OpenMP in `addParticles(other)`

[ax3l]

I am running workflows where I often store a reference particle container and copy it over with addParticles(other) for (many) repeated benchmarks (SciDAC) and interactive workflows (LDRD).

I noticed that addParticles(other) does not scale with OpenMP threads.

@atmyers @WeiqunZhang can we easily add OpenMP parallelism here? Ideally, we should try to speed up even tiled, single-box simulations.
https://github.com/AMReX-Codes/amrex/blob/25.08/Src/Particle/AMReX_ParticleContainerI.H#L1171-L1188

[WeiqunZhang]

This might be all we need.

diff --git a/Src/Particle/AMReX_ParticleContainerI.H b/Src/Particle/AMReX_ParticleContainerI.H
index 0f908ad8f2..bce5e7b265 100644
--- a/Src/Particle/AMReX_ParticleContainerI.H
+++ b/Src/Particle/AMReX_ParticleContainerI.H
@@ -1166,6 +1166,9 @@ addParticles (const PCType& other, F const& f, bool local)
 {
     BL_PROFILE("ParticleContainer::addParticles");
 
+#if defined(AMREX_USE_OMP) && !defined(AMREX_USE_GPU)
+#pragma omp parallel
+#endif
     for (int lev = 0; lev < other.numLevels(); ++lev)
     {
         const auto& plevel_other = other.GetParticles(lev);

[atmyers]

Could you try this: #4615

[ax3l]

Fixed via #4615