Hi Masa, Orjan, I do not have enough information to come with a diagnostic. There should be no problems in storing 2 million entries (even in memory) with two numbers only in each entry. Could you send me the result of tree.Print() showing the internal structure of your ntuple and an example of macro reading this ntuple where you see the memory limitations? Rene Brun On Sat, 14 Jul 2001, Masaharu Goto wrote: > Hello Orjan, > > Thank you for your message. > > The 200,000 elements limit may come from physical memory > resources and/or swap space you have on your computer. Since > it consumes 1.6Gbytes, it will be very difficult to load everything > on memory at once unless you have a big machine. > > Unfortunately, I do not have a direct solution to it now. > > Rene, > Do you have any comments? > > > Thank you > Masaharu Goto > > > > >Date: Wed, 11 Jul 2001 16:33:46 +0200 > >From: Orjan Nordhage <nordhage@tsl.uu.se> > >To: MXJ02154@niftyserve.or.jp > >Subject: A C++ question... > > > >Hello! > > > >Short presentation: I'm working at TSL at University of Uppsala in the > >WASA-group, looking at reactions like pp->pp + more. The detector result > >is a so called ntuple-file, which I let a C++-program read. The program > >then produces an outfile, wich contains 2 columns of numbers, which I > >use ROOT for plotting. > > > >However, this works just fine, ROOT is a great tool for this kind of > >things. But the problem is that the ntuple-file is very large, that is > >contains very many numbers, and I store them in a vector. That also work > >just fine until I want to read the whole file and store every important > >value, like 1 or 2 millions. Unfortunately, C++ has this limit of about > >200.000 elements for a double vector. Now, I wonder, do You know how to > >get around this problem? Can I include som directory <supervectors.h> or > >something? Or do You have any other suggestion? Of course, I can divide > >the reading into parts, and define several vectors, but I prefer not to, > >since the storing process is quite advanced as it is. > > > >Yours sincerely > > > > / ヨrjan Nordhage > > >
This archive was generated by hypermail 2b29 : Tue Jan 01 2002 - 17:50:52 MET