Roofit differences in LCG_105 and LCG_107

Dear ROOT experts,

I am trying to fit a distribution using a gaussian to describe the J/Psi resonance and an error function to describe the background.

I noticed that if I use LCG_105, I get a sensible result, while using LCG_107 the fit does not converge. Are there significant differences between the roofit version in the two LCG stack releases that can explain the two behaviors?

I send in attachment the plots showing the results, and the links to the dumps of the LCG_105 and LCG_107 fits. Sorry for copy-pasting here the first lines of the logs, but apparently I cannot put links in the post.

Best,
Nicolò

LCG_105:

Setting up environment
Setup complete
Will run:
run_single_fit ./example/fits_data/muon/generalTracks/JPsi/Run2025/Nominal/NUM_SoftID_DEN_TrackerMuons/NUM_SoftID_DEN_TrackerMuons_PV_1.root ./example/flat/muon/generalTracks/JPsi/Run2025/Run2025/Nominal/NUM_SoftID_DEN_TrackerMuons_PV.root NUM_SoftID_DEN_TrackerMuons_PV_1 ./example/flat/muon/generalTracks/JPsi/Run2025/JPsi_pythia8/Nominal/NUM_SoftID_DEN_TrackerMuons_PV.root Nominal data Nominal JPsi True
[#1] INFO:ObjectHandling -- RooWorkspace::import(wmain) importing dataset hGenPass
[#1] INFO:ObjectHandling -- RooWorkspace::import(wmain) importing dataset hGenFail
[#1] INFO:ObjectHandling -- RooWorkspace::import(wmain) importing dataset hPass
[#1] INFO:ObjectHandling -- RooWorkspace::import(wmain) importing dataset hFail
[#0] ERROR:Caching -- The internal binning of variable x is not uniform. The numerical FFT will likely yield wrong results.
[#1] INFO:Minimization -- p.d.f. provides expected number of events, including extended term in likelihood.
[#1] INFO:Eval -- RooRealVar::setRange(x) new range named 'refrange_fft_sigPass' created with bounds [2.7,3.5]
[#0] WARNING:Eval -- The FFT convolution 'sigPass' will run with 20 bins. A decent accuracy for difficult convolutions is typically only reached with n >= 1000. Suggest to increase the number of bins of the observable 'x'.
[#1] INFO:Caching -- RooAbsCachedPdf::getCache(sigPass) creating new cache 0xecd7000 with pdf sigPhysPass_CONV_sigResPass_CACHE_Obs[x]_NORM_x for nset (x) with code 0
[#1] INFO:NumericIntegration -- RooRealIntegral::init(bkgPass_Int[x]) using numeric integrator RooRombergIntegrator to calculate Int(x)
[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: activating const optimization
[#1] INFO:Minimization --  The following expressions have been identified as constant and will be precalculated and cached: (sigPhysPass)
 **********
 **    1 **SET PRINT           1
 **********
 **********
 **    2 **SET NOGRAD
 **********
 PARAMETER DEFINITIONS:
    NO.   NAME         VALUE      STEP SIZE      LIMITS
     1 acmsP        2.90000e+00  2.00000e-01    2.00000e+00  4.00000e+00
     2 betaP        5.00000e-02  9.90000e-03    1.00000e-03  1.00000e-01
     3 gammaP       1.00000e-01  4.00000e+00   -2.00000e+01  2.00000e+01
     4 meanP       -0.00000e+00  1.00000e+00   -5.00000e+00  5.00000e+00
     5 nBkgP        1.98600e+03  9.92750e+02    5.00000e-01  2.18460e+04
     6 nSigP        1.93803e+04  1.23285e+03    5.00000e-01  2.18460e+04
     7 sigmaP       9.00000e-01  4.47500e-01    5.00000e-03  5.00000e+00
