gmane.comp.lang.c++.root

Re: SetTitleOffset for TEfficiency

Marek Sirendi — 2013-05-16T16:48:51

SetTitleOffset for TEfficiency

Marek Sirendi — 2013-05-16T15:46:07

Re: Bug in TMatrixTBase

Lorenzo Moneta — 2013-05-15T07:28:27

Hello Omar,

 Thank you for reporting this bug. I will commit now the fix in the master and the 5.34 patches

 Best Regards

 Lorenzo
On May 13, 2013, at 6:35 PM, Omar Andrés Zapata Mesa <andresete.chaos< at >gmail.com<mailto:andresete.chaos< at >gmail.com>> wrote:

Hello dear ROOTers,
I found a bug in TMatrixTBase using Option "F" for fortran representation in the method GetMatrix2Array(Element *data,Option_t *option).

When I get tha matrix's data in an array with fortran representation and I try to create a matrix again using TMatrixT(nrows,ncols,*data,"F"), the matrix is corrupted.

I attach the macro to test the bug and the patch.

Thanks.

--
Omar Andres Zapata Mesa
Fundamental Interaction Phenomenology Gruop (Gfif)
Head Developer
http://gfif.udea.edu.co<http://gfif.udea.edu.co/>
Gfif Developers (Gfif Dev)
Division of computer science
System Engineering Student
Universidad de Antioquia At Medellin - Colombia
Usuario Linux  #490962
<bug.C><TMatrixTBase.cxx.diff>

Re: GetConfidenceIntervals in Minuit2?

Lorenzo Moneta — 2013-05-15T06:54:29

Hello Marcos, 

 This method is implemented if you use the ROOT::Math::Fitter class as interface for fitting and not directly Minuit2. The ROOT::Math::FitResult class which is produced after fitting has then a method, GetConfidenceIntervals, 
see  http://root.cern.ch/root/html/ROOT__Fit__FitResult.html#ROOT__Fit__FitResult:GetConfidenceIntervals< at >1

If you need an example on how to fit using the ROOT::Math::Fitter class, look at the tutorial fit/combinedFit.C

http://root.cern.ch/root/html534/tutorials/fit/combinedFit.C.html

  Best Regards

 Lorenzo

 
On May 13, 2013, at 6:18 PM, Marcos Fernandez Garcia <mfg< at >mail.cern.ch> wrote:

Bug in TMatrixTBase

Omar Andrés Zapata Mesa — 2013-05-13T16:35:39

Hello dear ROOTers,
I found a bug in TMatrixTBase using Option "F" for fortran representation
in the method GetMatrix2Array(Element *data,Option_t *option).

When I get tha matrix's data in an array with fortran representation and I
try to create a matrix again using TMatrixT(nrows,ncols,*data,"F"), the
matrix is corrupted.

I attach the macro to test the bug and the patch.

Thanks.

GetConfidenceIntervals in Minuit2?

Marcos Fernandez Garcia — 2013-05-13T16:18:18

Hello

I have a chi2 fit using the TMinuit2 package, so I am doing things like:

    MnUserParameters upar;
    upar.Add(...);
    MnMigrad mn( mychi2 , upar );
    FunctionMinimum min = mn();
    for (Int_t i=0; i<parini.size();i++) {
      parini[i]=min.UserState().Value(i);
      parerr[i]=min.UserState().Error(i);
    }

I wanted to plot the fitted function with the confidence intervals as bin 
errors, as TFitter's GetConfidenceIntervals does. I checked 
the code in:

http://root.cern.ch/root/html/src/TFitter.cxx.html#RIPJhE

I expected TMinuit2 to have such a method since it can calculate the 
covariance matrix and parameter gradients.

Is it already done by anyone?

Thank you


 Marcos

Re: X axis title in a stack histogram

Olivier Couet — 2013-05-09T16:57:12

Try gPad->Update() after the stack drawing and set the title after the Update



Le 9 mai 2013 à 10:31, "Pilar Casado" <casado< at >ifae.es> a écrit :

Re: Re: X axis title in a stack histogram

Tian MA — 2013-05-09T08:59:25

I guess you should draw the stack at first, such as hs->Draw(), then update the canvas. 

