Level 1 vs Level 2 s-function execute time

I have the same algorithm implemented as both a level-1 and level-2 S-function. The Level-2 version gets its core functions called several hundred more times and as such, the simulation runs several hundred times slower (eg the L1 output function is called ~6000 times in one simulation, the L2 one is called ~450,000). I have seen others with the same problem, but no answers have been forthcoming. Any help would be appreciated.

Regards, --Ben.

Answer 1

[Ben]

Solved - for others having this problem, I had an input port direct feedthrough incorrectly set to true. This caused the creation of an agebraic loop.

[Kaustubha Govind]

Thanks for posting your answer, Ben!