#[repr(C)]pub struct GdnSeqVl {}Expand description
task #18 varlen GDN: per-seq args for gdn_chunk_{state,output}_mma_vl — one launch
runs all B<=8 sequences’ K4/K5 (CUDA gdnseq_t/gdnvl_t, layout-identical repr(C)).
Raw addresses are valid for the launch: every referenced buffer outlives the call and
all work is on the single compute stream (same discipline as the f16 GEMM FFI).
Fields§
§kb16: u64§gcum: u64§beta: u64§u: u64§wb16: u64§y: u64§ssnap: u64§state_in: u64§state_out: u64§q: u64§p: u64§o: u64§k: u64§v: u64§g: u64§a: u64§w: u64§t: i32§nc: i32Trait Implementations§
impl Copy for GdnSeqVl
impl DeviceRepr for GdnSeqVl
Auto Trait Implementations§
impl Freeze for GdnSeqVl
impl RefUnwindSafe for GdnSeqVl
impl Send for GdnSeqVl
impl Sync for GdnSeqVl
impl Unpin for GdnSeqVl
impl UnsafeUnpin for GdnSeqVl
impl UnwindSafe for GdnSeqVl
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more