Struct DynamicParallelismConfig 

pub struct DynamicParallelismConfig {
    pub max_nesting_depth: u32,
    pub max_pending_launches: u32,
    pub sync_depth: u32,
    pub child_grid: Dim3,
    pub child_block: Dim3,
    pub child_shared_mem: u32,
    pub sm_version: SmVersion,
}

Configuration for dynamic parallelism (device-side kernel launches).

Controls nesting depth, pending launch limits, synchronization behavior, and child kernel launch dimensions.

CUDA constraints

• max_nesting_depth must be in 1..=24.
• max_pending_launches must be at least 1.
• sync_depth must be less than or equal to max_nesting_depth.
• All grid and block dimensions must be non-zero.

Fields

max_nesting_depth: u32

Maximum nesting depth for child kernel launches (CUDA limit: 24).

max_pending_launches: u32

Maximum number of pending (un-synchronized) child launches (default 2048).

sync_depth: u32

Depth at which to insert synchronization barriers.

Child kernels launched at depths >= sync_depth will synchronize before returning to the parent, preventing unbounded pending launches.

child_grid: Dim3

Grid dimensions for child kernel launches.

child_block: Dim3

Block dimensions for child kernel launches.

child_shared_mem: u32

Dynamic shared memory allocation for child kernels (bytes).

sm_version: SmVersion

Target GPU architecture.

Implementations

impl DynamicParallelismConfig

pub fn new(sm_version: SmVersion) -> Self

Creates a new configuration with default values.

Defaults:

• max_nesting_depth: 4
• max_pending_launches: 2048
• sync_depth: 2
• child_grid: 128 blocks
• child_block: 256 threads
• child_shared_mem: 0

Trait Implementations

impl Clone for DynamicParallelismConfig

fn clone(&self) -> DynamicParallelismConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DynamicParallelismConfig

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

Formats the value using the given formatter.

impl Display for DynamicParallelismConfig

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

Formats the value using the given formatter.

impl PartialEq for DynamicParallelismConfig

fn eq(&self, other: &DynamicParallelismConfig) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Eq for DynamicParallelismConfig

impl StructuralPartialEq for DynamicParallelismConfig