2, W2, mVmMass, false, (mSettings->verboseLevel() > 2) )) {
if (mSettings->verboseLevel() > 2)
cout << "CrossSection::dsigdtdQ2dW2_total_checked(): warning, t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2)
<< " is outside of kinematically allowed region. Return 0." << endl;
return result;
}
//
// Total cross-section dsig2/(dQ2 dW2 dt)
// This is the probability density function needed for UNU.RAN
//
double csT = dsigdtdQ2dW2_total(t, Q2, W2, transverse);
double csL = dsigdtdQ2dW2_total(t, Q2, W2, longitudinal);
result = csT + csL;
//
// Polarization
//
if (mRandom->Uniform(result) <= csT)
mPolarization = transverse;
else
mPolarization = longitudinal;
//
// Diffractive Mode
//
double sampleRange = (mPolarization == transverse ? csT : csL);
double sampleDivider = dsigdtdQ2dW2_coherent(t, Q2, W2, mPolarization);
if (mRandom->Uniform(sampleRange) <= sampleDivider)
mDiffractiveMode = coherent;
else
mDiffractiveMode = incoherent;
//
// Print-out at high verbose levels
//
if (mSettings->verboseLevel() > 10) { // Spinal Tap ;-)
cout << "CrossSection::dsigdtdQ2dW2_total_checked(): " << result;
cout << " at t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2);
cout << " (" << (mPolarization == transverse ? "transverse" : "longitudinal");
cout << " ," << (mDiffractiveMode == coherent ? "coherent" : "incoherent");
cout << ')' << endl;
}
//
// Check validity of return value
//
if (std::isnan(result)) {
cout << "CrossSection::dsigdtdQ2dW2_total_checked(): Error, return value = NaN at" << endl;
cout << " t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2) << endl;
result = 0;
}
if (std::isinf(result)) {
cout << "CrossSection::dsigdtdQ2dW2_total_checked(): Error, return value = inf at" << endl;
cout << " t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2) << endl;
result = 0;
}
if (result < 0) {
cout << "CrossSection::dsigdtdQ2dW2_total_checked(): Error, negative cross-section at" << endl;
cout << " t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2) << endl;
result = 0;
}
return result;
}
double CrossSection::dsigdtdxp_total_checked(double t, double xpom)
{
double result = 0;
//
// Check if in kinematically allowed region
//
if (mCheckKinematics && !Kinematics::validUPC(mSettings->hadronBeamEnergy(),
mSettings->electronBeamEnergy(),
t, xpom, mVmMass,
(mSettings->verboseLevel() > 2) )) {
if (mSettings->verboseLevel() > 2)
cout << "CrossSection::dsigdtdxp_total_checked(): warning, t=" << t << ", xpom=" << xpom
<< " is outside of kinematically allowed region. Return 0." << endl;
return result;
}
//
// Total cross-section dsig2/(dt dxp)
// This is the probability density function needed for UNU.RAN
//
result = dsigdtdxp_total(t, xpom);
//
// Polarization
//
mPolarization = transverse; // always
//
// Diffractive Mode
//
double sampleRange = result;
double sampleDivider = dsigdtdxp_coherent(t, xpom);
if (mRandom->Uniform(sampleRange) <= sampleDivider)
mDiffractiveMode = coherent;
else
mDiffractiveMode = incoherent;
//
// Print-out at high verbose levels
//
if (mSettings->verboseLevel() > 10) {
cout << "CrossSection::dsigdtdxp_total_checked(): " << result;
cout << " at t=" << t << ", xp=" << xpom;
cout << " (" << (mDiffractiveMode == coherent ? "coherent" : "incoherent");
cout << ')' << endl;
}
//
// Check validity of return value
//
if (std::isnan(result)) {
