Added High Order and Low Order RK methods #2304
Conversation
| alg_order(alg::BS5) = 5 | ||
| alg_order(alg::OwrenZen4) = 4 | ||
| alg_order(alg::OwrenZen5) = 5 | ||
| alg_order(alg::Tsit5) = 5 |
There was a problem hiding this comment.
We will want to make Tsit5 be its own set.
There was a problem hiding this comment.
Ok so for this I will make a separate OrdinaryDiffEqTsit inside lib
There was a problem hiding this comment.
OrdinaryDiffEqTsit5. Since it's the most common method and it's commonly used alone, making it as lean as possible is a virtue. Thanks.
| @@ -818,8 +593,7 @@ for Alg in [:Exp4, :EPIRK4s3A, :EPIRK4s3B, :EPIRK5s3, :EXPRB53s3, :EPIRK5P1, :EP | |||
| iop) | |||
| end | |||
| end | |||
| struct SplitEuler <: | |||
There was a problem hiding this comment.
I went through the code for low order rk and it had some dependency on splitEuler and it also included it's caches. So I moved it there
There was a problem hiding this comment.
It should end up with the split methods group, with CNAB2 and such. What's the dependency?
There was a problem hiding this comment.
when I went through low_order_rk_caches.jl file I saw the splitEuler Cache here so I thought that splitEuler might be an algorithm of low_order_rk. That was the reason
There was a problem hiding this comment.
That was a good guess but ultimately not the right one. There should be a SplitMethods which has CNAB2 and such, and SplitEuler is kind of a smaller test method that goes with that set.
Co-authored-by: Christopher Rackauckas <[email protected]>
|
Can we take this moment to split |
We can start working on splitting this one then |
|
@ChrisRackauckas Why did this happen ?? We didn't get these warnings previously |
|
Fixed that. |
|
Almost there! |
|
@oscardssmith I reverted your Rosenbrock commit because it broke some tests. |
|
@ChrisRackauckas Which solvers should I go with now ?? |
|
IMEXMultistep, Linear, and ExponenitalRK are the last 3 sets. |
Checklist
contributor guidelines, in particular the SciML Style Guide and
COLPRAC.
Additional context
Add any other context about the problem here.
solves a part of #2177