* basic support for multicore CPU via a .evalOMP() which

internaly uses OpenMP if enabled at compile time.
 * added a bench/ folder with a couple benchmarks and benchmark tools.

bench/benchmarkX.cpp

@@ -0,0 +1,22 @@
+// g++ -O3 -DNDEBUG benchmarkX.cpp -o benchmarkX && time ./benchmarkX
+
+#include <Eigen/Core>
+
+using namespace std;
+USING_PART_OF_NAMESPACE_EIGEN
+
+int main(int argc, char *argv[])
+{
+	MatrixXd I = MatrixXd::identity(20,20);
+	MatrixXd m(20,20);
+	for(int i = 0; i < 20; i++) for(int j = 0; j < 20; j++)
+	{
+		m(i,j) = 0.1 * (i+20*j);
+	}
+	for(int a = 0; a < 100000; a++)
+	{
+		m = I + 0.00005 * (m + m*m);
+	}
+	cout << m << endl;
+	return 0;
+}