Add atomic_rmw! with syncscope
#431
Merged
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
pxl-th
force-pushed
the
pxl-th/atomic-syncscope
branch
from
76ac5f6
to
597fb37
Compare
You need to manually re-generate the wrappers by running |
|
Done! |
pxl-th
force-pushed
the
pxl-th/atomic-syncscope
branch
from
bb6907f
to
192e8a6
Compare
|
@maleadt, kind ping for review. |
maleadt
approved these changes
Jul 27, 2024
|
@maleadt, I've updated tests for LLVM < 16. Hopefully now they will pass :) |
maleadt
reviewed
Jul 27, 2024
|
LGTM. CI failure is EnzymeAD/Enzyme.jl#1688 |
|
Thanks! Can we tag a new release so that further packages can use it? |
|
First needs a JLL bump. I'll create a PR. |
nikic
pushed a commit
to llvm/llvm-project
that referenced
this pull request
Aug 20, 2024
…104775) Another upstreaming of C API extensions we have in Julia/LLVM.jl. Although [we went](maleadt/LLVM.jl#431) with a string-based API there, here I'm proposing something that's similar to existing metadata/attribute APIs: - explicit functions to map syncscope names to IDs, and back - `LLVM*SyncScope` versions of builder APIs that already take a `SingleThread` argument: atomic rmw, atomic xchg, fence - `LLVMGetAtomicSyncScopeID` and `LLVMSetAtomicSyncScopeID` for other atomic instructions - testing through `llvm-c-test`'s `--echo` functionality
dmpolukhin
pushed a commit
to dmpolukhin/llvm-project
that referenced
this pull request
Sep 2, 2024
…lvm#104775) Another upstreaming of C API extensions we have in Julia/LLVM.jl. Although [we went](maleadt/LLVM.jl#431) with a string-based API there, here I'm proposing something that's similar to existing metadata/attribute APIs: - explicit functions to map syncscope names to IDs, and back - `LLVM*SyncScope` versions of builder APIs that already take a `SingleThread` argument: atomic rmw, atomic xchg, fence - `LLVMGetAtomicSyncScopeID` and `LLVMSetAtomicSyncScopeID` for other atomic instructions - testing through `llvm-c-test`'s `--echo` functionality
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.
Add ability to specify syncscope for
atomic_rmw!op.This is needed to properly enable hardware floating-point atomics for AMDGPU (Ref: JuliaGPU/AMDGPU.jl#569) to avoid CAS loop.
Basically, expose
syncscopeargument here:https://github.com/llvm/llvm-project/blob/8ebe499e07760b0d18b5721b298efc9e4a241916/llvm/lib/IR/Core.cpp#L4299-L4308
Before:
%15 = atomicrmw fadd float addrspace(1)* %14, float 1.000000e+00 seq_cst, align 4, !dbg !148global_atomic_cmpswap_b32 v2, v[0:1], v[2:3], off offset:-4 glcWith syncscope
"agent":%15 = atomicrmw fadd float addrspace(1)* %14, float 1.000000e+00 syncscope("agent") seq_cst, align 4, !dbg !148global_atomic_add_f32 v[0:1], v2, off offset:-4@maleadt, I suppose wrappers in
lib/will be auto-updated or do I need to do this manually? I did it for a local build.