Speeding up triangulation about truck

[ytanimura]

Now, MeshableShape::triangulation has been parallelized!

There is a benchmark that compare new algorithm to old one.

[ytanimura]

Thank you for your issue. Actually we are thinking the same thing and have completed the technical verification. We have not yet been able to merge it into the master branch because of the following minor issues that remain.

• The struct ID, which widely used by topological algorithms, is not implemented Send and Sync because the raw pointer is used as the actual state.
• One of the goals of truck is to run in wasm. We recognize that parallel processing on a browser has a large overlap; in the case of wasm, it is necessary to apply a branching process that does not parallelize the process.

In fact, it won't take much time. Now that we have received the issue, we will try to merge it into master within the next week or a little later.