Closed asmeurer closed 4 years ago
Looks good
Another benchmark to include, but is currently way too slow, is
risch_integrate(Add(*[log(x + i) for i in range(10)]), x)
I think it's partly slow because of the polys (it operates on a Poly with 10 generators), but mostly because the height of the tower and certain ways the algorithm works (every log term is a separate extension in the tower).
There should be better benchmarks for the other algorithms too, but I didn't attempt to do that here. I also didn't attempt to refactor the other integration benchmarks code.
One question is if these benchmarks are too slow. If they are, we can reduce the
100
parameter.