Fitting histogram errors

ROOT
#1

Dear rooters,

I am facing a strange issue when I try to
fit some histogram with a custom made function. Globally, the fit provides good results, very consistent with what I expect and with other experimental data. But I notice an intermittent error dealing with Fatal in : storage area overwritten. Here is the output of the error I can see :

FCN=90.3587 FROM HESSE STATUS=NOT POSDEF 31 CALLS 207 TOTAL
EDM=9.80156e-14 STRATEGY= 1 ERR MATRIX NOT POS-DEF
EXT PARAMETER APPROXIMATE STEP FIRST
NO. NAME VALUE ERROR SIZE DERIVATIVE
1 p0 -1.56127e-05 2.42892e-08 3.72236e-12 -3.20897e+01
2 p1 5.59907e-03 3.48026e-06 1.33492e-09 -2.97977e-01
3 p2 -7.14542e-01 3.96131e-04 1.70360e-07 -2.62362e-03
4 p3 3.66635e+01 4.14804e-02 8.74126e-06 -1.90331e-05
5 p4 -5.47678e+02 2.38374e+00 3.63694e-04 1.62159e-08
E_1: 55 backgrd_E_1: 96.0044 Max_E_1: 134.97 max_peak_E_1: 38.9656
E_2: 60.9091 backgrd_E_2: 84.8945 Max_E_2: 88.3056 max_peak_E_2: 3.41114
FCN=2.47081 FROM MIGRAD STATUS=CONVERGED 875 CALLS 876 TOTAL
EDM=4.34333e-08 STRATEGY= 1 ERROR MATRIX UNCERTAINTY 2.2 per cent
EXT PARAMETER STEP FIRST
NO. NAME VALUE ERROR SIZE DERIVATIVE
1 p0 -8.12563e-06 1.98654e-06 2.85970e-09 3.99707e+03
2 p1 3.00569e-03 6.54098e-04 -8.83948e-07 3.42187e+01
3 p2 -3.93076e-01 7.74112e-02 9.52132e-05 2.32789e-01
4 p3 2.01342e+01 3.84243e+00 -4.12519e-03 2.41401e-04
5 p4 -2.63851e+02 6.55240e+01 5.88068e-02 -3.93476e-05
6 p5 4.84356e+01 7.28583e+00 -8.67274e-03 -6.86276e-05
7 p6 5.43893e+01 4.74220e-01 3.59787e-04 -6.56050e-06
8 p7 3.04316e+00 5.89669e-01 1.49985e-03 -8.35512e-04
9 p8 1.84284e+01 6.60498e+00 -1.68915e-02 -4.31161e-05
10 p9 3.61887e+00 1.12930e+00 8.38908e-04 -1.37695e-04
Fatal in : storage area overwritten
aborting
#0 0x00007fe46cfce07a in __GI___waitpid (pid=6166, stat_loc=stat_loc
entry=0x7ffe735defc0, options=options
entry=0) at …/sysdeps/unix/sysv/linux/waitpid.c:29
#1 0x00007fe46cf46fbb in do_system (line=) at …/sysdeps/posix/system.c:148
#2 0x00007fe46ecdb9dd in TUnixSystem::StackTrace() () from /home/linux/Bin/root/root-6.10.02/lib/libCore.so
#3 0x00007fe46ec169ba in DefaultErrorHandler(int, bool, char const*, char const*) () from /home/linux/Bin/root/root-6.10.02/lib/libCore.so
#4 0x00007fe46ec16352 in ErrorHandler () from /home/linux/Bin/root/root-6.10.02/lib/libCore.so
#5 0x00007fe46ec1684f in Fatal(char const*, char const*, …) () from /home/linux/Bin/root/root-6.10.02/lib/libCore.so
#6 0x00007fe46dd8c180 in operator delete(void*) () from /home/linux/Bin/root/root-6.10.02/lib/libNew.so
#7 0x00007fe46ba74dcf in ROOT::Fit::FitResult::~FitResult() () from /home/linux/Bin/root/root-6.10.02/lib/libMathCore.so
#8 0x00007fe46e5daa6b in std::_Sp_counted_ptr<TFitResult*, (__gnu_cxx::_Lock_policy)2>::_M_dispose() () from /home/linux/Bin/root/root-6.10.02/lib/libHist.so
#9 0x00007fe46ba8ea4a in ROOT::Fit::Fitter::~Fitter() () from /home/linux/Bin/root/root-6.10.02/lib/libMathCore.so
#10 0x00007fe46e5d42a2 in std::_Sp_counted_ptr<ROOT::Fit::Fitter*, (__gnu_cxx::_Lock_policy)2>::_M_dispose() () from /home/linux/Bin/root/root-6.10.02/lib/libHist.so
#11 0x00007fe46e67573a in TBackCompFitter::~TBackCompFitter() () from /home/linux/Bin/root/root-6.10.02/lib/libHist.so
#12 0x00007fe46e675789 in TBackCompFitter::~TBackCompFitter() () from /home/linux/Bin/root/root-6.10.02/lib/libHist.so
#13 0x00007fe46e5df7b7 in TFitResultPtr HFit::Fit(TH1*, TF1*, Foption_t&, ROOT::Math::MinimizerOptions const&, char const*, ROOT::Fit::DataRange&) () from /home/linux/Bin/root/root-6.10.02/lib/libHist.so
#14 0x00007fe46e5d3b3a in ROOT::Fit::FitObject(TH1*, TF1*, Foption_t&, ROOT::Math::MinimizerOptions const&, char const*, ROOT::Fit::DataRange&) () from /home/linux/Bin/root/root-6.10.02/lib/libHist.so
#15 0x00007fe46e6c8613 in TH1::Fit(TF1*, char const*, char const*, double, double) () from /home/linux/Bin/root/root-6.10.02/lib/libHist.so
#16 0x00007fe46e6c7889 in TH1::Fit(char const*, char const*, char const*, double, double) () from /home/linux/Bin/root/root-6.10.02/lib/libHist.so
#17 0x0000000000404b91 in main ()

