Re: [ROOT] Minuit questions (seg violation error + covariance)

From: Georges Lobo (lobo@irmm.jrc.be)
Date: Wed May 29 2002 - 15:35:35 MEST


Thanks for the answer but this does not solve the segmentation violation
error I have.
I have this message even when I don't try to call mnemat.
Thanks

----- Original Message -----
From: "Rene Brun" <brun@pcbrun.cern.ch>
To: "lobo" <lobo@irmm.jrc.be>
Cc: <roottalk@pcroot.cern.ch>
Sent: Tuesday, May 28, 2002 9:57 PM
Subject: Re: [ROOT] Minuit questions (seg violation error + covariance)


> Hi George,
>
> Replace the line:
> gMinuit->mnemat(matrix,3);
> by
> gMinuit->mnemat(&matrix[0][0],3);
>
> Rene Brun
>
> On Mon, 27 May 2002, lobo wrote:
>
> > Hi rooters,
> > I have a problem since some days concerning Minuit in root.
> > I have written a small testfile to explain my problems.
> > // testfit.C
> > void testfit() {
> >
> > Double_t x1[]  = {1000,16100,83000,223426,425938};
> > Double_t y1[]  =
{0.0194932,0.00299029,0.00168717,0.00167152,0.00134782};
> > Double_t ex1[] = {40,500,1000,1574.5,3062.5};
> > Double_t ey1[] =
> > {0.00232989,0.000269625,0.000142085,0.000110943,7.1839e-05};
> > TGraphErrors *fluxbkg = new TGraphErrors(5,x1,y1,ex1,ey1);
> > fluxbkg->Print();
> > TF1 *fitFcn = new TF1("fitFcn",Background,0,450000,3);
> > fitFcn->SetParameters(0.,1.,-0.5.);
> > fitFcn->Draw();
> > fluxbkg->Fit("fitFcn");
> > double matrix[3][3];
> > gMinuit->mnemat(matrix,3);
> > }
> >
> > Double_t Background(Double_t *x, Double_t *par) {
> > Double_t fitval = TMath::Power(x[0],par[2]);
> >   return(par[0]+par[1]*fitval);
> >
> >
> > }
> >
> > When I load this file (.L testfit.C) and I execute it I got a
segmentation
> > violation error after the fit.
> >
> > [root][1] .L testfit.C
> > [root][2] testfit();
> > x[0]=1000, y[0]=0.0194932, ex[0]=40, ey[0]=0.00232989
> > x[1]=16100, y[1]=0.00299029, ex[1]=500, ey[1]=0.000269625
> > x[2]=83000, y[2]=0.00168717, ex[2]=1000, ey[2]=0.000142085
> > x[3]=223426, y[3]=0.00167152, ex[3]=1574.5, ey[3]=0.000110943
> > x[4]=425938, y[4]=0.00134782, ex[4]=3062.5, ey[4]=7.1839e-05
> > <TCanvas::MakeDefCanvas>: created default TCanvas with name c1
> >  **********
> >  **    1 **SET ERR           1
> >  **********
> >  **********
> >  **    5 **MIGRAD        5000   5.539e-12
> >  **********
> >  MACHINE ACCURACY LIMITS FURTHER IMPROVEMENT.
> >  MIGRAD MINIMIZATION HAS CONVERGED.
> >  FCN=3.59539 FROM MIGRAD    STATUS=CONVERGED     258 CALLS         259
> > TOTAL
> >                      EDM=6.36245e-15    STRATEGY= 1  ERROR MATRIX
> > UNCERTAINTY
> > 1.6 per cent
> >   EXT PARAMETER                                   STEP         FIRST
> >   NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
> >    1  p0           1.30041e-03   8.98106e-05  -0.00000e+00   1.21824e-03
> >    2  p1           6.12043e+00   3.81528e+00   0.00000e+00  -1.66377e-07
> >    3  p2          -8.43509e-01   7.80422e-02  -0.00000e+00  -6.64307e-06
> >  FCN=3.59539 FROM MIGRAD    STATUS=CONVERGED     258 CALLS         259
> > TOTAL
> >                      EDM=6.36245e-15    STRATEGY= 1  ERROR MATRIX
> > UNCERTAINTY
> > 1.6 per cent
> >   EXT PARAMETER                                   STEP         FIRST
> >   NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
> >    1  p0           1.30041e-03   8.98106e-05  -0.00000e+00   1.21824e-03
> >    2  p1           6.12043e+00   3.81528e+00   0.00000e+00  -1.66377e-07
> >    3  p2          -8.43509e-01   7.80422e-02  -0.00000e+00  -6.64307e-06
> >
> >  *** Break *** segmentation violation
> > Root > Function testfit() busy flag cleared
> >
> > The starnge thing is if I type all the instruction one by one in the
root
> > session I got no error, except it I don't type the line fitFcn->Draw();
> > before.
> >
> > example if I type
> > ...
> > [root] TF1 *fitFcn = new TF1("fitFcn",Background,0,450000,3);
> > [root] fitFcn->SetParameters(0.,1.,-0.5.);
> > [root] fitFcn->Draw();
> > [root] fluxbkg->Fit("fitFcn");
> > ....
> > I have no no segvi error.
> > But if I type
> >
> > [root] TF1 *fitFcn = new TF1("fitFcn",Background,0,450000,3);
> > [root] fitFcn->SetParameters(0.,1.,-0.5.);
> > [root] fluxbkg->Fit("fitFcn");
> > I got a
> > *** Break *** segmentation violation
> >
> > The other thing is about the mnemat call. When thing are working well I
> > got an error when retrieveing the matrix.
> > [root] double matrix[3][3];
> > [root] gMinuit->mnemat(matrix,3);
> >
> > Error: Can't call TMinuit::mnemat(matrix,3) in current scope
> > FILE:/tmp/fileOCWcoy_cint LINE:1
> > Possible candidates are...
> > filename       line:size busy function type and name  (in TMinuit)
> > (compiled)        0:0    0 public: virtual void mnemat(Double_t*
> > emat,Int_t ndim);
> > filename       line:size busy function type and name  (in TNamed)
> > filename       line:size busy function type and name  (in TObject)
> > *** Interpreter error recovered ***
> >
> > Could please someone tell me what is wrong ? I surely made a mistake but
I
> > can't figure out which.
> > Thanks in advance.
> >
> >
> >
>
>



This archive was generated by hypermail 2b29 : Sat Jan 04 2003 - 23:50:54 MET