DistributedClosestPoint GPU-related optimizations - #1974

Open

publixsubfan wants to merge 12 commits into

feature/yang39/dcp-gpu-opt

Commits on Sep 1, 2026

Sort indices by morton index to try and speed up closest point kernels
publixsubfan
committed
340fffa
Add a BVH2Node struct for improved HIP performance
publixsubfan
committed
12b7823
BVH: add support for shared memory-based stack
publixsubfan
committed
db86043
Add block size parameter to axom::execution_space
publixsubfan
committed
8426dc3
BVH: add shared memory traversal interface
publixsubfan
committed
b180bf6
GPU optimizations for squared_distance(Point, BoundingBox)
publixsubfan
committed
aa6d92a
BVH: use device-side traverser interface in findCandidatesImpl
publixsubfan
committed
90cb8c4
BVH: fix child slot index layouts used by GWN reduction
publixsubfan
committed
df7d910
Fixup: guard SharedBVHStack on non-GPU platforms
publixsubfan
committed
d89a4be
Document internal BVH stack classes
publixsubfan
committed
a561e5c

Commits on Sep 2, 2026

Merge branch 'develop' into feature/yang39/dcp-gpu-opt
rhornung67
authored
e9339c8
Silence build warning
publixsubfan
committed
bf4072b