Skip to content
/ GooFit Public
forked from GooFit/GooFit

Massively-parallel framework for maximum-likelihood fits, implemented in CUDA.

License

Notifications You must be signed in to change notification settings

hassec/GooFit

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Fork of GooFit used for development during my masters thesis

The plan is to clean this up a little in the future and extend the documentation. Nevertheless this is the place where one can find the state of GooFit at the end of my thesis. Further improvement and fixes won't be propagated back into this branch.


GooFit is a massively-parallel framework, written in CUDA, for doing maximum-likelihood fits. It is also possible to build GooFit using OpenMP.

Requirements

  • If using CUDA:
  • CUDA 4.2 or 5.x
  • An nVidia GPU supporting compute capability at least 2.0
  • If building for OpenMP:
  • a compiler supporting OpenMP

Installation

  • Clone with git:

      git clone git://github.com/GooFit/GooFit.git
    
  • check that the path setting in Makefile for CUDALOCATION is correct for your system

  • If using OpenMP, CUDALOCATION should point at ./fakecuda

  • Set TARGET_OMP = 1 in the Makefile

  • Put ./fakcuda on your PATH, e.g. export PATH=$PATH:$HOME/GooFit/fakecuda (n.b. the use of g++ is hardwired in fakecuda/nvcc; this is bad)

  • Install a copy of CUDA's thrust in ./thrust (there's nothing to compile)

  • then compile with gmake.

As an alternative to modifying the Makefile, you can pass the values as e.g. gmake CUDALOCATION=$HOME/Src/GooFit/fakecuda TARGET_OMP=1

Building the Examples Fits

  • If you have Root on your system,

  • set the path to full root installation (bash syntax) export ROOTSYS=

  • If you don't have Root, do NOT set ROOTSYS

  • check that the path setting in examples/simpleFit/Makefile for CUDALOCATION is correct for your system

  • If using OpenMP, CUDALOCATION should point at ./fakecuda

  • Set TARGET_OMP = 1 in the Makefile

  • change to the examples/simpleFit directory and buid the program(s): cd examples/simpleFit ; gmake all (you'll recall that you can set CUDALOCATION and TARGET_OMP on the command line instead of modifying the Makefile)

Running the Example Fit

  • Setup the bin and library paths for root

      export PATH=$PATH:$ROOTSYS/bin:.
      # On Linux:
      export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$ROOTSYS/lib:<GooFit path>/rootstuff
      # On Mac:
      export DYLD_LIBRARY_PATH=$DYLD_LIBRARY_PATH:$ROOTSYS/lib:<GooFit path>/rootstuff
    

(If you don't have Root, don't modify the PATH and omit $ROOTSYS/lib:)

  • To run the simple fit example type: ./simpleFitExample (If you didn't define ROOTSYS the data and fits will be written to stdout)

Acknowledgement

GooFit's development has been supported by the National Science Foundation under grant number NSF-1005530. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the developers and do not necessarily reflect the views of the National Science Foundation.

About

Massively-parallel framework for maximum-likelihood fits, implemented in CUDA.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • C++ 57.1%
  • Cuda 31.8%
  • TeX 10.5%
  • Other 0.6%