GitLab

This should be refactored, but heh?!

Empty CMakeLists added to BenchEfficiency repertory in order to be able to compile FMMApi, CKernelApi and/or others addons

# By Florent Pruvost (2) and Samuel Pitoiset (2)
# Via Samuel Pitoiset
* 'master' of git+ssh://scm.gforge.inria.fr//gitroot/scalfmm/scalfmm:
  GroupTree: fix start/stop profiling with KStar
  GroupTree: allow to start/stop StarPU profiling
  add the possibility to build scalfmm as a dynamic lib (static by default)
  sync with https://scm.gforge.inria.fr/anonscm/svn/morse/trunk/morse_distrib/cmake_modules/morse/find

This will also be used when building ScalFMM with KStar to compute
efficiencies from the generated trace.rec file.

sync with https://scm.gforge.inria.fr/anonscm/svn/morse/trunk/morse_distrib/cmake_modules/morse/find

Changes in API, new call back for M2M and L2L where we provide the array of cells instead of calling the M2M over each child

Major modification! Defined rigoreous target-source convention: p(x)=sum_y K(x,y)w(y) with x=target and y=source. Direct computation is done by calling MutualParticles(Target params,Source params) or P2P(Target, Source); Kernel matrices evaluation is done by calling evaluate(x=target,y=source) which returns k(x,y)=k(x-y).

# By Luka Stanisic
# Via Luka Stanisic
* 'master' of git+ssh://scm.gforge.inria.fr//gitroot/scalfmm/scalfmm:
  minor fix of the memcpy function call