 **********
 **    3 **SET ERR         0.5
 **********
 **********
 **    4 **SET PRINT           1
 **********
 **********
 **    5 **SET STR           1
 **********
 NOW USING STRATEGY  1: TRY TO BALANCE SPEED AGAINST RELIABILITY
 **********
 **    6 **MINIMIZE        3500           1
 **********
 FIRST CALL TO USER FUNCTION AT NEW START POINT, WITH IFLAG=4.
 START MIGRAD MINIMIZATION.  STRATEGY  1.  CONVERGENCE WHEN EDM .LT. 1.00e-03
 FCN=-170252 FROM MIGRAD    STATUS=INITIATE      152 CALLS         153 TOTAL
                     EDM= unknown      STRATEGY= 1      NO ERROR MATRIX       
  EXT PARAMETER               CURRENT GUESS       STEP         FIRST   
  NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE 
   1  acmsP        3.99906e+00   2.00000e-01   0.00000e+00   4.29184e-02
   2  betaP        9.98574e-02   9.90000e-03   0.00000e+00  -3.80316e-01
   3  gammaP       1.00000e-01   4.00000e+00   0.00000e+00   1.52361e+03
   4  meanP        0.00000e+00   1.00000e+00   0.00000e+00  -1.08044e+07
   5  nBkgP        1.98600e+03   9.92750e+02   0.00000e+00  -4.31787e+03
   6  nSigP        1.38921e+04   1.23285e+03  -6.10284e-01  -1.73572e+02
   7  sigmaP       5.02489e-03   4.47500e-01   0.00000e+00  -5.69635e+02
                               ERR DEF= 0.5
 MIGRAD FAILS TO FIND IMPROVEMENT
 MIGRAD MINIMIZATION HAS CONVERGED.
 FCN=-172054 FROM MIGRAD    STATUS=CONVERGED     323 CALLS         324 TOTAL
                     EDM=0.000373228    STRATEGY= 1  ERROR MATRIX UNCERTAINTY   0.2 per cent
  EXT PARAMETER                                   STEP         FIRST   
  NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE 
   1  acmsP        4.00000e+00   1.23001e+00  -0.00000e+00   1.25207e-05
   2  betaP        9.99997e-02   4.86760e-03   0.00000e+00  -6.05775e-04
   3  gammaP       1.59499e-01   6.98619e-02   0.00000e+00   3.14645e-01
   4  meanP        1.84845e-03   1.53809e-04   0.00000e+00  -8.99264e+01
   5  nBkgP        4.29212e+03   9.00131e+01   0.00000e+00   1.03257e-01
   6  nSigP        1.31814e+04   2.06597e+02  -0.00000e+00   3.37733e-01
   7  sigmaP       5.00437e-03   3.97835e-05   0.00000e+00  -8.81927e+00
                               ERR DEF= 0.5
 EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  7    ERR DEF=0.5
  2.600e-06 -5.230e-12 -5.129e-08 -2.477e-10 -1.112e-04  4.373e-04 -5.320e-11 
 -5.230e-12  5.376e-09  4.550e-08  2.966e-10  1.204e-04 -4.455e-04  5.831e-11 
 -5.129e-08  4.550e-08  4.881e-03  2.574e-06  1.154e+00 -3.283e+00  5.210e-07 
 -2.477e-10  2.966e-10  2.574e-06  2.366e-08  6.467e-03 -1.992e-02  3.468e-09 
 -1.112e-04  1.204e-04  1.154e+00  6.467e-03  8.103e+03 -9.677e+03  1.394e-03 
  4.373e-04 -4.455e-04 -3.283e+00 -1.992e-02 -9.677e+03  4.269e+04 -5.042e-03 
 -5.320e-11  5.831e-11  5.210e-07  3.468e-09  1.394e-03 -5.042e-03  6.952e-10 
 PARAMETER  CORRELATION COEFFICIENTS  
       NO.  GLOBAL      1      2      3      4      5      6      7
        1  0.00140   1.000 -0.000 -0.000 -0.001 -0.001  0.001 -0.001
        2  0.03165  -0.000  1.000  0.009  0.026  0.018 -0.029  0.030
        3  0.33338  -0.000  0.009  1.000  0.240  0.184 -0.227  0.283
        4  0.96362  -0.001  0.026  0.240  1.000  0.467 -0.627  0.855
        5  0.63121  -0.001  0.018  0.184  0.467  1.000 -0.520  0.587
        6  0.98091   0.001 -0.029 -0.227 -0.627 -0.520  1.000 -0.926
        7  0.99225  -0.001  0.030  0.283  0.855  0.587 -0.926  1.000

