Seismic data processing and interpretation are central to modern exploration workflows, where efficient parsing and visualization of SEG-Y files directly impacts operational success 8, 9. The extraction of inline and crossline sections from three-dimensional seismic post-stack cubes enables identification of subtle subsurface features including faults, horizons, facies, and potential hydrocarbon traps. Current processing solutions struggle with modern seismic datasets, particularly as real-time interpretation and machine learning applications demand sustained high-throughput data delivery to maintain computational efficiency, all the while existing processing tools frequently exhibit throughput limitations that constrain these advanced analytical approaches 10. Traditional implementations like segyio, while widely adopted, struggle particularly with crossline slicing operations where data must be gathered from non-contiguous file locations 1. The segfast library with memmap engine attempts to address these issues through memory-mapped file handling but encounters inconsistent performance across different scenarios 2. OpenVDS offers a brick- based volumetric format that can achieve superior performance in specific cases, but requires scanning and importing SEG-Y files into a proprietary data store before optimal access is possible, creating preprocessing overhead and workflow integration challenges 3. The Rust programming language combines memory safety, performance optimization, and concurrency capabilities that address these specific challenges 4, 12, 13. Unlike C-based implementations that risk memory management errors, Rust's ownership model provides guaranteed memory safety without performance penalties 14. The language's zero-cost abstractions enable fine-grained optimization of critical processing paths, while sophisticated concurrency primitives allow efficient utilization of multi-core processing environments common in modern interpretation workstations. This research presents a Rust-based engine for accelerating SEG-Y file slicing through optimized memory management, buffer allocation, and concurrent processing. We benchmark performance across different access patterns, dataset sizes, and storage configurations, comparing against segyio, segfast with memmap, and OpenVDS using controlled testing with statistical validation. The Rust-based approach achieves extraction speeds up to 4x faster than traditional implementations, with the most pronounced gains in random crossline access operations. Unlike format-conversion approaches, this implementation reads SEG-Y files directly without preprocessing, allowing immediate integration into existing workflows.
Дубовик et al. (Mon,) studied this question.