David Widmann
David Widmann
@gdalle Tests still fail due to ADTypes it seems.
Fair enough. It just felt to me like the README > So this package ends up about 30% or so faster than Distributions.jl (the method at the far edge is...
To get up to the same speed, you have to use the samplers in Distributions and should not run `rand(Poisson(x))`. As the benchmark in the first comment above shows, `rand(rng,...
> I also understand the documentation such that dt sets an upper bound to the step size irrespective of additional points set in tstops or the points set in saveat....
> `expected = solve( dde,SROCK2(),dt=dt,saveat=save_times)` `DDEProblem`s require a `MethodOfSteps` algorithm. I'm not sure why this does not throw an error.
I'm not too surprised that it is quite a bit slower since the number of delays scales quadratically with `n`. Since you specify `constant_lags` (BTW I've never specified them as...
> That reminds me though, @YingboMa was mentioning that in some of Enright or Harier's work that discontinuities were not check for if steps were accepted, which is something we...
> Required by JuliaMath/SpecialFunctions.jl/issues/371 It seems it's only used in the tests but not required by the implementation in the PR?
I am not hesitant per se, I just want to understand the motivation for this PR so I can review it properly. You provided one argument for adding it but...
I'm in favour of adding such improvements in principle. However, I think we should be consistent with Julia base and non-breaking, ie all shortcuts should return values of the same...