LCG_107:

Setting up environment
Setup complete
Will run:
run_single_fit ./example/fits_data/muon/generalTracks/JPsi/Run2025/Nominal/NUM_SoftID_DEN_TrackerMuons/NUM_SoftID_DEN_TrackerMuons_PV_1.root ./example/flat/muon/generalTracks/JPsi/Run2025/Run2025/Nominal/NUM_SoftID_DEN_TrackerMuons_PV.root NUM_SoftID_DEN_TrackerMuons_PV_1 ./example/flat/muon/generalTracks/JPsi/Run2025/JPsi_pythia8/Nominal/NUM_SoftID_DEN_TrackerMuons_PV.root Nominal data Nominal JPsi True
[#1] INFO:ObjectHandling -- RooWorkspace::import(wmain) importing dataset hGenPass
[#1] INFO:ObjectHandling -- RooWorkspace::import(wmain) importing dataset hGenFail
[#1] INFO:ObjectHandling -- RooWorkspace::import(wmain) importing dataset hPass
[#1] INFO:ObjectHandling -- RooWorkspace::import(wmain) importing dataset hFail
[#0] ERROR:Caching -- The internal binning of variable x is not uniform. The numerical FFT will likely yield wrong results.
[#1] INFO:Minimization -- p.d.f. provides expected number of events, including extended term in likelihood.
[#1] INFO:Eval -- RooRealVar::setRange(x) new range named 'refrange_fft_sigPass' created with bounds [2.7,3.5]
[#0] WARNING:Eval -- The FFT convolution 'sigPass' will run with 20 bins. A decent accuracy for difficult convolutions is typically only reached with n >= 1000. Suggest to increase the number of bins of the observable 'x'.
[#1] INFO:Caching -- RooAbsCachedPdf::getCache(sigPass) creating new cache 0xec2af20 with pdf sigPhysPass_CONV_sigResPass_CACHE_Obs[x]_NORM_x for nset (x) with code 0
[#1] INFO:Fitting -- RooAbsPdf::fitTo(pdfPass) fixing normalization set for coefficient determination to observables in data
[#1] INFO:Fitting -- using CPU computation library compiled with -mavx2
[#1] INFO:Fitting -- RooAddition::defaultErrorLevel(nll_pdfPass_hPass) Summation contains a RooNLLVar, using its error level
[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: activating const optimization
 **********
 **    1 **SET PRINT           1
 **********
 **********
 **    2 **SET NOGRAD
 **********
 PARAMETER DEFINITIONS:
    NO.   NAME         VALUE      STEP SIZE      LIMITS
     1 acmsP        2.90000e+00  2.00000e-01    2.00000e+00  4.00000e+00
     2 betaP        5.00000e-02  9.90000e-03    1.00000e-03  1.00000e-01
     3 gammaP       1.00000e-01  4.00000e+00   -2.00000e+01  2.00000e+01
     4 meanP       -0.00000e+00  1.00000e+00   -5.00000e+00  5.00000e+00
     5 nBkgP        1.98600e+03  9.92750e+02    5.00000e-01  2.18460e+04
     6 nSigP        1.93803e+04  1.23285e+03    5.00000e-01  2.18460e+04
     7 sigmaP       9.00000e-01  4.47500e-01    5.00000e-03  5.00000e+00
 **********
 **    3 **SET ERR         0.5
 **********
 **********
 **    4 **SET PRINT           1
 **********
 **********
 **    5 **SET STR           1
 **********
 NOW USING STRATEGY  1: TRY TO BALANCE SPEED AGAINST RELIABILITY
 **********
 **    6 **MINIMIZE        3500           1
 **********
 FIRST CALL TO USER FUNCTION AT NEW START POINT, WITH IFLAG=4.
[#1] INFO:NumericIntegration -- RooRealIntegral::init(bkgPass_Int[x]) using numeric integrator RooIntegrator1D to calculate Int(x)
[#0] WARNING:Eval -- The FFT convolution 'sigPass' will run with 20 bins. A decent accuracy for difficult convolutions is typically only reached with n >= 1000. Suggest to increase the number of bins of the observable 'x'.
[#1] INFO:Caching -- RooAbsCachedPdf::getCache(sigPass) creating new cache 0xedb63b0 with pdf sigPhysPass_CONV_sigResPass_CACHE_Obs[x] for nset () with code 1 from preexisting content.
 START MIGRAD MINIMIZATION.  STRATEGY  1.  CONVERGENCE WHEN EDM .LT. 1.00e-03
RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=2.9	betaP=0.00100888	gammaP=0.1	meanP=-0.00831991	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502489
RooAbsPdf::sigPass_over_sigPass_Int[x][ numerator=sigPass denominator=sigPass_Int[x] ]
     p.d.f value is less than zero, trying to recover @ numerator=sigPass=17458.1, denominator=sigPass_Int[x]=2927.24
     p.d.f value is less than zero, trying to recover @ numerator=sigPass=17458.1, denominator=sigPass_Int[x]=2927.24
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 5.96401,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 5.96401,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=2.9	betaP=0.00100888	gammaP=0.1	meanP=0.000831991	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502489
