KernelAbstractions 0.10 - #3302
Closed
christiangnrd wants to merge 2 commits into
Closed
christiangnrd wants to merge 2 commits into
christiangnrd wants to merge 2 commits into
Conversation
christiangnrd
added this pull request to stack #3303
September 26, 2026 19:51
maleadt
added a commit
to JuliaGPU/KernelAbstractions.jl
that referenced
this pull request
Sep 27, 2026
The fork branch this used to point to is gone; the KernelAbstractions 0.10 port is now JuliaGPU/CUDA.jl#3302, and the branch this points to builds on it.
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## interface #3302 +/- ##
==============================================
- Coverage 85.77% 38.59% -47.18%
==============================================
Files 188 188
Lines 19130 18660 -470
==============================================
- Hits 16408 7202 -9206
- Misses 2722 11458 +8736 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
This was referenced Sep 27, 2026
maleadt
added a commit
to JuliaGPU/KernelAbstractions.jl
that referenced
this pull request
Sep 27, 2026
The fork branch this used to point to is gone; the KernelAbstractions 0.10 port is now JuliaGPU/CUDA.jl#3302, and the branch this points to builds on it.
maleadt
added a commit
to JuliaGPU/KernelAbstractions.jl
that referenced
this pull request
Sep 27, 2026
The fork branch this used to point to is gone; the KernelAbstractions 0.10 port is now JuliaGPU/CUDA.jl#3302, and the branch this points to builds on it.
maleadt
added a commit
to JuliaGPU/KernelAbstractions.jl
that referenced
this pull request
Sep 27, 2026
The fork branch this used to point to is gone; the KernelAbstractions 0.10 port is now JuliaGPU/CUDA.jl#3302, and the branch this points to builds on it.
maleadt
added a commit
to JuliaGPU/KernelAbstractions.jl
that referenced
this pull request
Sep 27, 2026
The fork branch this used to point to is gone; the KernelAbstractions 0.10 port is now JuliaGPU/CUDA.jl#3302, and the branch this points to builds on it.
maleadt
added a commit
to JuliaGPU/KernelAbstractions.jl
that referenced
this pull request
Sep 27, 2026
The fork branch this used to point to is gone; the KernelAbstractions 0.10 port is now JuliaGPU/CUDA.jl#3302, and the branch this points to builds on it.
Closed
maleadt
added a commit
to JuliaGPU/KernelAbstractions.jl
that referenced
this pull request
Sep 27, 2026
The fork branch this used to point to is gone; the KernelAbstractions 0.10 port is now JuliaGPU/CUDA.jl#3302, and the branch this points to builds on it.
With KernelAbstractions 0.10, back-ends implement KernelInterface, and KernelAbstractions builds on top of that. `CUDABackend` is now the KernelInterface back-end from CUDACore (`CUDAInterface.jl` replaces `CUDAKernels.jl`, and its module is renamed to `CUDAKernels`), and only the KernelAbstractions-specific bits (kernel launch, `@index` bounds checking, scratch memory, `@Const` adaptation) remain, in a `KernelAbstractionsExt` package extension. - require KernelAbstractions 0.10, taken from its main branch for the time being ([sources] in test/Project.toml). The Julia 1.10 and 1.11 CI jobs develop it from git, as 1.10 ignores [sources] and 1.11 doesn't see the test project's when instantiating the workspace, and the GPU-less job moves to Julia 1.12 for the same reason; - cuSPARSE extends `KernelInterface.get_backend`, which is the same function as `KernelAbstractions.get_backend` in KA 0.10, so it no longer depends on KernelAbstractions; - `CUDA.versioninfo` reports KernelAbstractions only when it's loaded; - skip the KA tests that launch kernels on the POCL-based CPU back-end. Co-authored-by: Tim Besard <tim.besard@gmail.com>
Dogfood the KernelInterface device API (`get_local_id`, `get_group_id`, `get_local_size`, `get_num_groups`, `barrier`, `localmemory`) in the mapreduce, scan, `findall` and RNG kernels. The indices are requested as `Int32`, matching CUDA's hardware indices: KernelInterface defaults to `Int`, which turns the index arithmetic into 64-bit operations and made reductions 2x, scans and `findall` 4-7x, and in-kernel `rand` 15x slower. Co-authored-by: Tim Besard <tim.besard@gmail.com>
Member
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.
#2944 but on this repo so it can stack