Module parallel_executor 

Parallel executor implementation using Rayon for multi-threaded execution.

This module provides a parallel implementation of the TensorLogic executor that can execute independent operations concurrently using thread pools.

Key Features

• Level-by-level execution: Operations are grouped by execution level
• Rayon thread pools: Configurable thread pool for parallel execution
• Automatic dependency handling: Uses DependencyAnalysis for safe parallelization
• Performance monitoring: Tracks parallel vs sequential execution times

Example

use tensorlogic_scirs_backend::ParallelScirs2Exec;
use tensorlogic_infer::TlAutodiff;

let mut executor = ParallelScirs2Exec::new();
executor.set_num_threads(4); // Use 4 threads

let result = executor.forward(&graph)?;

Structs

ParallelConfig

Configuration for parallel execution.

ParallelScirs2Exec

Parallel executor using Rayon for multi-threaded execution.

ParallelStats

Statistics about parallel execution.