RooAbsPdf::sigPass_over_sigPass_Int[x][ numerator=sigPass denominator=sigPass_Int[x] ]
     p.d.f value is less than zero, trying to recover @ numerator=sigPass=372.723, denominator=sigPass_Int[x]=38.3079
     p.d.f value is less than zero, trying to recover @ numerator=sigPass=372.723, denominator=sigPass_Int[x]=38.3079
     p.d.f value is less than zero, trying to recover @ numerator=sigPass=372.723, denominator=sigPass_Int[x]=38.3079
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 9.72967,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 9.72967,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 9.72967,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=2.9	betaP=0.00100888	gammaP=0.1	meanP=-0.000831991	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502489
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 6.9725,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=2.9	betaP=0.00100888	gammaP=0.1	meanP=8.31991e-05	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502489
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.5085,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=2.9	betaP=0.00100888	gammaP=0.1	meanP=-8.31991e-05	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502489
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.19444,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=2.9	betaP=0.00100888	gammaP=0.1	meanP=0	nBkgP=1986	nSigP=19380.3	sigmaP=0.00500226
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35148,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=2.9	betaP=0.00100888	gammaP=0.1	meanP=0	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502153
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35148,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=2.9	betaP=0.00100888	gammaP=0.1	meanP=0	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502524
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35148,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99906	betaP=0.00100888	gammaP=0.1	meanP=-8.31991e-06	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502489
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.33575,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99906	betaP=0.00100888	gammaP=0.1	meanP=-6.04829e-06	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502489
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.34004,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99906	betaP=0.00100888	gammaP=0.1	meanP=0	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502485
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35148,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99906	betaP=0.00100888	gammaP=0.1	meanP=0	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502492
RooAbsPdf::sigPass_over_sigPass_Int[x][ numerator=sigPass denominator=sigPass_Int[x] ]
     p.d.f value is less than zero, trying to recover @ numerator=sigPass=264.695, denominator=sigPass_Int[x]=31.6944
     p.d.f value is less than zero, trying to recover @ numerator=sigPass=264.695, denominator=sigPass_Int[x]=31.6944
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35148,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35148,bkgPass_over_bkgPass_Int[x] = 1.24967), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99906	betaP=0.0998574	gammaP=0.1	meanP=-6.04829e-07	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502489
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35033,bkgPass_over_bkgPass_Int[x] = 1.25042), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99906	betaP=0.0998574	gammaP=0.1	meanP=-6.04829e-08	nBkgP=1986	nSigP=19380.3	sigmaP=0.00502489
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35136,bkgPass_over_bkgPass_Int[x] = 1.25042), !coefficients=(nSigP = 19380.3,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99906	betaP=0.0998574	gammaP=0.1	meanP=-6.04829e-08	nBkgP=1986	nSigP=13892.1	sigmaP=0.00502489
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35136,bkgPass_over_bkgPass_Int[x] = 1.25042), !coefficients=(nSigP = 13892.1,nBkgP = 1986)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99906	betaP=0.0998574	gammaP=0.1	meanP=-6.04829e-08	nBkgP=1986	nSigP=13892.1	sigmaP=0.00502489
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35136,bkgPass_over_bkgPass_Int[x] = 1.25042), !coefficients=(nSigP = 13892.1,nBkgP = 1986)

