Struct TaskQueueSnapshot 

pub struct TaskQueueSnapshot {
    pub ready_count: u32,
    pub paused_count: u32,
    pub yielded_count: u32,
    pub requeued_count: u32,
    pub running_count: u32,
    pub faulted_count: u32,
    pub total_requeues: u64,
    pub max_priority_age: u32,
}

Queue telemetry derived from device-visible continuation task slots.

Fields

ready_count: u32

Count of ready slots.

paused_count: u32

Count of paused slots.

yielded_count: u32

Count of yielded slots.

requeued_count: u32

Count of requeued slots.

running_count: u32

Count of running slots.

faulted_count: u32

Count of faulted slots.

total_requeues: u64

Sum of per-slot requeue counters.

max_priority_age: u32

Maximum priority-aging tick observed in the queue.

Implementations

impl TaskQueueSnapshot

pub fn from_tasks(tasks: &[TaskWorkItem]) -> Result<Self, BackendError>

Build a queue snapshot from continuation task slots.

Errors

Returns BackendError when the slice contains an unknown task state or a count cannot fit the u32 ABI.

pub fn schedulable_count(&self) -> u32

Number of slots immediately eligible for GPU scheduling.

pub fn try_schedulable_count(&self) -> Result<u32, BackendError>

Checked number of slots immediately eligible for GPU scheduling.

Errors

Returns BackendError when the summed schedulable count exceeds the u32 launch ABI.

pub fn continuation_pressure_count(&self) -> u64

Number of slots carrying continuation pressure.

pub fn try_continuation_pressure_count(&self) -> Result<u64, BackendError>

Checked number of slots carrying continuation pressure.

Errors

Returns BackendError when continuation pressure exceeds u64.

pub fn build_state_convergence_program( transfer_body: Vec<Node>, current_buffer: &str, next_buffer: &str, changed_buffer: &str, words: u32, max_iterations: u32, ) -> Program

Build a Program that runs a one-shot persistent fixpoint over the queue snapshot’s state-counter buffer, converging the counts to a stable equilibrium representing the queue’s long-run distribution. Wires the self-substrate persistent fixpoint builder for observability collectors that want stable signals over transient queue jitter.

current_buffer / next_buffer / changed_buffer are caller-supplied buffer names for the persistent_fixpoint ping-pong; transfer_body is the per-iteration body that reads current and writes next. Returns a Program suitable for one dispatch.

P-RUNTIME-4: replaces a host-side jitter-smoothing loop with a single GPU-side fixpoint dispatch.

pub fn apply_to_launch_request( &self, request: MegakernelLaunchRequest, ) -> MegakernelLaunchRequest

Merge this queue telemetry into a launch request.

pub fn try_apply_to_launch_request( &self, request: MegakernelLaunchRequest, ) -> Result<MegakernelLaunchRequest, BackendError>

Checked merge of queue telemetry into a launch request.

Errors

Returns BackendError when schedulable count or continuation pressure cannot fit the launch request ABI.

Trait Implementations

impl Clone for TaskQueueSnapshot

fn clone(&self) -> TaskQueueSnapshot

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for TaskQueueSnapshot

impl Debug for TaskQueueSnapshot

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

Formats the value using the given formatter.

impl Default for TaskQueueSnapshot

fn default() -> TaskQueueSnapshot

Returns the “default value” for a type.

impl Eq for TaskQueueSnapshot

impl PartialEq for TaskQueueSnapshot

fn eq(&self, other: &TaskQueueSnapshot) -> 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 StructuralPartialEq for TaskQueueSnapshot