CPU support #45
Merged
CPU support #45
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
|
This feature could become very useful in near future. 100x slower is actually not too slow. With an extra CPU based optimizations like Intel AVX-512 and AMX, I think it should be feasible to give another 10x performance boost. |
|
Agree! I was actually quite pleased with the performance (and it's a first implementation, so lots of room for improvements). |
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Outdated suggestions cannot be applied.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.
Adds CPU support, which allows to generate splats without any GPUs (if you have time to wait).
Currently it runs about ~100x slower than CUDA, but there's probably room for improvement in performance. In particular, there might be ways to parallelize some calculations in the renderer, as well as evaluating whether the project_forward and compute_sh_forward functions could benefit from a manual C++ implementation (without relying on libtorch, which conveniently gives us the backward passes for free).
The results should be numerically equivalent to the CUDA implementation (although some numerical precision differences will be present).
Closes #5
Try it by passing the
--cpuflag to./opensplat