cout << "CrossSection::dsigdtdxp_total_checked(): Error, return value = NaN at" << endl;
cout << " t=" << t << ", xp=" << xpom << endl;
result = 0;
}
if (std::isinf(result)) {
cout << "CrossSection::dsigdtdxp_total_checked(): Error, return value = inf at" << endl;
cout << " t=" << t << ", xp=" << xpom << endl;
result = 0;
}
if (result < 0) {
cout << "CrossSection::dsigdtdxp_total_checked(): Error, negative cross-section at" << endl;
cout << " t=" << t << ", xp=" << xpom << endl;
result = 0;
}
return result;
}
double CrossSection::dsigdt_total(double t, double Q2, double W2, GammaPolarization pol) const
{
double result = 0;
if (mSettings->tableSetType() == coherent_and_incoherent) {
result = dsigdt_coherent(t, Q2, W2, pol) + dsigdt_incoherent(t, Q2, W2, pol);
}
else if (mSettings->tableSetType() == total_and_coherent) {
result = mTableCollection->get(Q2, W2, t, pol, mean_A2); // units now are fm^4
result /= (16*M_PI);
result /= hbarc2; // in fm^2/GeV^2
result *= 1e7; // in nb/GeV^2
double lambda = 0;
if (mSettings->correctForRealAmplitude()){
lambda = logDerivateOfAmplitude(t, Q2, W2, pol);
result *= realAmplitudeCorrection(lambda);
}
if (mSettings->correctSkewedness()){
lambda = logDerivateOfGluonDensity(t, Q2, W2, pol);
result *= skewednessCorrection(lambda);
}
}
return result;
}
//
// UPC Version
//
double CrossSection::dsigdt_total(double t, double xpom) const
{
double result = 0;
if (mSettings->tableSetType() == coherent_and_incoherent) {
result = dsigdt_coherent(t, xpom) + dsigdt_incoherent(t, xpom);
}
else if (mSettings->tableSetType() == total_and_coherent) {
result = mTableCollection->get(xpom, t, mean_A2); // units now are fm^4
result /= (16*M_PI);
result /= hbarc2; // in fm^2/GeV^2
result *= 1e7; // in nb/GeV^2
double lambda = 0;
if (mSettings->correctForRealAmplitude()){
lambda = logDerivateOfAmplitude(t, xpom);
result *= realAmplitudeCorrection(lambda);
}
if (mSettings->correctSkewedness()){
lambda = logDerivateOfGluonDensity(t, xpom);
result *= skewednessCorrection(lambda);
}
}
return result;
}
double CrossSection::dsigdt_coherent(double t, double Q2, double W2, GammaPolarization pol) const
{
double val = mTableCollection->get(Q2, W2, t, pol, mean_A); // fm^2
double result = val*val/(16*M_PI); // units now are fm^4
result /= hbarc2; // in fm^2/GeV^2
result *= 1e7; // in nb/GeV^2
double lambda = 0;
if (mSettings->correctForRealAmplitude()){
lambda = logDerivateOfAmplitude(t, Q2, W2, pol);
result *= realAmplitudeCorrection(lambda);
}
if (mSettings->correctSkewedness()){
lambda = logDerivateOfGluonDensity(t, Q2, W2, pol);
result *= skewednessCorrection(lambda);
}
return result;
}
//
// UPC Version
//
double CrossSection::dsigdt_coherent(double t, double xpom) const
{
double val = mTableCollection->get(xpom, t, mean_A); // fm^2
double result = val*val/(16*M_PI); // units now are fm^4
result /= hbarc2; // in fm^2/GeV^2
result *= 1e7; // in nb/GeV^2
double lambda = 0;
if (mSettings->correctForRealAmplitude()){
lambda = logDerivateOfAmplitude(t, xpom);
result *= realAmplitudeCorrection(lambda);
}
if (mSettings->correctSkewedness()){
lambda = logDerivateOfGluonDensity(t, xpom);
result *= skewednessCorrection(lambda);
}
return result;
}
double CrossSection::dsigdtdQ2dW2_total(double t, double Q2, double W2, GammaPolarization pol) const
{
double result = dsigdt_total(t, Q2, W2, pol);
