Hugo Granström

Nice! Do you see any difference in performance after the changes? 😄 (I'm not currently at a computer)

Is Vern7 consistently slower than Vern6 or is it only for bigger tol? It does sound suspicious otherwise. 🤔 I'll do some testing as well tomorrow. Nothing is like finding...

I can confirm that Vern7 is really slow.... Took it 16 second while Vern6 & Tsit5 took less than my system could measure 😅 so there is probably something wrong...

Yeah we should change it back, the reason we changed it last time wasn't even a valid reason 😅 it was just a misunderstanding in the Tsit5. If it gets...

I did the same kind of test as I did before measuring how many steps each method took and what error it thought it had (`error` is what the method...

Yes I can confirm that as well 😄👍

It is really odd that it doesn't work though

If you test to return `yLow` instead of `yHigh`, do you get acceptable errors if you force the minDt pretty high (1e-1) if you compare it to now when we...

That's really odd as well... 😵 The time shouldn't change between the to as we have already done all the time step adjustments when we return `yLow`. But that the...

Tomorrow I will check the coefficients for the y's against JuliaDiffEq to see if I can see any difference