vector.atomic.rmw

← vector dialect

Atomic read-modify-write at per-lane signed element offsets. Each lane atomically combines its value with origin + offsets[lane] and the result lane is the old memory value observed by that atomic operation.

Operation contract

Property	Value
Semantic phase	—
Interfaces	MemoryAccess
Memory effects	readwrite view

Signature

Kind	Name	Type	Cardinality	Description
Operand	value	vector	required	Vector update value for each lane.
Operand	view	view	required	Typed destination view.
Operand	offsets	vector	required	Per-lane signed element offsets from the logical origin.
Operand	indices	index	variadic	Dynamic logical origin indices.
Result	result	vector	required	Old memory values read by the atomic operations.
Attribute	kind	enum AtomicKind	required	—
Attribute	ordering	enum AtomicOrdering	required	Required atomic memory ordering.
Attribute	scope	enum AtomicScope	required	Required atomic synchronization scope.
Attribute	cache_scope	enum CacheScope	optional	Optional cache/coherency scope required by target lowering.
Attribute	cache_temporal	enum CacheTemporal	optional	Optional temporal cache policy required by target lowering.
Attribute	static_indices	i64_array	required	Static logical origin indices with INT64_MIN sentinels for dynamics.

Verification constraints

• HasIndexOrNonI1IntegerElement(offsets)
• SameElementType(value, view, result)
• SameShape(offsets, value, result)
• SameType(value, result)

Examples

%old = vector.atomic.rmw<addi> %v, %view[%row, %col][%offsets] {ordering = relaxed, scope = workgroup} : vector<4xi32>, view<[%m]x[%n]xi32, %layout>, vector<4xindex>