feat(cuda): add C-level simple API mirroring CPU one-shot wrappers#863
Open
DiamonDinoia wants to merge 1 commit into
Open
feat(cuda): add C-level simple API mirroring CPU one-shot wrappers#863DiamonDinoia wants to merge 1 commit into
DiamonDinoia wants to merge 1 commit into
Conversation
Adds 36 entry points (3 dims x 3 types x {single,many} x {single,double})
that wrap the cuFINUFFT plan -> setpts -> execute -> destroy sequence
into a single call. Pointers are device pointers; argument order matches
the CPU finufft1d1 family in src/c_interface.cpp. The plan is held in a
unique_ptr so error paths still release GPU resources.
A cross-check test exercises every (dim, type, precision) combination
against the 4-step plan path, and the C ABI is smoke-tested from
test/cuda/public_api_test.c. The new surface is documented in
docs/c_gpu.rst.
Reference: src/c_interface.cpp (CPU finufft1d1 family) and the prior
cufinufft simple-interface draft by @janden on janden/update_cu_api
(commit 61e2ae9 "cuda: first attempt at simple cufinufft interface").
Co-authored-by: Joakim Andén <3976052+janden@users.noreply.github.com>
|
FFT backend: Numbers are advisory: GitHub-hosted runners have variable hardware. Treat <1.10× as noise. CPU and compiler configurationCPU name: Arch: Core count: ISA extensions: Compiler version: Compiler flags: perftest commands |
Collaborator
|
Good idea! |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 36 entry points (3 dims x 3 types x {single,many} x {single,double}) that wrap the cuFINUFFT plan -> setpts -> execute -> destroy sequence into a single call. Pointers are device pointers; argument order matches the CPU finufft1d1 family in src/c_interface.cpp. The plan is held in a unique_ptr so error paths still release GPU resources.
A cross-check test exercises every (dim, type, precision) combination against the 4-step plan path, and the C ABI is smoke-tested from test/cuda/public_api_test.c. The new surface is documented in docs/c_gpu.rst.
Reference: the prior cufinufft simple-interface draft by @janden on janden/update_cu_api