if (mSettings->applyPhotonFlux()) result *= mPhotonFlux(Q2,W2,pol);
return result;
}
//
// UPC version
//
double CrossSection::dsigdtdxp_total(double t, double xpom) const
{
double result = dsigdt_total(t, xpom);
if (mSettings->applyPhotonFlux()) {
result *= UPCPhotonFlux(t, xpom);
}
return result;
}
double CrossSection::dsigdt_incoherent(double t, double Q2, double W2, GammaPolarization pol) const
{
double result = mTableCollection->get(Q2, W2, t, pol, variance_A); // fm^4
result /= (16*M_PI);
result /= hbarc2; // in fm^2/GeV^2
result *= 1e7; // in nb/GeV^2
double lambda = 0;
if (mSettings->correctForRealAmplitude()){
lambda = logDerivateOfAmplitude(t, Q2, W2, pol);
result *= realAmplitudeCorrection(lambda);
}
if (mSettings->correctSkewedness()){
lambda = logDerivateOfGluonDensity(t, Q2, W2, pol);
result *= skewednessCorrection(lambda);
}
return result;
}
//
// UPC Version
//
double CrossSection::dsigdt_incoherent(double t, double xpom) const
{
double result = mTableCollection->get(xpom, t, variance_A); // fm^4
result /= (16*M_PI);
result /= hbarc2; // in fm^2/GeV^2
result *= 1e7; // in nb/GeV^2
double lambda = 0;
if (mSettings->correctForRealAmplitude()){
lambda = logDerivateOfAmplitude(t, xpom);
result *= realAmplitudeCorrection(lambda);
}
if (mSettings->correctSkewedness()){
lambda = logDerivateOfGluonDensity(t, xpom);
result *= skewednessCorrection(lambda);
}
return result;
}
double CrossSection::dsigdtdQ2dW2_coherent(double t, double Q2, double W2, GammaPolarization pol) const
{
double result = dsigdt_coherent(t, Q2, W2, pol);
if (mSettings->applyPhotonFlux()) result *= mPhotonFlux(Q2,W2,pol);
return result;
}
//
// UPC Version
//
double CrossSection::dsigdtdxp_coherent(double t, double xpom) const
{
double result = dsigdt_coherent(t, xpom);
if (mSettings->applyPhotonFlux()) {
result *= UPCPhotonFlux(t, xpom);
}
return result;
}
double CrossSection::UPCPhotonFlux(double t, double xpom) const{
double Egamma = Kinematics::Egamma(xpom, t, mVmMass,
mSettings->hadronBeamEnergy(), mSettings->electronBeamEnergy());
double result = mPhotonFlux(Egamma);
//
// Jacobian = dEgamma/dN, such that dsig/dtdxp = dsigd/tdEgam*dEgam/dxp
//
double h=xpom*1e-3;
double Egamma_p = Kinematics::Egamma(xpom+h, t, mVmMass,
mSettings->hadronBeamEnergy(), mSettings->electronBeamEnergy());
double Egamma_m = Kinematics::Egamma(xpom-h, t, mVmMass,
mSettings->hadronBeamEnergy(), mSettings->electronBeamEnergy());
double jacobian = (Egamma_p-Egamma_m)/(2*h);
result *= fabs(jacobian);
return result;
}
GammaPolarization CrossSection::polarizationOfLastCall() const {return mPolarization;}
DiffractiveMode CrossSection::diffractiveModeOfLastCall() const {return mDiffractiveMode;}
double CrossSection::logDerivateOfAmplitude(double t, double Q2, double W2, GammaPolarization pol) const
{
double lambda = 0;
bool lambdaFromTable = true;
Table *table = 0;
if (!mProtonTableCollection) {
cout << "CrossSection::logDerivateOfAmplitude(): no proton table defined to obtain lambda." << endl;
cout << " Corrections not available. Should be off." << endl;
return 0;
}
//
// If the lambda table is present we use the more accurate and numerically
// stable table value. Otherwise we calculate it from the <A> table(s).
//
// Note (TU): if the lambda value from a table is not valid and not > 0,
// get() returns lambda=0 and table=0. This enforces a renewed
// calculation.