Does anyone can provide some hint about this ?

Thanks for your support,

best regards,

Loic

#2

This indicates a double delete and/or read of un-initialized memory.

To investigate further, we would need to see the failing code.

Even better you can run valgrind (for example valgrind --suppressions=$ROOTSYS/etc/valgrind-root.supp root.exe -b -l -q myMacro.C which might tell us exactly what the problem is.

Cheers,
Philippe.

#3

Hello Philippe,

I ran valgrind yesterday. Log file is attached to this post.

Here is an extract of my code:

Double_t BackGrd(Double_t *x, Double_t *par) {

return par[0]*x[0]*x[0]*x[0]*x[0]+par[1]*x[0]*x[0]*x[0]+par[2]*x[0]*x[0]+par[3]*x[0]+par[4];

}

Double_t fit_peak_1(Double_t *x, Double_t *par) {

Double_t arg=x[0]-par[1];
Double_t sig=par[2];
return par[0] * TMath::Exp(-0.5*arg*arg/sig/sig);

}

Double_t fit_peak_2(Double_t *x, Double_t *par) {

Double_t arg=x[0]-67.0/1.1;
Double_t sig=par[1];
return par[0] * TMath::Exp(-0.5*arg*arg/sig/sig);

}

Double_t fit_fnc(Double_t *x, Double_t *par){

return BackGrd(x,par)+fit_peak_1(x,&par[5])+fit_peak_2(x,&par[8]);

}

Nrj_1D[300]->Fit(“BACKGRD”,“0”,“L”,30,120);

  Double_t par_0=BACKGRD->GetParameter(0);
  Double_t par_1=BACKGRD->GetParameter(1);
  Double_t par_2=BACKGRD->GetParameter(2);
  Double_t par_3=BACKGRD->GetParameter(3);
  Double_t par_4=BACKGRD->GetParameter(4);
  
  
  Double_t E_1=Nrj_1D[300]->GetMaximumBin();;
  Double_t backgrd_E_1=par_0*E_1*E_1*E_1*E_1+par_1*E_1*E_1*E_1+par_2*E_1*E_1+par_3*E_1+par_4;
  Double_t Max_E_1=Nrj_1D[300]->GetMaximum();
  Double_t max_peak_E_1=Max_E_1-backgrd_E_1;
  
  Double_t E_2=67/1.1;
  Double_t backgrd_E_2=par_0*E_2*E_2*E_2*E_2+par_1*E_2*E_2*E_2+par_2*E_2*E_2+par_3*E_2+par_4;
  
  Double_t Max_E_2=Nrj_1D[300]->GetBinContent((Int_t)E_2+1);
  Double_t max_peak_E_2=Max_E_2-backgrd_E_2;
  
  cout<<"E_1: "<<E_1<<" backgrd_E_1: "<<backgrd_E_1<<" Max_E_1: "<<Max_E_1<<" max_peak_E_1: "<<max_peak_E_1<<endl;
  cout<<"E_2: "<<E_2<<" backgrd_E_2: "<<backgrd_E_2<<" Max_E_2: "<<Max_E_2<<" max_peak_E_2: "<<max_peak_E_2<<endl;
  
  TF1* Fit_FNC=new TF1("Fit_FNC", fit_fnc, 30.,120.,10);
  Fit_FNC->SetParameters(par_0,par_1,par_2,par_3,par_4,max_peak_E_1,E_1,3.,max_peak_E_2,3.);      
  Fit_FNC->SetNpx(1001);
  
   
  
  Nrj_1D[300]->Fit(Fit_FNC,"0","L",30,120);
  Fit_FNC->SetLineColor(1);
  Fit_FNC->Draw("same");		
	
TF1* Photo_Peak=new TF1("Photo_Peak",fit_peak_1,0,128,3);
Photo_Peak->SetParameter(0,Fit_FNC->GetParameter(5));
Photo_Peak->SetParameter(1,Fit_FNC->GetParameter(6));
Photo_Peak->SetParameter(2,Fit_FNC->GetParameter(7));
Photo_Peak->SetLineColor(2);
Photo_Peak->Draw("same");
Photo_Peak->SetNpx(1001);

Whatever the type of fitting function I select, incluted predefined functions, I get this intermittent failure.

Apparently I miss something, probably too obvious, but I do not get its nature yet.

Thanks for your support,

Best regards,

Loic
output.tar.gz (1.1 MB)

#4

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.