Struct KernelProfile 

pub struct KernelProfile {
    pub tile_m: u32,
    pub tile_n: u32,
    pub tile_k: u32,
    pub unroll_factor: u32,
    pub use_shared_memory: bool,
    pub register_target: u32,
    pub split_k: u32,
}

Mutable kernel configuration that the profile-guided optimizer adjusts.

Downstream builders read these fields after optimisation to generate the final PTX code.

Fields

tile_m: u32

Tile size in the M dimension.

tile_n: u32

Tile size in the N dimension.

tile_k: u32

Tile size in the K dimension.

unroll_factor: u32

Loop unroll factor.

use_shared_memory: bool

Whether shared memory staging is enabled.

register_target: u32

Target register count per thread (0 = no constraint).

split_k: u32

Number of split-K slices (1 = disabled).

Implementations

impl KernelProfile

pub const fn new() -> Self

Creates a new KernelProfile with sensible defaults.

Trait Implementations

impl Clone for KernelProfile

fn clone(&self) -> KernelProfile

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for KernelProfile

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for KernelProfile

fn default() -> Self

Returns the “default value” for a type.

impl Display for KernelProfile

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.