============
Using OpenMP                                                    [TOC]
============

Changes:

- `gemm.hpp` uses OpenMP if you compile with `-fopenmp`.

Tar-Ball for this Session
=========================
The tar-ball __session3.tgz__ contains the files:

---- SHELL ---------------------------------------------------------------------
tar tfvz session3.tgz
--------------------------------------------------------------------------------

:links: session3.tgz -> http://www.mathematik.uni-ulm.de/~lehn/test_ublas/session3.tgz


Compile and Run Benchmark
=========================
For a quick benchmark we again limit $m$ to $500$ with `-DM_MAX=500`. This
time we also enable the optimized micro kernel for AVX with `-DHAVE_AVX`:

---- SHELL (path=session3,hostname=heim) ---------------------------------------
g++ -Ofast -mavx -Wall -std=c++11 -DNDEBUG -DHAVE_AVX -fopenmp -DM_MAX=500 -I ../boost_1_60_0 matprod.cc
./a.out
--------------------------------------------------------------------------------

Core Function for Matrix-Matrix Produkt
=======================================
:import: session3/gemm.hpp

Benchmark Results
=================

---- SHELL (path=session3,hostname=heim) ---------------------------------------
g++ -Ofast -mavx -Wall -std=c++11 -DHAVE_AVX -fopenmp -DNDEBUG -I ../boost_1_60_0 matprod.cc
./a.out  > report.session3
cat report.session3
gnuplot plot.session3.mflops
gnuplot plot.session3.time
gnuplot plot.session3.time_log
--------------------------------------------------------------------------------


MFLOPS
------
---- IMAGE -----------------------
session3/bench.session3.mflops.svg
----------------------------------


Time
----

---- IMAGE -----------------------
session3/bench.session3.time.svg
----------------------------------

Time with Logarithmic scale
---------------------------

---- IMAGE -------------------------
session3/bench.session3.time_log.svg
------------------------------------


:navigate: back  -> doc:session2/page01
           next  -> doc:session4/page01