Sync with fe86ed4 #48
Conversation
qnzhou
force-pushed
the
qnzhou/v6.29.0
branch
from
8302679 to
cb9718a
Compare
| @@ -108,12 +108,12 @@ std::vector<Index> split_edges( | |||
| std::vector<std::array<Index, 2>> parent_edge( | |||
| num_output_vertices, | |||
| {invalid<Index>(), invalid<Index>()}); | |||
| tbb::parallel_for(Index(0), num_edges, [&](Index eid) { | |||
| for (Index eid = 0; eid < num_edges; eid++) { | |||
There was a problem hiding this comment.
I don't understand why this change is necessary?
There was a problem hiding this comment.
Thread sanitizer is complaining about race condition. But yes, I think it is a false alarm.
There was a problem hiding this comment.
It's actually not a false alarm. Looking more closely at the code you are iterating over each edges, writing the vertex->edge relationship in the parent_edge array. When multiple edges are incident to the same vertex this parent will be overwritten multiple time, with different edge indices each time.
There was a problem hiding this comment.
Actually, I think you are right. For split_long_edges, every edge split point has a unique parent edge, so there is no race condition. However, for remove_degenerate_facets, an edge split points may belong to multiple collinear edges, which leads to the race condition. Let me fix that.
No description provided.