Skip to content

[Metal] Simdgroup intrinsic attributes and tweak target triple #744

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
christiangnrd wants to merge 1 commit into
base: master
Choose a base branch
from
Open

[Metal] Simdgroup intrinsic attributes and tweak target triple #744

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 10 additions & 2 deletions src/metal.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@ source_code(target::MetalCompilerTarget) = "text"
# Metal is not supported by our LLVM builds, so we can't get a target machine
llvm_machine(::MetalCompilerTarget) = nothing

llvm_triple(target::MetalCompilerTarget) = "air64-apple-macosx$(target.macos)"
llvm_triple(target::MetalCompilerTarget) = "air64_v$(target.air.major)$(target.air.minor)-apple-macosx$(target.macos)"
Copy link
Member

@maleadt maleadt Nov 24, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this what Apple does in AIR files?

Copy link
Member Author

@christiangnrd christiangnrd Nov 24, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The air version gets appended to “air64” when -mmacosx-version-min is specified, or when a newer feature (like bfloat) is used. I can revert this and wait until we gave a feature that actually requires a newer air version is in (bfloat or logging) so that we can test if this is actually needed or not.


llvm_datalayout(target::MetalCompilerTarget) =
"e-p:64:64:64"*
Expand Down Expand Up @@ -1018,7 +1018,7 @@ function annotate_air_intrinsics!(@nospecialize(job::CompilerJob), mod::LLVM.Mod

# synchronization
if fn == "air.wg.barrier" || fn == "air.simdgroup.barrier"
add_attributes("nounwind", "convergent")
add_attributes("nounwind", "mustprogress", "convergent", "willreturn")

# atomics
elseif match(r"air.atomic.(local|global).load", fn) !== nothing
Expand All @@ -1033,6 +1033,14 @@ function annotate_air_intrinsics!(@nospecialize(job::CompilerJob), mod::LLVM.Mod
elseif match(r"^air.atomic.(local|global).(add|sub|min|max|and|or|xor)", fn) !== nothing
# TODO: "memory(argmem: readwrite)" on LLVM 16+
add_attributes("argmemonly", "nounwind")

# simdgroup
elseif match(r"air.simdgroup_matrix_8x8_multiply_accumulate", fn) !== nothing
add_attributes("convergent", "mustprogress", "nounwind", "willreturn")
elseif match(r"air.simdgroup_matrix_8x8_load", fn) !== nothing
add_attributes("convergent", "mustprogress", "nofree", "nounwind", "readonly", "willreturn")
elseif match(r"air.simdgroup_matrix_8x8_store", fn) !== nothing
add_attributes("convergent", "mustprogress", "nounwind", "willreturn", "writeonly")
end
end

Expand Down
Loading