 FCN=-170252 FROM MIGRAD    STATUS=INITIATE      156 CALLS         157 TOTAL
                     EDM= unknown      STRATEGY= 1      NO ERROR MATRIX       
  EXT PARAMETER               CURRENT GUESS       STEP         FIRST   
  NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE 
   1  acmsP        3.99906e+00   2.00000e-01   0.00000e+00   4.29182e-02
   2  betaP        9.98574e-02   9.90000e-03   0.00000e+00  -3.80316e-01
   3  gammaP       1.00000e-01   4.00000e+00   0.00000e+00   1.52361e+03
   4  meanP        0.00000e+00   1.00000e+00   0.00000e+00  -5.46274e+11
   5  nBkgP        1.98600e+03   9.92750e+02   0.00000e+00  -4.31787e+03
   6  nSigP        1.38921e+04   1.23285e+03  -6.10284e-01  -1.73572e+02
   7  sigmaP       5.02489e-03   4.47500e-01   0.00000e+00   3.46777e-05
                               ERR DEF= 0.5
RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=4	betaP=0.1	gammaP=-0.394922	meanP=3.0242e-08	nBkgP=3801.06	nSigP=14132.5	sigmaP=0.0050251
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35153,bkgPass_over_bkgPass_Int[x] = 1.24149), !coefficients=(nSigP = 14132.5,nBkgP = 3801.06)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99928	betaP=0.0998909	gammaP=0.0380115	meanP=-2.26068e-09	nBkgP=2185.34	nSigP=13922.2	sigmaP=0.00502491
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35147,bkgPass_over_bkgPass_Int[x] = 1.25019), !coefficients=(nSigP = 13922.2,nBkgP = 2185.34)

RooAbsMinimizerFcn: Minimized function has error status.
Returning maximum FCN so far (-157036) to force MIGRAD to back out of this region. Error log follows.
Parameter values: 	acmsP=3.99928	betaP=0.0998909	gammaP=0.0380115	meanP=-2.26068e-09	nBkgP=2185.34	nSigP=13922.2	sigmaP=0.00502491
RooAddPdf::pdfPass[ nSigP * sigPass_over_sigPass_Int[x] + nBkgP * bkgPass_over_bkgPass_Int[x] ]
     getLogVal() top-level p.d.f evaluates to NaN @ !refCoefNorm=(x = 3.1), !pdfs=(sigPass_over_sigPass_Int[x] = 8.35147,bkgPass_over_bkgPass_Int[x] = 1.25019), !coefficients=(nSigP = 13922.2,nBkgP = 2185.34)
```

Hi @nicolo.trevisani,

thank you for your question. I will add that LCG_105 should have ROOT 6.30.02 while LCG_107 should have ROOT 6.32.02. I will let @jonas comment on which changes in between these ROOT releases could have caused the behaviour you are seeing.

Cheers,

Marta