vo002_VectorCalculations.C File Reference

Detailed Description

In this tutorial we learn how the RVec class can be used to express easily mathematical operations involving arrays and scalars.

 
void vo002_VectorCalculations()
{
 
   // Operations on RVec instances are made to be *fast* (vectorisation is exploited)
   // and easy to use. RVecF, RVecD, RVecI, ..., are handy aliases for RVec<float>, RVec<double>, RVec<int>, ...
   ROOT::RVecF v1{1., 2., 3.};
   ROOT::RVecF v2{4., 5., 6.};
 
   // Arithmetic operations are to be intended on pairs of elements with identical index
   auto v_sum = v1 + v2;
   auto v_mul = v1 * v2;
 
   // Easy to inspect:
   std::cout << "v1 = " << v1 << "\n"
             << "v2 = " << v2 << "\n"
             << "v1 + v2 = " << v_sum << "\n"
             << "v1 * v2 = " << v_mul << std::endl;
 
   // It's also possible to mix scalars and RVecs
   auto v_diff_s_0 = v1 - 2;
   auto v_diff_s_1 = 2 - v1;
   auto v_div_s_0 = v1 / 2.;
   auto v_div_s_1 = 2. / v1;
 
   std::cout << v1 << " - 2 = " << v_diff_s_0 << "\n"
             << "2 - " << v1 << " = " << v_diff_s_1 << "\n"
             << v1 << " / 2 = " << v_div_s_0 << "\n"
             << "2 / " << v1 << " = " << v_div_s_1 << std::endl;
 
   // Dot product and the extraction of quantities such as Mean, Min and Max
   // are also easy to express (see here for the full list:
   // https://root.cern/doc/master/namespaceROOT_1_1VecOps.html)
   auto v1_mean = Mean(v1);
   auto v1_dot_v2 = Dot(v1, v2);
 
   std::cout << "Mean of " << v1 << " is " << v1_mean << "\n"
             << "Dot product of " << v1 << " and " << v2 << " is " << v1_dot_v2 << std::endl;
 
   // Most used mathematical functions are supported
   auto v_exp = exp(v1);
   auto v_log = log(v1);
   auto v_sin = sin(v1);
 
   std::cout << "exp(" << v1 << ") = " << v_exp << "\n"
             << "log(" << v1 << ") = " << v_log << "\n"
             << "sin(" << v1 << ") = " << v_sin << std::endl;
 
   // Even an optimised version of the functions is available
   // provided that VDT is not disabled during the configuration
#ifdef R__HAS_VDT
   auto v_fast_exp = fast_exp(v1);
   auto v_fast_log = fast_log(v1);
   auto v_fast_sin = fast_sin(v1);
 
   std::cout << "fast_exp(" << v1 << ") = " << v_fast_exp << "\n"
             << "fast_log(" << v1 << ") = " << v_fast_log << "\n"
             << "fast_sin(" << v1 << ") = " << v_fast_sin << std::endl;
 
   // It may happen that a custom operation needs to be applied to the RVec.
   // In this case, the Map utility can be used:
   auto v_transf = Map(v1, [](double x) { return x * 2 / 3; });
 
   std::cout << "Applying [](double x){return x * 2 / 3;} to " << v1 << " leads to " << v_transf << "\n";
#endif
}

 
v1 = { 1, 2, 3 }
v2 = { 4, 5, 6 }
v1 + v2 = { 5, 7, 9 }
v1 * v2 = { 4, 10, 18 }
{ 1, 2, 3 } - 2 = { -1, 0, 1 }
2 - { 1, 2, 3 } = { 1, 0, -1 }
{ 1, 2, 3 } / 2 = { 0.5, 1, 1.5 }
2 / { 1, 2, 3 } = { 2, 1, 0.666667 }
Mean of { 1, 2, 3 } is 2
Dot product of { 1, 2, 3 } and { 4, 5, 6 } is 32
exp({ 1, 2, 3 }) = { 2.71828, 7.38906, 20.0855 }
log({ 1, 2, 3 }) = { 0, 0.693147, 1.09861 }
sin({ 1, 2, 3 }) = { 0.841471, 0.909297, 0.14112 }
fast_exp({ 1, 2, 3 }) = { 2.71828, 7.38906, 20.0855 }
fast_log({ 1, 2, 3 }) = { 0, 0.693147, 1.09861 }
fast_sin({ 1, 2, 3 }) = { 0.841471, 0.909297, 0.14112 }
Applying [](double x){return x * 2 / 3;} to { 1, 2, 3 } leads to { 0.666667, 1.33333, 2 }

Date

May 2018

Author

Danilo Piparo

Definition in file vo002_VectorCalculations.C.