#copy this directory
# scp -r esc09-master:/home/innocente/ExercisesMT .
cp -r  /nfsmaster/innocente/ExercisesMT .

#copy tcmalloc
# scp  esc09-master:/home/innocente/tcmalloc.tgz
cp -r  /nfsmaster/innocente/tcmalloc.tgz .
tar -zxf tcmalloc.tgz
export LD_PRELOAD=$HOME/tcmalloc/lib/libtcmalloc.so
export LD_PRELOAD=
echo $LD_PRELOAD 

# ---------------------------------

setCXX : set c++ flags

RealTime.h: a precise measurement of wall clock time (actually clock cycles...)


future.h:  gcc <future> header file missing in 4.4.1
future.cc:  library component of the above

to compile and run:
cd ExercisesMT
source setCXX 
g++441 $CXXFLAGS exFuture.cpp future.cc
./a.out 

-----------------------------------------

exAtomics.cpp: example of use of atomics
exFuture.cpp : examples of use of future
exRandom.cpp : example of use of <random>

exercises:
  look in these files, "play" with code constructs and syntax to solve you own preferred "code-puzzle"


helloWordMT.cpp : example of thread fork, join and concurrency

exercises
   introduce an atomic, a lock to solve the concurrency problem (and transform it in a trivial sequential program :-(

-----------------------------------------

DistributedLocks.cpp: examples of shared memory access patterns

exercises:
   try various memory access pattern, 
  global  lock,  
  distributed locks, 
  atomics, 
  local memory

  why it aborts?

-----------------------------------------


 SharedQueue.h:  simple lock free implementation of a FIFO
  exQueue.cpp: example of use od the above

exercises:
  modify it to use condition variables, 
  try it in multiple "pushers" 

-----------------------------------------

ParallelSort.cpp: Divide&Conquer implementation of sorting


exercises:

    use futures,
    use OpenMP 

-----------------------------------------

exSharedList.cpp: minimal implementation of lock free linked-list (mainly a malloc exercise

use tcmalloc
use a memory pool


-----------------------------------------


HistoMapReduce.cpp Map/Reduce, Master/Worker, pull-mode implementation of parallel histogramming


HistoFuture.cpp  Map/Reduce, Master/Worker, round-robin push-mode implementation of parallel histogramming using futures


profile
observe thread scheduling, starvation
parallelize the generation of the image

-----------------------------------------


ClusterizerPart1.cpp  some components to implelement the kmean clusterizer

using Histograming as example, implement the iterative kmean clusterizer

-----------------------------------------