arXiv subjects

Stavros Kalafatis

Publications and source records attributed to Stavros Kalafatis.

4 recordsLinked to original sources

arXiv2026

Splaxel: Efficient Distributed Training of 3D Gaussian Splatting for Large-scale Scene Reconstruction via Pixel-level Communication

3D Gaussian Splatting (3DGS) enables high-fidelity and real-time 3D scene reconstruction, but scaling training to large-scale scenes requires optimizing hundreds of millions of Gaussians across multiple GPUs. Existing distributed approaches either partition scenes into isolated regions, causing global inconsistency, or rely on global Gaussian-level exchanges, which lead to substantial growth in inter-GPU communication and quickly dominate iteration time. We propose Splaxel, a communication-efficient distributed 3DGS training framework based on pixel-level local rendering and global composition. Instead of synchronizing Gaussians, each GPU renders its local subset and exchanges only partial pixel values, maintaining mathematical consistency while keeping communication cost stable as the scene size increases. Splaxel further reduces pixel-level redundancy through geometric and transmittance visibility prediction and improves GPU utilization via conflict-free camera-view consolidation. Evaluated on large-scale datasets with up to 120M Gaussians, Splaxel achieves up to 7.6$\times$ speedup over the state-of-the-art distributed 3DGS framework while preserving high reconstruction quality.

cs.DC↗

arXiv2026

ORAP: Optimized Row Access Prefetching for Rowhammer-mitigated Memory

Rowhammer is a well-studied DRAM phenomenon wherein multiple activations to a given row can cause bit flips in adjacent rows. Many mitigation techniques have been introduced to address Rowhammer, with some support being incorporated into the JEDEC DDR5 standard for per-row-activation-counter (PRAC) and refresh-management (RFM) systems. Mitigation schemes built on these mechanisms claim to have various levels of area, power, and performance overheads. To date the evaluation of existing mitigation schemes typically neglects the impact of other memory system components such as hardware prefetchers. Nearly all modern systems incorporate hardware prefetching and these can significantly improve processor performance through speculative cache population. These prefetchers induce higher numbers of downstream memory requests and increase DRAM activation rates. The performance overhead of Rowhammer mitigations are tied directly to memory access patterns, exposing both hardware prefetchers and Rowhammer mitigations to cross-interaction. We find that the performance improvement provided by prior-work hardware prefetchers is often severely impacted by Rowhammer mitigations. In effect, much of the benefit of speculative memory references from prefetching lies in accelerating and reordering DRAM references in ways that trigger mitigations, significantly reducing the benefits of prefetching. This work proposes the Optimized Row Access Prefetcher (ORAP), leveraging last-level-cache (LLC) space to cache large portions of DRAM rowbuffer contents to reduce the need for future activations. Working with the state-of-the-art Berti prefetcher, ORAP reduces DRAM activation rates by 51.3% and achieves a 4.6% speedup over the prefetcher configuration of Berti and SPP-PPF when prefetching in an RFM-mitigated memory system. Under PRAC mitigations, ORAP reduces energy overheads by 11.8%.

cs.AR↗

arXiv2026

SPPAM: Signature Pattern Prediction and Access-Map Prefetcher

The discrepancy between processor speed and memory system performance continues to limit the performance of many workloads. To address the issue, one effective and well studied technique is cache prefetching. Many prefetching designs have been proposed, with varying approaches and effectiveness. For example, SPP is a popular prefetcher that leverages confidence throttled recursion to speculate on the future path of program's references, however it is very susceptible to the reference reordering of higher-level caches and the out-of-order core. Orthogonally, AMPM is another popular approach to prefetching which uses reordering-resistant access maps to identify patterns within a region, but is unable to speculate beyond that region. In this paper, we propose SPPAM, a new approach to prefetching, inspired by prior works such as SPP and AMPM, while addressing their limitations. SPPAM utilizes online-learning to build a set of access-map patterns. These patterns are used in a speculative lookahead which is throttled by a confidence metric. Targeting the second-level cache, SPPAM alongside state-of-the-art prefetchers Berti and Bingo improves system performance by 31.4% over no prefetching and 6.2% over the baseline of Berti and Pythia.

cs.AR↗

arXiv2024

NeurLZ: An Online Neural Learning-Based Method to Enhance Scientific Lossy Compression

Large-scale scientific simulations generate massive datasets, posing challenges for storage and I/O. Traditional lossy compression struggles to advance more in balancing compression ratio, data quality, and adaptability to diverse scientific data features. While deep learning-based solutions have been explored, their common practice of relying on large models and offline training limits adaptability to dynamic data characteristics and computational efficiency. To address these challenges, we propose NeurLZ, a neural method designed to enhance lossy compression by integrating online learning, cross-field learning, and robust error regulation. Key innovations of NeurLZ include: (1) compression-time online neural learning with lightweight skipping DNN models, adapting to residual errors without costly offline pertaining, (2) the error-mitigating capability, recovering fine details from compression errors overlooked by conventional compressors, (3) $1\times$ and $2\times$ error-regulation modes, ensuring strict adherence to $1\times$ user-input error bounds strictly or relaxed 2$\times$ bounds for better overall quality, and (4) cross-field learning leveraging inter-field correlations in scientific data to improve conventional methods. Comprehensive evaluations on representative HPC datasets, e.g., Nyx, Miranda, Hurricane, against state-of-the-art compressors show NeurLZ's effectiveness. During the first five learning epochs, NeurLZ achieves an 89% bit rate reduction, with further optimization yielding up to around 94% reduction at equivalent distortion, significantly outperforming existing methods, demonstrating NeurLZ's superior performance in enhancing scientific lossy compression as a scalable and efficient solution.

cs.DC↗