Best Wishes,
Martin

在2013-05-09 16:31:28,Tian MA<mat< at >ihep.ac.cn>写道：

Re: X axis title in a stack histogram

Pilar Casado — 2013-05-09T08:31:28

Hello Marcos,

It doesn't work. I get:
Error: illegal pointer to class object GetXaxis() 0x0 446  StackPlots.C:254:
*** Interpreter error recovered ***

Thanks, Pilar

El 09/05/2013 10:01, Marcos Fernandez Garcia escribió:

Re: X axis title in a stack histogram

Marcos Fernandez Garcia — 2013-05-09T08:01:24

In principle, if hs is a pointer to your THStack, it should be enough 
doing:

hs->GetXaxis()->SetTitle( "The X axis title" );

 Marcos

On Thu, 9 May 2013, Pilar Casado wrote:

X axis title in a stack histogram

Pilar Casado — 2013-05-09T07:58:23

Hello,

How can I set the X axis title of a stack histogram.
I have tried the usual SetTitle on the stack histogram,
have also set two of the histogram in the stack, but
I don't manage.

Many thanks, Pilar

Re: TTimer time resolution

Konstantin Olchanski — 2013-05-07T23:43:30

10ms is 100Hz is the normal time resolution of the UNIX/Linux clock, related
timers and timeouts (select(), sundry sleep(), etc).

For higher precision sleep and time delay, you need to use specialized
Linux "real time" facilities. Google for HZ=1000, high-precision timers, etc.

If you need very short delays, say under 0.1ms, you may find that you cannot
make linux sleep so little and you have to use delay loops. In which case
you can use the VME bus access as a delay - on most hardware, a VMEbus single word
read access takes around 1us. So delay of 100us (0.1ms) is 100 VMEbus reads in a loop.

TTimer time resolution

Irakli Keshelashvili — 2013-05-07T11:36:44

Dear ROOTers,

I'm using TTimer for VMEbus readout. As I learned, the minimum time resolution is 10ms as it is defined in TSystem.h -

kItimerResolution = 10      // interval-timer resolution in ms

Question: is it O.K. if one changes it to 1ms? Or is there some other method to decrease time interval?

Best regards
Irakli

* * * * * * * * * * * * * * *
 Dr. Irakli Keshelashvili
 Department of Physics
 University of Basel
 Klingelbergstrasse 82,
 CH-4056 Basel, Switzerland
 Tel. +41 6126 73775 (Office)
      +41 7887 06657 (Mobile)
 Fax: +41 6126 73784
 * * * * * * * * * * * * * * *

Re: ROOT and GlusterFS

Philippe Canal — 2013-05-06T18:58:34

Hi Doug,

Is this still an issue with 5.34?

Philippe.

On 1/8/13 6:58 PM, Doug Schouten wrote:

Re: extracting systematics from profile

Nabil Ghodbane — 2013-05-03T14:03:04

hi,
just to mention that i got a reply on the forum and the reply can be 
found at this link:
http://root.cern.ch/phpBB3/viewtopic.php?f=15&t=16415&p=70269#p70269
see you.

Nabil Ghodbane
Skype: nabilghodbane

On 05/02/2013 06:02 PM, Nabil Ghodbane wrote:

Re: TH1F* hists froma THStack plot

Olivier Couet — 2013-05-03T13:34:50

In my collection of examples I found one using GetHists which may help you:

----- file stackerrors2.C
void DrawStackE1 (THStack *hs)
{
   hs->Draw();
   TH1D *h0 = (TH1D*)hs->GetHistogram();
   if(h0)h0->Draw();

   TList *hl;
   hl = hs->GetHists();
   TIter iter1(hl);
   TH1D *h;
   Double_t x1=9999.,y1=9999.,x2=-9999.,y2=-9999.;
   while((h=((TH1D*)iter1()))){
      h->Draw("E1 SAME");
   }
}

void stackerrors2()
{
   TH1D *h1 = new TH1D("h1","h1",10,0,10);
   TH1D *h2 = new TH1D("h2","h2",10,0,10);
   THStack *h = new THStack();

   h1->SetLineColor(kRed);
   h1->SetMarkerStyle(20),
   h2->SetLineColor(kBlue);
   h2->SetMarkerStyle(21);

   for(int i=0; i<11; i++){
      h1->SetBinContent(i,1.5-i/10);
      h1->SetBinError(i,0.5*i);
      h2->SetBinContent(i,10.5-i/10);
      h2->SetBinError(i,0.7*i);
   }

   h->Add(h1);
   h->Add(h2);
   DrawStackE1 (h);
}
-----




On May 3, 2013, at 3:29 PM, Maria Pilar Casado <casado< at >ifae.es<mailto:casado< at >ifae.es>>
 wrote:

Dear root expert,

I'm using root version 5.34/05  14 February 2013

I would need to retrieve the list of histograms from a THStack histogram
and add them into another histogram. I'm trying to do it in this way by
I get the complain you can see at the end. How can I do it?

