Distance tracking on stack causes OpenMP traversal to fail
Nobody has claimed this yet.
- Dominant language
- C++
- Stars
- 89
- Forks
- 24
- PR merge metrics
- No merged PRs in 30d
Description
PR #28 added distance tracking on stack to the ray traversal kernels. When trying to use cuBQL in our own downstream app (XDG) which is making use of the OpenMP BVH builder and ShrinkingRayTraversal template the changes made in #28 cause rays that would previously return correct hits to now return misses.
Disabling the distance tracking by setting this line to #define CUBQL_DIST_STACK 0 resolves the issue and i get the expected successful ray hits again. I'm not sure about how to fix it for OpenMP traversal as clearly there is no issue when using the CUDA backend. But in the absence of an actual fix, perhaps the ability to toggle on or off CUBQL_DIST_STACK would be helpful to ensure that OpenMP runtimes can still work?
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start in cuBQL/traversal/rayQueries.h at the CUBQL_DIST_STACK definition, then inspect the distance-tracking changes from PR #28 and the ShrinkingRayTraversal path used by the OpenMP BVH builder. Reproduce the downstream ray-hit failure with distance tracking enabled and compare it with CUBQL_DIST_STACK set to 0; done means OpenMP traversal returns the expected hits without breaking the CUDA path.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- computer-graphics
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 48/100