RE: [ROOT] Event handling or graphing huge datasets

From: Fons Rademakers (Fons.Rademakers@cern.ch)
Date: Mon Mar 04 2002 - 17:25:36 MET


Clearly some X11 limit is reached here. We should break up the single
polyline in a set of polylines.

Cheers, Fons.



On Mon, 2002-03-04 at 16:53, Colley, Tony wrote:
> Hello again.
> 
> I was mistaken... I was accidently testing with 10^6 points and not 10^7
> points. That extra order of magnitude causes a crash in ROOT with the
> following message:
> 
>   *** BREAK *** write on a pipe with no one to read it
>   Error in <RootX11IOErrorHandler>: fatal X11 error (connection to server
> lost?)
>   *** Save data and exit application ***
> 
> So I did a little testing with how many points it would work with.
> 
> If I have 1048570 (or fewer) data points, everything works fine.
> 
> If I have 1048571 data points, the graph is displayed fine, but I get the
> following message:
>    Error in <RootX11ErrorHandler>: BadMatch (invalid parameter attributes)
> (XID: 37750999, XREQ: 42)
> 
> If I have 1048572 data points, the border/axes of the graph are displayed
> fine, but there is no plot of the data, and I get the following message:
>    Error in <RootX11ErrorHandler>: BadLength (poly request too large or
> internal Xlib length error) (XID: 0, XREQ: 65)
> 
> If I have 1048573 (or more) data points, ROOT crashes and I get the same
> error message as for 10^7 points.
> 
> This failure limit is suspiciously close to 2^20 (1048576).
> 
> Anyone else successful at plotting more than 2^20 data points?
> 
> In case you don't have the original email that started this thread... I am
> using TGraph for plotting the x,y data points.
> 
> +------------------------------------------------+
> | Tony Colley                ITT Industries A/CD |
> | Modeling/Simulation Group  Fort Wayne, IN  USA |
> +------------------------------------------------+
> | Using ROOT 3.03/02 on RedHat Linux 7.2         |
> | Last CVS update: 26 Feb 2002 at 2218 GMT       |
> +------------------------------------------------+
> 
> 
> -----Original Message-----
> From: Faine, Valeri [mailto:fine@bnl.gov]
> Sent: Saturday, March 02, 2002 21:06
> To: 'Eddy Offermann '; 'Tony.Colley@itt.com '
> Cc: 'roottalk@pcroot.cern.ch '
> Subject: RE: [ROOT] Event handling or graphing huge datasets
> 
> 
>  
>   Hello Eddy,
> 
>   I am not sure you correct. 
> 
> It is true "TGraph::PaintGrapHist" does take in account the screen 
> resolution but ExecuteEvent.
> see:
> http://root.cern.ch/root/htmldoc/src/TGraph.cxx.html#TGraph:ExecuteEvent
> 
>  When one tries to move the graph (may be accidently, just the wrong mouse
> button was pressed) this entails the graph of the all fNpoints
> to be redrawn at least twice: The first time to clear the old graph position
> and the second time to draw it at the new place. Each of 10^7 points is
> drawn with 6 calls to TVirtualX::DrawLine sending at list 4 coordinate
> numbers as an X11-message. This means just one touched 
> (occasionally)  the middle button of the mouse the (4*2) * 2 * 6 * 10^7
> bytes will be sent out to X-terminal. I believe this may shock the
> networking.
> Assuming we have 10Mbit / sec network one can calculate that 10^10 / 10 ^7
> = 1000 sec per single mouse step (this assumes the X-terminal response time
> is as fast as its networking connection) 
> 
> 
>  
> 
> -----Original Message-----
> From: Eddy Offermann
> To: Tony.Colley@itt.com
> Cc: roottalk@pcroot.cern.ch
> Sent: 2/26/02 12:46 PM
> Subject: Re: [ROOT] Event handling or graphing huge datasets
> 
> Hi Tony,
> 
> Take a look at "void TGraph::PaintGrapHist" and scan for "lowRes".
> You will see around line 2350 that if you draw the graph with option L
> or P,
> The number of data points is checked versus your screen resolution.
> In case of many data points it will not draw every point but just
> calculate the
> average pixel position. 10^7 should not be a problem
> 
> 
> Eddy A.J.M. Offermann
> Renaissance Technologies Corp.
> Route 25A, East Setauket NY 11733
> e-mail: eddy@rentec.com
> http://www.rentec.com
> 
> --- END OF INCLUDED MESSAGES ---
> 
> \I'm not responsible for anything below this line/
> 
> ************************************ 
> If this email is not intended for you, or you are not responsible for the
> delivery of this message to the addressee, please note that this message may
> contain ITT Privileged/Proprietary Information.  In such a case, you may not
> copy or deliver this message to anyone.  You should destroy this message and
> kindly notify the sender by reply email.  Information contained in this
> message that does not relate to the business of ITT is neither endorsed by
> nor attributable to ITT. 
> ************************************ 
-- 
Org:    CERN, European Laboratory for Particle Physics.
Mail:   1211 Geneve 23, Switzerland
E-Mail: Fons.Rademakers@cern.ch              Phone: +41 22 7679248
WWW:    http://root.cern.ch/~rdm/            Fax:   +41 22 7679480



This archive was generated by hypermail 2b29 : Sat Jan 04 2003 - 23:50:43 MET