Skip to content

[MLIR][NVVM] [NFC] Update Docs for shfl.sync Op #89044

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

Merged
merged 1 commit into from
Apr 18, 2024
Merged

[MLIR][NVVM] [NFC] Update Docs for shfl.sync Op #89044

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
19 changes: 16 additions & 3 deletions mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -619,20 +619,33 @@ def ShflKindAttr : EnumAttr<NVVM_Dialect, ShflKind, "shfl_kind">;
def NVVM_ShflOp :
NVVM_Op<"shfl.sync">,
Results<(outs LLVM_Type:$res)>,
Arguments<(ins I32:$dst,
Arguments<(ins I32:$thread_mask,
LLVM_Type:$val,
I32:$offset,
I32:$mask_and_clamp,
ShflKindAttr:$kind,
OptionalAttr<UnitAttr>:$return_value_and_is_valid)> {
let summary = "NVVM Dialect Op for shfl.sync";
let description = [{
The `shfl.sync` Op implements data shuffle within threads of a warp.
The `thread_mask` denotes the threads participating in the Op where
the bit position corresponds to a particular thread’s laneid.
The `offset` specifies a source lane or source lane offset
(depending on `kind`). The `val` is the input value to be copied from
the source. The `mask_and_clamp` contains two packed values specifying
a mask for logically splitting warps into sub-segments and an upper bound
for clamping the source lane index.
[For more information, refer PTX ISA]
(https://docs.nvidia.com/cuda/parallel-thread-execution/#data-movement-and-conversion-instructions-shfl-sync)
}];
string llvmBuilder = [{
auto intId = getShflIntrinsicId(
$_resultType, $kind, static_cast<bool>($return_value_and_is_valid));
$res = createIntrinsicCall(builder,
intId, {$dst, $val, $offset, $mask_and_clamp});
intId, {$thread_mask, $val, $offset, $mask_and_clamp});
}];
let assemblyFormat = [{
$kind $dst `,` $val `,` $offset `,` $mask_and_clamp attr-dict
$kind $thread_mask `,` $val `,` $offset `,` $mask_and_clamp attr-dict
`:` type($val) `->` type($res)
}];
let hasVerifier = 1;
Expand Down