Recent advances in AI-generated content have driven widespread adoption of Diffusion Transformers (DiTs) for high-resolution, long-duration content generation. While parallelization techniques accelerate diffusion inference, they face significant scalability challenges due to excessive communication overhead in multi-node environments. We observe that sequence partitions in Context Parallelism (CP) exhibit distinct heterogeneity: spatially proximate partitions contribute more significantly to attention computation results. By mapping this heterogeneous pattern to hierarchical communication topology, we can access high-contribution partitions with reduced communication cost. This insight motivates our novel selective attention state mechanism that strategically balances partial attention computation and historical result reuse across denoising steps. We present DiTango, an efficient parallel framework for DiT generation. DiTango features an anchor-guided state selection planner that optimizes computation-reuse decisions for each partition, complemented by a runtime that orchestrates efficient state-centric operations. This design achieves superior system efficiency while preserving generation quality. Experimental evaluation on popular diffusion models demonstrates that DiTango achieves up to 1.9x end-to-end and 3.2x attention speedup with near-linear scaling in multi-node settings, while maintaining generation quality comparable to state-of-the-art approaches.

We present Odin, a training system for distributed point-based neural rendering (PBNR) that accelerates training by removing unnecessary global barriers. Odin observes that two training steps need synchronization only when they update overlapping visible parameters. It uses lightweight visibility metadata and a two-phase design with ahead-of-time scheduling plus runtime refinement to handle irregular and dynamic behavior. Odin hides 82% of exposed communication and achieves up to 1.89 × throughput speedup on 64 GPUs while preserving baseline accuracy.

Recent advances in 3D Gaussian Splatting (3DGS) have demonstrated significant potential over traditional rendering techniques, attracting widespread attention from both industry and academia. However, real-time rendering with 3DGS remains a challenging problem, particularly in large-scale, high-resolution scenes due to the presence of numerous anisotropic Gaussian representations, and it has not been extensively explored. To address this challenge, we introduce FlashGS, an open-source CUDA library with Python bindings, featuring comprehensive algorithm design and optimizations, including redundancy elimination, adaptive scheduling, and efficient pipelining. First, we eliminate substantial redundant computations through precise Gaussian intersection tests, leveraging the intrinsic mechanism of the 3DGS rasterizer. During task partitioning, we propose an adaptive scheduling strategy that accounts for variations in Gaussian size and shape. Additionally, we design a multi-stage pipelining strategy for color computation in the rendering process, further accelerating performance. We conduct an extensive evaluation of FlashGS across a diverse range of synthetic and real-world 3D scenes, encompassing scene sizes of up to 2.7 km2cityscape and resolutions of up to over 4K. Our approach improves 3DGS rendering performance by an order of magnitude, achieving an average speedup of 7.2×, and rendering at a minimum of 125.9 FPS, setting a new state-of-the-art in real-time 3DGS rendering. https://github.com/InternLandMark/FlashGS.

3D Gaussian Splatting (3DGS) renders pixels by rasterizing Gaussian primitives, where conditional alpha-blending dominates the computational cost in the rendering pipeline. This paper proposes TC-GS, an algorithm-independ-ent universal module that expands the applicability of Tensor Core (TCU) for 3DGS, leading to substantial speedups and seamless integration into existing 3DGS optimization frameworks. The key innovation lies in mapping alpha computation to matrix multiplication, fully utilizing otherwise idle TCUs in existing 3DGS implementations. TC-GS provides plug-and-play acceleration for existing top-tier acceleration algorithms and integrates seamlessly with rendering pipeline designs, such as Gaussian compression and redundancy elimination algorithms. Additionally, we introduce a global-to-local coordinate transformation to mitigate rounding errors from quadratic terms of pixel coordinates caused by Tensor Core half-precision computation. Extensive experiments demonstrate that our method maintains rendering quality while providing an additional 2.18× speedup over existing Gaussian acceleration algorithms, thereby achieving a total acceleration of up to 5.6×.

Pipeline parallelism is a crucial technique for large-scale model training, enabling parameter splitting and performance enhancement. However, creating effective pipeline schedules often requires significant manual effort and coding skills, leading to practical inconveniences and complex debugging. Major frameworks such as DeepSpeed and ColossalAI simplify the process by adopting predefined pipeline schedule strategies, such as GPipe and 1F1B. The use of predefined schedules offers limited flexibility and suboptimal training efficiency, as the limited number of manually set candidates cannot provide the optimal strategy for arbitrary model training. To deal with the issue, this article aims to automatically search for the optimal strategy with high efficiency. Since current frameworks only support a limited set of fixed strategies, lacking the technical capability to create a comprehensive strategy search space, we first design a novel domain-specific language (DSL) for pipeline schedule development. The DSL exhibits great understandability, agility, and reusability, supporting the development of all known pipeline schedule strategies and their variants. Second, we are the first to model the complete pipeline schedule strategy space via the DSL, enabling an automated end-to-end globally optimal pipeline schedule searching, while past work may get stuck in a local optimum. Finally, we propose to optimize pipeline performance by modeling and solving the pipeline schedule as a Binary-Tree-Traversing (BTT) optimization problem. Based on the formalization, we further adopt a Dynamic Try-Test Genetic Algorithm to search for the best pipeline schedule strategy, which overwhelms a variety of pre-defined ones. Experimental results show that Koala achieves an enhanced performance by up to \(1.53\times\) over state-of-the-art approaches. Besides, the pipeline schedule strategy searched by Koala outperforms pre-defined pipeline schedule strategies by \(1.10\times \sim 1.55\times\) . Moreover, Koala has superior scalability and effectiveness in combining with data parallelism and tensor parallelism.

Strategies for partitioning an application’s data and computation play fundamental role in determining the efficiency of parallelization. This paper describes a sophisticated strategy for partitioning data and computation known as multi-partitioning, which can support the best parallelization for some applications such as the line sweep computations. However, the implementation of multi-partitioning is very difficult and, as we know, there is none automatic parallelizing compiler supports such partitioning strategy. Though the dHPF compiler implemented multi-partitioning as a special extension for block style HPF partitioning, it still needs the programmer’s participation to analyze the application and decide the data distribution scheme. In this paper, we present a global tiling transformation algorithm and a tile-to-processors mapping strategy called hyper-diagonal modular mapping, to implement the multi-partitioning strategy. The experimentation with NPB2.3-serial SP shows that the code generated by the compiler achieves scalable performance.