Adding parallel for loops for A matrix compute

HPImaging / sv-mbirct

High Performance Model Based Image Reconstruction for Computed Tomography

BSD 3-Clause "New" or "Revised" License

6 stars 7 forks source link

Adding parallel for loops for A matrix compute #13

Closed svvenkatakrishnan closed 3 years ago

svvenkatakrishnan commented 3 years ago

Given its takes time to compute the A-matrix especially for large volumes, would it be possible to make the core for loop a parallel for ?

sjkisner commented 3 years ago

A_comp() and A_piecewise() were parallelized with omp. These are very memory intensive with a lot of malloc calls. A_comp() works good on a lot of threads. A_piecewise() is set to a maximum of 4 threads..beyond that the run time can increase.