//
lambda = mProtonTableCollection->get(Q2, W2, t, pol, lambda_A, table);
if (!table) { // no lambda value from correct table, use fallback solution
lambdaFromTable = false;
double value = mProtonTableCollection->get(Q2, W2, t, pol, mean_A, table); // use obtained table from here on
/*
if (value <= 0) {
cout << "CrossSection::logDerivateOfAmplitude(): got invalid value from table, value=" << value << '.' << endl;
cout << " t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2) << ", pol="
<< (pol == transverse ? 'T' : 'L') << endl;
return 0;
}
*/
if (!table) {
cout << "CrossSection::logDerivateOfAmplitude(): got invalid pointer to lookup table." << endl;
return 0;
}
//
// Note: the derivate taken from values in the table is a delicate issue.
// The standard interpolation method(s) used in Table::get() are
// at times not accurate and causes ripples on lambda(W2).
// However, interpolation methods or robust fitting is by far too
// expensive in terms of time per point.
//
//
// Calculate the derivative using simple method.
//
double derivative;
double hplus, hminus;
double dW2 = table->binWidthW2();
hplus = hminus = 0.5*dW2; // half bin width is found to be the best choice after some testing
hplus = min(hplus, table->maxW2()-W2);
hminus = min(hminus, W2-table->minW2());
hminus -= numeric_limits<float>::epsilon();
hplus -= numeric_limits<float>::epsilon();
if (hminus < 0) hminus = 0;
if (hplus < 0) hplus = 0;
if (hplus == 0 && hminus == 0) {
cout << "CrossSection::logDerivateOfAmplitude(): Warning, cannot find derivative." << endl;
return 0;
}
double a = table->get(Q2, W2+hplus, t);
double b = table->get(Q2, W2-hminus, t);
derivative = (a-b)/(hplus+hminus);
//
// Finally calculate lambda
//
double jacobian = (W2-protonMass2+Q2)/value;
lambda = jacobian*derivative;
} // end fall back solution
if (mSettings->verboseLevel() > 3) {
cout << "CrossSection::logDerivateOfAmplitude(): ";
if (lambdaFromTable)
cout << "Info, lambda taken from table." << endl;
else
cout << "Info, lambda derived numerically from proton amplitude table" << endl;
}
//
// At a lambda of ~0.6 both corrections have equal value
// of around 2.9. This will yield excessive large (unphysical)
// corrections. Large values are typically caused by fluctuations
// and glitches in the tables and should be rare.
// In all cases where something goes wrong we set lambda to
// its most probably value (0.2)
//
const double lambdaInCaseOfError = 0.2;
double maxLambda = mSettings->maxLambdaUsedInCorrections();
if (fabs(lambda) > maxLambda) {
if (mSettings->verboseLevel() > 2) {
cout << "CrossSection::logDerivateOfAmplitude(): ";
cout << "Warning, lambda is excessively large (" << lambda << ") at " ;
cout << "Q2=" << Q2 << ", W2=" << W2 << ", t=" << t << endl;
cout << "Set to " << (lambda > 0 ? maxLambda : -maxLambda) << "." << endl;
}
lambda = lambda > 0 ? maxLambda : -maxLambda;
}
/*
if (lambda < 0) {
if (mSettings->verboseLevel() > 2) {
cout << "CrossSection::logDerivateOfAmplitude(): ";
cout << "Warning, lambda is negative (" << lambda << "). " ;
cout << "Set to " << (-lambda) << "." << endl;
}
lambda = -lambda;
}
*/
if (std::isinf(lambda)) {
cout << "CrossSection::logDerivateOfAmplitude(): error, lambda = inf for pol=" << (pol == transverse ? 'T' : 'L') << endl;
cout << " t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2) << endl;
lambda = lambdaInCaseOfError;
}
if (std::isnan(lambda)) {
cout << "CrossSection::logDerivateOfAmplitude(): error, lambda is NaN for pol=" << (pol == transverse ? 'T' : 'L') << endl;
cout << " t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2) << endl;
lambda = lambdaInCaseOfError;
}
return lambda;
}
double CrossSection::logDerivateOfGluonDensity(double t, double Q2, double W2, GammaPolarization pol) const
{
double lambda = 0;
bool lambdaFromTable = true;
Table *table = 0;
if (!mProtonTableCollection) {
cout << "CrossSection::logDerivateOfGluonDensity(): no proton table defined to obtain lambda." << endl;
cout << " Corrections not available. Should be off." << endl;
return 0;
}
//
// If the lambda table is present we use the more accurate and numerically
// stable table value. Otherwise we calculate it from the <A> table(s).
