Castro

Castro copied to clipboard

(on Cori Haswell nodes, using the 3e8 K setup) The tracer particles implementation with the XRB flame wave simulations fails in the case of 2 refinement levels, ref_ratio of 4...

nmford20

## PR summary ## PR motivation ## PR checklist - [ ] test suite needs to be run on this PR - [ ] this PR will change answers in...

zingale

some normalization efforts for true SDC

4

## PR summary ## PR motivation ## PR checklist - [ ] test suite needs to be run on this PR - [ ] this PR will change answers in...

zingale

revisit CMA

1

Currently we do the "consistent multifluid advection" method by constraining the species fluxes to sum to the density flux after we have done all of the transverse flux differences. But...

zingale

consider changing dtnuc_e to use average de/dt over timestep instead of instantaneous

4

Currently we use the instantaneous energy release from the network in determining the burning limited timestep. This could be problematic if we are in a region where we are trying...

zingale

add a fourth-order burn option

2

Even for the 2nd order algorithm, it can be beneficial to do the reactions 4th order accurate in space. This will correct for the fact that != enuc(, , )....

zingale

For true SDC (`USE_TRUE_SDC = TRUE`), when we finish a step and enter the sum_utils, we get an FPE in the `gwstrain` calculation. This is likely because the true SDC...

zingale

The gravitational acceleration (g, not phi) is most naturally expressed at cell edges rather than cell centers in the Poisson gravity formulation, and it is straightforward to express it at...

maxpkatz

A new Athena++ paper (https://arxiv.org/abs/2012.01340v1) argues that when g has truncation error that is larger than roundoff (which is true for multigrid), we can correct the lack of momentum conservation...

maxpkatz

state_nghost gives different answers

4

Run `reacting_bubble` with `inputs_2d_test` as: ``` ./Castro2d.gnu.MPI.ex inputs_2d_test castro.state_nghost=0 max_step=10 ``` and then ``` ./Castro2d.gnu.MPI.ex inputs_2d_test castro.state_nghost=1 max_step=10 ``` and there are nontrivial differences: ``` variable name absolute error relative...

zingale