Many thanks, Pilar

 TH1F * h_signalf = new TH1F("h_signalf","h_signalf",50,0.,500.);
 TList *thisList = h_stack->GetHists();

 TObjLink *lnk = thisList->FirstLink();
 while (lnk) {
  //h_signalf->Add( (TH1F*) lnk->GetObject());
  h_signalf->Add( (TH1F*) lnk);
  lnk = lnk->Next();
 }
--------------------------------
Error:
Error: Can't call TObjLink::Next() in current scope KolmoTest.C:45:
Possible candidates are...
(in TObjLink)

--
----------------------------------------------------------------
Dept. de Física & IFAE,
Facultat Ciències Nord UAB, E-08193 Bellaterra (Barcelona) Spain

Tel. (Barcelona): +34 93 581 42 85
Tel. (CERN):      +41 22 767 11 25

Avís - Aviso - Legal Notice - (LOPD) - http://legal.ifae.es

TH1F* hists froma THStack plot

Maria Pilar Casado — 2013-05-03T13:29:14

Dear root expert,

I'm using root version 5.34/05  14 February 2013

I would need to retrieve the list of histograms from a THStack histogram
and add them into another histogram. I'm trying to do it in this way by
I get the complain you can see at the end. How can I do it?

Many thanks, Pilar

   TH1F * h_signalf = new TH1F("h_signalf","h_signalf",50,0.,500.);
   TList *thisList = h_stack->GetHists();

   TObjLink *lnk = thisList->FirstLink();
   while (lnk) {
    //h_signalf->Add( (TH1F*) lnk->GetObject());
    h_signalf->Add( (TH1F*) lnk);
    lnk = lnk->Next();
   }
--------------------------------
Error:
Error: Can't call TObjLink::Next() in current scope KolmoTest.C:45:
Possible candidates are...
(in TObjLink)

Re: ROOT has moved to Git...

Konstantin Olchanski — 2013-05-03T05:25:53

Confirmed. Looks good. This is much better than svn - I do not have to hunt down the tag names on the web.


My complete source+binary kits take about 3GB for root_v5.34.01_SL62_64
and about 4GB for today's root_master_SL64_64 (much disk space is consumed
by libroota.a and roota).

So I concur - I do not need to worry about space taken by the sources.

(I guess this is one more battle between whale and elephant - "git clone"
stores the complete revision history deduped and compressed - while
"svn checkout" keeps 2 copies of each file, uncompressed, unduped. Plus
on an ext3/ext4 filesystem, "svn checkout" consumes more disk blocks
because git blobs are all inside one large "pack" file while svn files
are all individual files (a "1 byte" file consumes 4Kibytes (1 block)
of disk space).

(We are moving to convert the MIDAS DAQ sources from svn to git,
so have great interest in this type of conversions).

Re: Mangled files? Re: ROOT has moved to Git...

Konstantin Olchanski — 2013-05-03T05:12:48


Confirmed. Clean git clone built without errors. SL6.4-64. Thanks.

(I never figured out how to fix the botched git clone - had
to erase the whole thing...)


K.O.

Re: Mangled files? Re: ROOT has moved to Git...

Fons Rademakers — 2013-05-02T23:18:52

Hi Konstantin,

   I've done that.

Cheers, Fons.


On 03/05/2013 01:02, Konstantin Olchanski wrote:

Re: Mangled files? Re: ROOT has moved to Git...

Konstantin Olchanski — 2013-05-02T23:02:11


I now think it is your system that's broken. According, to GIT documentation at
ftp://www.kernel.org/pub/software/scm/git/docs/gitattributes.html

"Any byte sequence that begins with $Id: and ends with $ in the worktree file is replaced with $Id$ upon check-in".

I read this as "$Id(.*)$" is replaced by "$Id hexhexhex $".

Observe how in the example file above, the first "$" is in "$Id", the next "$"
is part of the first "$ROOTSYS".

Indeed, that is what I observe in my broken files: the text between
the first "$Id" and the "$" in the first "$ROOTSYS" disappears.

The files that are not mangled, have an extra "$" in them and the part
of the copyright notice up to the first "$ROOTSYS" survives.

This may be a change in git between 1.7.1 (my version) and 1.7.10 (your version), have no idea
what Fons uses (but I note that git on lxplus is 1.7.4, again, different from vanilla SL6).

As I understand, Linus Torvalds declared "$Id$ is evil", so I vote to remove the "ident" keyword from .gitattributes.