//
lambda = mProtonTableCollection->get(Q2, W2, t, pol, lambda_skew, table);
if (!table) {
//
// no lambda value from correct table, use fallback solution
// lambda_skew ~ lambda_real
//
lambda = logDerivateOfAmplitude(t, Q2, W2, pol) ; // This is an approximation in this case
} // end fall back solution
if (mSettings->verboseLevel() > 3) {
cout << "CrossSection::logDerivateOfGluonDensity(): ";
if (lambdaFromTable)
cout << "Info, lambda taken from table." << endl;
else
cout << "Info, lambda taken from logDerivateOfAmplitude as an approximation" << endl;
}
//
// At a lambda of ~0.6 both corrections have equal value
// of around 2.9. This will yield excessive large (unphysical)
// corrections. Large values are typically caused by fluctuations
// and glitches in the tables and should be rare.
// In all cases where something goes wrong we set lambda to
// its most probably value (0.2)
//
const double lambdaInCaseOfError = 0.2;
double maxLambda = mSettings->maxLambdaUsedInCorrections();
if (fabs(lambda) > maxLambda) {
if (mSettings->verboseLevel() > 2) {
cout << "CrossSection::logDerivateOfGluonDensity(): ";
cout << "Warning, lambda is excessively large (" << lambda << ") at " ;
cout << "Q2=" << Q2 << ", W2=" << W2 << ", t=" << t << endl;
cout << "Set to " << (lambda > 0 ? maxLambda : -maxLambda) << "." << endl;
}
lambda = lambda > 0 ? maxLambda : -maxLambda;
}
if (std::isinf(lambda)) {
cout << "CrossSection::logDerivateOfGluonDensity(): error, lambda = inf for pol=" << (pol == transverse ? 'T' : 'L') << endl;
cout << " t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2) << endl;
lambda = lambdaInCaseOfError;
}
if (std::isnan(lambda)) {
cout << "CrossSection::logDerivateOfGluonDensity(): error, lambda is NaN for pol=" << (pol == transverse ? 'T' : 'L') << endl;
cout << " t=" << t << ", Q2=" << Q2 << ", W=" << sqrt(W2) << endl;
lambda = lambdaInCaseOfError;
}
return lambda;
}
//
// UPC only version
//
double CrossSection::logDerivateOfAmplitude(double t, double xpom) const
{
double lambda = 0;
bool lambdaFromTable = true;
Table *table = 0;
if (!mProtonTableCollection) {
cout << "CrossSection::logDerivateOfAmplitude(): no proton table defined to obtain lambda." << endl;
cout << " Corrections not available. Should be off." << endl;
return 0;
}
//
// Usual numeric issues at boundaries.
// Subtracting an eps in log(xpom) does the trick.
//
if (xpom > mProtonTableCollection->maxX()) {
- xpom = exp(log(xpom)-numeric_limits<float>::epsilon());
+ xpom = exp(log(xpom)-numeric_limits<float>::epsilon());
}
if (xpom < mProtonTableCollection->minX()) {
xpom = exp(log(xpom)+numeric_limits<float>::epsilon());
}
//
// If the lambda table is present we use the more accurate and numerically
// stable table value. Otherwise we calculate it from the <A> table(s).
