Optimizations #106
Merged
Optimizations #106
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
elalish
reviewed
May 10, 2022
| @@ -50,12 +53,10 @@ void Manifold::Impl::Face2Tri(const VecDH<int>& faceEdge, | |||
| ALWAYS_ASSERT(numEdge >= 3, topologyErr, "face has less than three edges."); | |||
| const glm::vec3 normal = faceNormal[face]; | |||
|
|
|||
| std::map<int, int> vertBary; | |||
| for (int j = firstEdge; j < lastEdge; ++j) | |||
| vertBary[halfedge[j].startVert] = halfedgeBary.H()[j]; | |||
There was a problem hiding this comment.
Good call, I bet this could be a big part of why the triangulation step has been scaling worse than it should.
| // Stable sort is required here so that halfedges from the same face are | ||
| // paired together (the triangles were created in face order). In some | ||
| // degenerate situations the triangulator can add the same internal edge in | ||
| // two different faces, causing this edge to not be 2-manifold. We detect this | ||
| // and fix it by swapping one of the identical edges, so it is important that | ||
| // we have the edges paired according to their face. | ||
| std::stable_sort(edge.begin(), edge.end()); | ||
| thrust::stable_sort(edge.beginD(), edge.endD()); |
There was a problem hiding this comment.
Is this new? I vaguely recall using std:: because thrust didn't have a stable_sort.
There was a problem hiding this comment.
From git blame of https://github.com/NVIDIA/thrust/blame/fa54f2c6f1217237953f27ddf67f901b6b34fbdd/testing/stable_sort.cu, it seems that stable_sort is in thrust for over 10 years. Perhaps it is related to cudatoolkit version?
There was a problem hiding this comment.
cool, I must have remembered wrong.
pca006132
force-pushed
the
optimizations
branch
2 times, most recently
from
ad1e096
to
02ce2b4
Compare
elalish
approved these changes
May 10, 2022
There was a problem hiding this comment.
Really love how much simpler this is! And perfTest for CUDA is indeed running ~30% faster for me:
nTri = 512, time = 0.00647049 sec
nTri = 2048, time = 0.00899614 sec
nTri = 8192, time = 0.0170416 sec
nTri = 32768, time = 0.0450551 sec
nTri = 131072, time = 0.144635 sec
nTri = 524288, time = 0.524747 sec
nTri = 2097152, time = 1.89367 sec
nTri = 8388608, time = 7.68228 sec
| @@ -266,7 +266,7 @@ Collider::Collider(const VecDH<Box>& leafBB, | |||
| */ | |||
| template <typename T> | |||
| SparseIndices Collider::Collisions(const VecDH<T>& querriesIn) const { | |||
| int maxOverlaps = 1 << 20; | |||
| int maxOverlaps = querriesIn.size() * 4; | |||
| // Stable sort is required here so that halfedges from the same face are | ||
| // paired together (the triangles were created in face order). In some | ||
| // degenerate situations the triangulator can add the same internal edge in | ||
| // two different faces, causing this edge to not be 2-manifold. We detect this | ||
| // and fix it by swapping one of the identical edges, so it is important that | ||
| // we have the edges paired according to their face. | ||
| std::stable_sort(edge.begin(), edge.end()); | ||
| thrust::stable_sort(edge.beginD(), edge.endD()); |
There was a problem hiding this comment.
cool, I must have remembered wrong.
elalish
approved these changes
May 11, 2022
| @@ -282,7 +282,11 @@ SparseIndices Collider::Collisions(const VecDH<T>& querriesIn) const { | |||
| break; | |||
| else { // if not enough memory was allocated, guess how much will be needed | |||
| int lastQuery = querryTri.Get(0).H().back(); | |||
| maxOverlaps *= 2; | |||
| float ratio = static_cast<float>(querriesIn.size()) / lastQuery; | |||
| if (ratio > 1000) // do not trust the ratio if it is too large | |||
cartesian-theatrics
pushed a commit
to SovereignShop/manifold
that referenced
this pull request
Mar 11, 2024
Optimizations
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
optimizations as mentioned in #105