Add isotropic thermal conduction and RKL2 supertimestepping #1
Conversation
…or using prolong/restric with non-base containers)
|
It looks like this also adds a new reconstruction method. Would it make sense to split that off into its own PR? or is that needed for the test problems here? |
|
That recon is already in |
pgrete
changed the title
|
@forrestglines @BenWibking I just fixed the tests and updated to latest |
| pkg->AddParam<>("thermal_diff", thermal_diff); | ||
|
|
||
| const auto mu = pkg->Param<Real>("mu"); | ||
| conduction_sat_prefac = 6.86 * std::sqrt(mu) * conduction_sat_phi; |
There was a problem hiding this comment.
Is this 6.86 a magic number?
There was a problem hiding this comment.
Yes, it assumes a fully ionized hydrogen plasma. I added an appropriate comment to clarify.
| if (parthenon::Globals::my_rank == 0) { | ||
| const auto ratio = 2.0 * tau / mindt_diff; | ||
| std::cout << "STS ratio: " << ratio << " Taking " << s_rkl << " steps." << std::endl; | ||
| if (ratio > 400.1) { | ||
| std::cout << "WARNING: ratio is > 400. Proceed at own risk." << std::endl; | ||
| } | ||
| } |
There was a problem hiding this comment.
Do we always want to output this? Writing to STDOUT every cycle? Could this cause a slow down on some machines?
There was a problem hiding this comment.
I found this useful when running simulations (as it explains/is related to the performance per cycle).
Given that std:cout is buffered I also don't expect any significant slowdown.
| return TaskStatus::complete; | ||
| } | ||
|
|
||
| // Assumes that prim and cons are in sync initially. |
There was a problem hiding this comment.
Is this assumption affected by other user/pgen source terms that might be added?
There was a problem hiding this comment.
Currently not as it's both the very first task called before any other sources and the very last task called after FillDerived.
I added notes that if the position of the tasks changes, this should be updated.
| @@ -2,5 +2,5 @@ | |||
|
|
|||
| Executes 2D ring diffusion problem following Sharma & Hammett (2007) and calculates convergence rate. | |||
| Errors are calculated based on comparison to steady state solution. | |||
| Convergence for this problem is not great, but matches other numbers reported in literate, e.g., Balsara, Tilley & Howk MNRAS (2008) doi:10.1111/j.1365-2966.2008.13085.x . | |||
| Convergence for this problem is not great, but matches other numbers reported in literature, e.g., Balsara, Tilley & Howk MNRAS (2008) doi:10.1111/j.1365-2966.2008.13085.x . | |||
There was a problem hiding this comment.
Do you have an image of the problem to share? Any obvious issues with AthenaPK's solution?
There was a problem hiding this comment.
The resulting image from that test can directly be compared to the number in literature ( I extracted the numbers), i.e.
New features
Open items