Struct Flow 

pub struct Flow<'a> {
    pub enki: &'a Enki,
    pub queue: TaskQueue<'a>,
    pub cmd: CommandBuffer,
    pub slot_idx: usize,
    pub timeline_value: u64,
    pub submitted: AtomicBool,
    pub sticky_error: Option<Error>,
    pub borrow_ledger: FrameBorrowLedger,
    pub timestamp_count: u32,
}

An active, coherent GPU compute and presentation execution stream.

A Flow batches nam (kernel) dispatches and display presentation tasks into an atomic command buffer sequence synchronized by timeline semaphores.

Fields

enki: &'a Enki

queue: TaskQueue<'a>

cmd: CommandBuffer

slot_idx: usize

timeline_value: u64

submitted: AtomicBool

sticky_error: Option<Error>

borrow_ledger: FrameBorrowLedger

timestamp_count: u32

Implementations

impl<'a> Flow<'a>

pub fn new(enki: &'a Enki) -> Self

Initializes a new Flow, acquiring a command buffer from the ring and resetting query pools.

pub fn present<T: Copy + Send + Sync + 'static>(&mut self, pixels: &GpuVec<T>)

Queues an owned GPU vector for direct presentation to the display window.

Temporal Presentation Hazard

Once queued for presentation, any subsequent dispatch (.run() or .run_unchecked()) within the same flow that attempts to mutate (&mut) this buffer halts execution with diagnostic error[E1010]. Read-only access (&) after presentation remains permitted.

Diagnostics

Halts execution with diagnostic error[E2002] if element count is less than width * height.

pub fn present_slice<T: Copy + Send + Sync + 'static>( &mut self, pixels: &Slice<'_, T>, )

Queues a contiguous GPU slice for direct presentation to the display window.

Temporal Presentation Hazard

Once queued for presentation, any subsequent dispatch (.run() or .run_unchecked()) within the same flow that attempts to mutate (&mut) this buffer halts execution with diagnostic error[E1010]. Read-only access (&) after presentation remains permitted.

Diagnostics

Halts execution with diagnostic error[E2002] if element count is less than width * height. #[track_caller]

pub fn end_flow(self)

Finalizes and submits the recorded flow to the GPU queue, halting on failure.

pub fn try_end_flow(self) -> Result<()>

Finalizes and submits the recorded flow to the GPU queue, returning a Result.

pub fn mark(&mut self) -> GpuInstant

Records a hardware timestamp query mark on the GPU timeline for profiling.

pub fn write_timestamp(&mut self, query_index: u32, stage: PipelineStageFlags2)

Records a timestamp query into the active flow.

pub fn push_task(&mut self, task: RawTask<'a>)

Pushes a low-level task onto the internal execution queue.

Trait Implementations

impl<'a> Drop for Flow<'a>

fn drop(&mut self)

Executes the destructor for this type.

fn pin_drop(self: Pin<&mut Self>)

🔬This is a nightly-only experimental API. (pin_ergonomics)

Execute the destructor for this type, but different to Drop::drop, it requires self to be pinned.