Struct SLTToSIRLowerer 

pub struct SLTToSIRLowerer { /* private fields */ }

Implementations

impl SLTToSIRLowerer

pub fn new(four_state: bool) -> Self

pub fn with_unpacked_input_types<A: Hash + Eq + Clone>( self, arena: &SLTNodeArena<A>, element_widths: &FxHashMap<A, usize>, ) -> Self

pub fn lower<A: Hash + Eq + Clone + Debug + Display>( &self, builder: &mut SIRBuilder<A>, node: NodeId, arena: &SLTNodeArena<A>, cache: &mut FxHashMap<NodeId, RegisterId>, ) -> RegisterId

Recursively expand SLT nodes into SIR instructions

pub fn lower_fold_groups_jointly<A: Hash + Eq + Clone + Debug + Display>( &self, builder: &mut SIRBuilder<A>, roots: &[NodeId], arena: &SLTNodeArena<A>, cache: &mut FxHashMap<NodeId, RegisterId>, ) -> bool

Lower several independent recovered folds as one counted loop.

This entry point is intentionally transactional: a rejected family leaves both the builder and the materialization cache unchanged. The scheduler remains responsible for proving that the roots are mutually unordered in its dependency graph; this method rechecks every local property needed by the joint loop itself.

pub fn lower_with_inputs<A: Hash + Eq + Clone + Debug + Display>( &self, builder: &mut SIRBuilder<A>, node: NodeId, arena: &SLTNodeArena<A>, cache: &mut FxHashMap<NodeId, RegisterId>, inputs: FxHashMap<VarAtomBase<A>, RegisterId>, ) -> RegisterId

pub fn project_materialized<A>( &self, builder: &mut SIRBuilder<A>, value: RegisterId, access: BitAccess, ) -> RegisterId

Project a value which has already been lowered without retaining every projection at once. Grouped folds use this after the packed result has been computed, immediately before the corresponding state Store.

Trait Implementations

impl Drop for SLTToSIRLowerer

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.