Struct PtxMicrokernelSpec 

pub struct PtxMicrokernelSpec {
    pub ptx_version: &'static str,
    pub sm_target: &'static str,
    pub registers_per_thread: u32,
    pub smem_bytes: usize,
    pub block_dim: (u32, u32, u32),
    pub tile_dim: (usize, usize),
}

PTX microkernel definition (for documentation and future CUDA support)

This is a specification for the GPU microkernel. Actual PTX code generation would be done by the trueno-ptx crate.

References

• NVIDIA PTX ISA Reference Manual
• Volkov, V. (2010). Better Performance at Lower Occupancy.

Fields

ptx_version: &'static str

PTX version (e.g., “8.0”)

sm_target: &'static str

Target SM architecture (e.g., “sm_80”)

registers_per_thread: u32

Register count per thread

smem_bytes: usize

Shared memory bytes per block

block_dim: (u32, u32, u32)

Thread block dimensions

tile_dim: (usize, usize)

Tile dimensions (MR, NR)

Trait Implementations

impl Clone for PtxMicrokernelSpec

fn clone(&self) -> PtxMicrokernelSpec

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for PtxMicrokernelSpec

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

Formats the value using the given formatter.

impl Default for PtxMicrokernelSpec

fn default() -> Self

Returns the “default value” for a type.