//
lambda = mProtonTableCollection->get(xpom, t, lambda_A, table);
if (!table) { // no lambda value from correct table, use fallback solution
lambdaFromTable = false;
- (void) mProtonTableCollection->get(xpom, t, mean_A, table); // use obtained table from here on
+ (void) mProtonTableCollection->get(xpom, t, mean_A, table); // use obtained table from here on
if (!table) {
cout << "CrossSection::logDerivateOfAmplitude(): got invalid pointer to lookup table." << endl;
return 0;
}
//
// Here's the tricky part (see comments in non-UPC version above)
//
double theLogxpom = log(xpom);
double dlogxpom = table->binWidthX(); // assuming table is in log x ???????? FIX later
double maxLogxpom = log(table->maxX());
double derivative;
double hplus, hminus;
hplus = hminus = 0.5*dlogxpom;
hplus = min(hplus, fabs(maxLogxpom-theLogxpom));
hminus = min(hminus, fabs(theLogxpom-log(table->minX())));
hminus -= numeric_limits<float>::epsilon();
hplus -= numeric_limits<float>::epsilon();
if (hminus < 0) hminus = 0;
if (hplus < 0) hplus = 0;
if (hplus == 0 && hminus == 0) {
cout << "CrossSection::logDerivateOfAmplitude(): Warning, cannot find derivative." << endl;
return 0;
}
- double a = table->get(theLogxpom+hplus, t);
- double b = table->get(theLogxpom-hminus, t);
- derivative = (a-b)/(hplus+hminus);
-
+ double a = table->get(exp(theLogxpom+hplus), t);
+ double b = table->get(exp(theLogxpom-hminus), t);
+ derivative = log(a/b)*xpom/(hplus+hminus);
+
//
// Finally calculate lambda
// Directly dlog(A)/-dlog(xpom) here.
- // #TT Are we missing the 1/A part in this derivative??
+ //
lambda = -derivative;
}
if (mSettings->verboseLevel() > 3) {
cout << "CrossSection::logDerivateOfAmplitude(): ";
if (lambdaFromTable)
cout << "Info, lambda taken from table." << endl;
else
cout << "Info, lambda derived numerically from proton amplitude table" << endl;
+ cout << " t=" << t << ", xpom=" << xpom << endl;
}
//
// At a lambda of ~0.6 both corrections have equal value
// of around 2.9. This will yield excessive large (unphysical)
// corrections. Large values are typically caused by fluctuations
// and glitches in the tables and should be rare.
// In all cases where something goes wrong we set lambda to
// its most probably value (0.2)
//
const double lambdaInCaseOfError = 0.2;
double maxLambda = mSettings->maxLambdaUsedInCorrections();
if (fabs(lambda) > maxLambda) {
if (mSettings->verboseLevel() > 2) {
cout << "CrossSection::logDerivateOfAmplitude(): ";
cout << "Warning, lambda is excessively large (" << lambda << ") at " ;
cout << "xpom=" << xpom << ", t=" << t << endl;
cout << "Set to " << (lambda > 0 ? maxLambda : -maxLambda) << "." << endl;
}
lambda = lambda > 0 ? maxLambda : -maxLambda;
}
if (std::isinf(lambda)) {
cout << "CrossSection::logDerivateOfAmplitude(): error, lambda = infinity for" << endl;
cout << " t=" << t << ", xpom=" << xpom << endl;
lambda = lambdaInCaseOfError;
}
if (std::isnan(lambda)) {
cout << "CrossSection::logDerivateOfAmplitude(): error, lambda is NaN for" << endl;
cout << " t=" << t << ", xpom=" << xpom << endl;
lambda = lambdaInCaseOfError;
}
return lambda;
}
//
// UPC version
//
double CrossSection::logDerivateOfGluonDensity(double t, double xpom) const
{
double lambda = 0;
bool lambdaFromTable = true;
Table *table = 0;
if (!mProtonTableCollection) {
cout << "CrossSection::logDerivateOfGluonDensity(): no proton table defined to obtain lambda." << endl;
cout << " Corrections not available. Should be off." << endl;
return 0;
}
//
// Usual numeric issues at boundaries.
// Subtracting an eps in log(xpom) does the trick.
//
if (xpom > mProtonTableCollection->maxX()) {
xpom = exp(log(xpom)-numeric_limits<float>::epsilon());
}
if (xpom < mProtonTableCollection->minX()) {
xpom = exp(log(xpom)+numeric_limits<float>::epsilon());
}
//
// If the lambda table is present we use the more accurate and numerically
// stable table value. Otherwise we calculate it from the <A> table(s).
