Expand description
Planning (docs/SPEC.md §2.5): the HardwarePlacementRule — a DataFusion
PhysicalOptimizerRule that rewrites eligible FilterExec /
ProjectionExec / HashJoinExec / AggregateExec nodes into Gpu*Exec
nodes — and the OxidePhysicalCodec, a PhysicalExtensionCodec that lets
Ballista ship those nodes to executors.
Placement is decided against a target backend: the local hardware in
embedded mode, the declared cluster capability (OXIDE_CLUSTER_BACKEND) on
the Ballista scheduler. EXPLAIN shows the target as a tag, e.g.
GpuFilterExec[cuda]; every node still selects its real backend at
execute time and falls back to the CPU path when the device is absent.
Dependency direction: depends on oxidelake-core, oxidelake-compute and
datafusion-proto (the single datafusion-* sub-crate exception, ADR-0009).
Re-exports§
pub use codec::OxidePhysicalCodec;pub use rule::HardwarePlacementRule;pub use rule::VectorDistancePlacementRule;pub use rule::physical_optimizer_rules;pub use target::CLUSTER_TARGET_ENV;pub use target::cluster_target_from_env;pub use target::resolve_cluster_target;