pub struct Nvfp4MatmulSpec<'a> {
pub activation: Operand,
pub packed: &'a [Operand],
pub block_scales: &'a [Operand],
pub tensor_scale: Operand,
pub output: Operand,
pub m: u32,
pub n: u32,
pub k: u32,
pub epilogue: u32,
pub weight_mode: u32,
}Expand description
Specialization payload for native row-major NVFP4: activation, packed
weights, block scales, tensor scale, output, then m, n, k, epilogue.
Fields§
§activation: Operand§packed: &'a [Operand]§block_scales: &'a [Operand]§tensor_scale: Operand§output: Operand§m: u32§n: u32§k: u32§epilogue: u32§weight_mode: u32Implementations§
Auto Trait Implementations§
impl<'a> Freeze for Nvfp4MatmulSpec<'a>
impl<'a> RefUnwindSafe for Nvfp4MatmulSpec<'a>
impl<'a> Send for Nvfp4MatmulSpec<'a>
impl<'a> Sync for Nvfp4MatmulSpec<'a>
impl<'a> Unpin for Nvfp4MatmulSpec<'a>
impl<'a> UnsafeUnpin for Nvfp4MatmulSpec<'a>
impl<'a> UnwindSafe for Nvfp4MatmulSpec<'a>
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