To generate the LUT (at the moment for R<400, with layers above 270 cm not optimized) run
root -b -q O2/Detectors/Base/test/buildMatBudLUT.C+
The generation is quite time consuming (may take ~30 min).
It can be filled in parallel, one TGeoNavigator per thread, by passing a thread count as the
5th argument of buildMatBudLUT or by setting the environment variable:
export NTHREADS_MATBUD=16
The result does not depend on the number of threads. Scaling beyond a few threads needs ROOT >= v6-36-10-alice3, which removes the per-query thread-id lookup and the false sharing between the per-thread scratch buffers of TGeo shapes; with older ROOT the parallel path is still correct, just slower.
The optimized LUT will be stored in the matbud.root file.
Load it as:
auto mbr = o2::base::MatLayerCylSet::loadFromFile("matbud.root");
To query mat. budget between 2 points use:
float xyz0[3] = {0.,0.,0.};
float xyz1[3] = {70.,80.,20.};
auto mb = mbl.getMatBudget(xyz0[0],xyz0[1],xyz0[2], xyz1[0],xyz1[1],xyz1[2]);
// alternatively, use MatCell getMatBudget(const math_utils::Point3D<float> &point0,const math_utils::Point3D<float> &point1) method
std::cout << "<rho>= " << mb.meanRho << " <x/X0>= " << mb.meanX2X0 << "\n";
Macro extractLUTLayers.C can be used to extract layers covering certain radius range to obtain more compact LUT.