Support for k-mers up to 128.
Currently it is about 4 times slower for k-mers >64 than for k-mers up to 64.
My guess that the additional 2 times slowdown (2 times slower is required as software implementation of the type) comes from additional operations needed to compute hashes etc.
I.g. this could be optimized but i don't think it's a priority. Even the fact that those k are supported and they are not terribly slow should be an advantage.
Currently the only downside to this PR is rouhgly 2 times slowdown of compilation. I'll see what can be done with it.
EDIT: the compilation now takes about 3sec on my PC which i.g. should be okay.
Support for k-mers up to 128. Currently it is about 4 times slower for k-mers >64 than for k-mers up to 64. My guess that the additional 2 times slowdown (2 times slower is required as software implementation of the type) comes from additional operations needed to compute hashes etc.
I.g. this could be optimized but i don't think it's a priority. Even the fact that those k are supported and they are not terribly slow should be an advantage.
Currently the only downside to this PR is rouhgly 2 times slowdown of compilation. I'll see what can be done with it. EDIT: the compilation now takes about 3sec on my PC which i.g. should be okay.