Skip to content

amdgpu: document that ds_permute/ds_bpermute take a byte address #2258

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
adamgemmell merged 1 commit into from
Oct 27, 2026
Merged

amdgpu: document that ds_permute/ds_bpermute take a byte address #2258

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
10 changes: 7 additions & 3 deletions crates/core_arch/src/amdgpu/mod.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -702,10 +702,12 @@ pub fn s_memrealtime() -> u64 {

/// Scatter data across all lanes in a wavefront.
///
/// Writes `value` to the lane `lane`.
/// Writes `value` to the lane addressed by `lane`.
///
/// `lane` is a byte address, not a lane index: pass `n * 4` to write to lane `n`.
///
/// Reading from inactive lanes returns `0`.
/// In case multiple values get written to the same `lane`, the value from the source lane with the higher index is taken.
/// In case multiple values get written to the same lane, the value from the source lane with the higher index is taken.
///
#[doc = include_str!("intrinsic_is_convergent.md")]
#[inline]
Expand All @@ -715,7 +717,9 @@ pub unsafe fn ds_permute(lane: u32, value: u32) -> u32 {
}
/// Gather data across all lanes in a wavefront.
///
/// Returns the `value` given to `ds_permute` by lane `lane`.
/// Returns the `value` given to `ds_bpermute` by the lane addressed by `lane`.
///
/// `lane` is a byte address, not a lane index: pass `n * 4` to read from lane `n`.
///
/// Reading from inactive lanes returns `0`.
///
Expand Down
Loading