arXiv subjects

Albert-Jan Yzelman

Publications and source records attributed to Albert-Jan Yzelman.

6 recordsLinked to original sources

arXiv2026

Nonlinear spectral clustering with C++ GraphBLAS

Nonlinear reformulations of the spectral clustering method have gained a lot of recent attention due to their increased numerical benefits and their solid mathematical background. However, the estimation of the multiple nonlinear eigenvectors is associated with an increased computational cost. We present an implementation of a direct multiway spectral clustering algorithm in the $p$-norm, for $p\in(1,2]$, using a novel C++ GraphBLAS API. The key operations are expressed in linear algebraic terms and are executed over the resulting sparse matrices and dense vectors, parameterized in the algebra pertinent to the computation. We demonstrate the effectiveness and accuracy of our shared-memory algorithm on several artificial test cases. Our numerical examples and comparative results against competitive methods indicate that the proposed implementation attains high quality clusters in terms of the balanced graph cut metric. The strong scaling capabilities of our algorithm are showcased on a range of datasets with up to $8$ million nodes and $48$ million edges.

cs.DC↗

arXiv2025

Faster Distributed Inference-Only Recommender Systems via Bounded Lag Synchronous Collectives

Recommender systems are enablers of personalized content delivery, and therefore revenue, for many large companies. In the last decade, deep learning recommender models (DLRMs) are the de-facto standard in this field. The main bottleneck in DLRM inference is the lookup of sparse features across huge embedding tables, which are usually partitioned across the aggregate RAM of many nodes. In state-of-the-art recommender systems, the distributed lookup is implemented via irregular all-to-all (alltoallv) communication, and often presents the main bottleneck. Today, most related work sees this operation as a given; in addition, every collective is synchronous in nature. In this work, we propose a novel bounded lag synchronous (BLS) version of the alltoallv operation. The bound can be a parameter allowing slower processes to lag behind entire iterations before the fastest processes block. In special applications such as inference-only DLRM, the accuracy of the application is fully preserved. We implement BLS alltoallv in a new PyTorch Distributed backend and evaluate it with a BLS version of the reference DLRM code. We show that for well balanced, homogeneous-access DLRM runs our BLS technique does not offer notable advantages. But for unbalanced runs, e.g. runs with strongly irregular embedding table accesses or with delays across different processes, our BLS technique improves both the latency and throughput of inference-only DLRM. In the best-case scenario, the proposed reduced synchronisation can mask the delays across processes altogether.

cs.DC↗

arXiv2025

HiCR, an Abstract Model for Distributed Heterogeneous Programming

We present HiCR, a model to represent the semantics of distributed heterogeneous applications and runtime systems. The model describes a minimal set of abstract operations to enable hardware topology discovery, kernel execution, memory management, communication, and instance management, without prescribing any implementation decisions. The goal of the model is to enable execution in current and future systems without the need for significant refactoring, while also being able to serve any governing parallel programming paradigm. In terms of software abstraction, HiCR is naturally located between distributed heterogeneous systems and runtime systems. We coin the phrase \emph{Runtime Support Layer} for this level of abstraction. We explain how the model's components and operations are realized by a plugin-based approach that takes care of device-specific implementation details, and present examples of HiCR-based applications that operate equally on a diversity of platforms.

cs.DC↗

arXiv2025

Distributed and heterogeneous tensor-vector contraction algorithms for high performance computing

The tensor-vector contraction (TVC) is the most memory-bound operation of its class and a core component of the higher-order power method (HOPM). This paper brings distributed-memory parallelization to a native TVC algorithm for dense tensors that overall remains oblivious to contraction mode, tensor splitting and tensor order. Similarly, we propose a novel distributed HOPM, namely dHOPM3, that can save up to one order of magnitude of streamed memory and is about twice as costly in terms of data movement as a distributed TVC operation (dTVC) when using task-based parallelization. The numerical experiments carried out in this work on three different architectures featuring multi-core and accelerators confirm that the performances of dTVC and dHOPM3 remain relatively close to the peak system memory bandwidth (50%-80%, depending on the architecture) and on par with STREAM benchmark figures. On strong scalability scenarios, our native multi-core implementations of these two algorithms can achieve similar and sometimes even greater performance figures than those based upon state-of-the-art CUDA batched kernels. Finally, we demonstrate that both computation and communication can benefit from mixed precision arithmetic also in cases where the hardware does not support low precision data types natively.

cs.DC↗

arXiv2023

Open Problems in (Hyper)Graph Decomposition

Large networks are useful in a wide range of applications. Sometimes problem instances are composed of billions of entities. Decomposing and analyzing these structures helps us gain new insights about our surroundings. Even if the final application concerns a different problem (such as traversal, finding paths, trees, and flows), decomposing large graphs is often an important subproblem for complexity reduction or parallelization. This report is a summary of discussions that happened at Dagstuhl seminar 23331 on "Recent Trends in Graph Decomposition" and presents currently open problems and future directions in the area of (hyper)graph decomposition.

cs.DS↗

arXiv2023

Effective implementation of the High Performance Conjugate Gradient benchmark on GraphBLAS

Applications in High-Performance Computing (HPC) environments face challenges due to increasing complexity. Among them, the increasing usage of sparse data pushes the limits of data structures and programming models and hampers the efficient usage of existing, highly parallel hardware. The GraphBLAS specification tackles these challenges by proposing a set of data containers and primitives, coupled with a semantics based on abstract algebraic concepts: this allows multiple applications on sparse data to be described with a small set of primitives and benefit from the many optimizations of a compile-time-known algebraic specification. Among HPC applications, the High Performance Conjugate Gradient (HPCG) benchmark is an important representative of a large body of sparse workloads, and its structure poses several programmability and performance challenges. This work tackles them by proposing and evaluating an implementation on GraphBLAS of HPCG, highlighting the main changes to its kernels. The results for shared memory systems outperforms the reference, while results in distributed systems highlight fundamental limitations of GraphBLAS-compliant implementations, which suggests several future directions.

cs.DC↗