pub fn predict_batch_parallel(
source: &(dyn ObservableEphemerisSource + Sync),
requests: &[PredictRequest],
options: PredictOptions,
) -> Vec<Result<PredictedObservables, ObservablesError>>Expand description
Predict observables for many (satellite, receiver, epoch) requests, fanning
the independent requests across a rayon thread pool.
Each request is evaluated by the same scalar predict kernel and the
indexed parallel collect preserves input order, so element i is
byte-for-byte identical to element i of predict_batch: the requests
share no mutable state and a single predict is internally sequential, so
throughput scales with cores while every value stays bit-exact. The
source must be Sync because it is read concurrently from every worker.