//
lambda = mProtonTableCollection->get(xpom, t, lambda_skew, table);
if (!table) {
//
// no lambda value from correct table, use fallback solution
// lambda_skew ~ lambda_real
//
lambdaFromTable = false;
lambda = logDerivateOfAmplitude(t, xpom);
}
if (mSettings->verboseLevel() > 3) {
cout << "CrossSection::logDerivateOfGluonDensity(): ";
if (lambdaFromTable)
cout << "Info, lambda taken from table." << endl;
else
cout << "Info, lambda taken from logDerivateOfAmplitude as approximation." << endl;
}
//
// At a lambda of ~0.6 both corrections have equal value
// of around 2.9. This will yield excessive large (unphysical)
// corrections. Large values are typically caused by fluctuations
// and glitches in the tables and should be rare.
// In all cases where something goes wrong we set lambda to
// its most probably value (0.2)
//
const double lambdaInCaseOfError = 0.2;
double maxLambda = mSettings->maxLambdaUsedInCorrections();
if (fabs(lambda) > maxLambda) {
if (mSettings->verboseLevel() > 2) {
cout << "CrossSection::logDerivateOfGluonDensity(): ";
cout << "Warning, lambda is excessively large (" << lambda << ") at " ;
cout << "xpom=" << xpom << ", t=" << t << endl;
cout << "Set to " << (lambda > 0 ? maxLambda : -maxLambda) << "." << endl;
}
lambda = lambda > 0 ? maxLambda : -maxLambda;
}
if (std::isinf(lambda)) {
cout << "CrossSection::logDerivateOfGluonDensity(): error, lambda = infinity for" << endl;
cout << " t=" << t << ", xpom=" << xpom << endl;
lambda = lambdaInCaseOfError;
}
if (std::isnan(lambda)) {
cout << "CrossSection::logDerivateOfGluonDensity(): error, lambda is NaN for" << endl;
cout << " t=" << t << ", xpom=" << xpom << endl;
lambda = lambdaInCaseOfError;
}
return lambda;
}
double CrossSection::realAmplitudeCorrection(double lambda) const
{
//
// Correction factor for real amplitude contribution
//
double beta = tan(lambda*M_PI/2.);
double correction = 1 + beta*beta;
return correction;
}
double CrossSection::realAmplitudeCorrection(double t, double Q2, double W2, GammaPolarization pol) const
{
double lambda = logDerivateOfAmplitude(t, Q2, W2, pol);
double correction = realAmplitudeCorrection(lambda);
// correction *= exp(-10*Kinematics::x(Q2, W2)); // damped
return correction;
}
//
// UPC version
//
double CrossSection::realAmplitudeCorrection(double t, double xpom) const
{
double lambda = logDerivateOfAmplitude(t, xpom);
double correction = realAmplitudeCorrection(lambda);
return correction;
}
double CrossSection::skewednessCorrection(double lambda) const
{
//
// Skewedness correction
//
double R = pow(2.,2*lambda+3)*TMath::Gamma(lambda+2.5)/(sqrt(M_PI)*TMath::Gamma(lambda+4));
double correction = R*R;
return correction;
}
double CrossSection::skewednessCorrection(double t, double Q2, double W2, GammaPolarization pol) const
{
double lambda = logDerivateOfAmplitude(t, Q2, W2, pol);
double correction = skewednessCorrection(lambda);
// correction *= exp(-10*Kinematics::x(Q2, W2)); // damped
return correction;
}
//
// UPC version
//
double CrossSection::skewednessCorrection(double t, double xpom) const
{
double lambda = logDerivateOfAmplitude(t, xpom);
double correction = skewednessCorrection(lambda);
return correction;
}

File Metadata

Mime Type
text/x-diff
Expires
Sat, May 3, 5:59 AM (1 d, 3 h)
Storage Engine
blob
Storage Format
Raw Data
Storage Handle
4982894
Default Alt Text
(33 KB)

Event Timeline