Search Results

We have built openGauss, an enterprise-grade open-source database system. openGauss has fulfilled its design goal of high performance, high availability, high security, and high intelligence. For high performance, it leverages NUMA (non-uniform ...

We delve into the performance of transformer encoder inference on low-power multi-core processors from two perspectives: First, we conduct a detailed profile of the inference process for two members of the BERT family on a modern multi-core ...

This study presents the first constrained sparse tensor factorization (cSTF) framework that optimizes and fully offloads computation to massively parallel GPU architectures, and the first performance characterization of cSTF on GPU architectures. In ...

This paper applies spatial and temporal adaptivity to an existing discrete element method (DEM) based snow simulation on the GPU. For spatial adaptivity, visually significant spatial regions are identified and simulated at varying resolutions. To this ...

This paper investigates the design of parallel general matrix multiplication (gemm) for a Versal Adaptive Compute Accelerated Platform (ACAP) equipped with a VC1902 system-on-chip and multiple Artificial Intelligence Engines (AIEs). Our efforts aim to ...

Data-parallel deep neural networks (DNN) training systems deployed across nodes have been widely used in various domains, while the system performance is often bottlenecked by the communication overhead among workers for synchronizing gradients. Top-k ...

The Edge Computing environments facilitate the creation of pervasive applications distributed across vast geographical regions, addressing particular challenges associated with centralized information processing, such as network bandwidth saturation and ...

We explore the utilization of the Apache TVM open source framework to automatically generate a family of algorithms that follow the approach taken by popular linear algebra libraries, such as GotoBLAS2, BLIS, and OpenBLAS, to obtain high-performance ...

The optimization of the matrix multiplication (or GEMM) has been a need during the last decades. This operation is considered the flagship of current linear algebra libraries such as BLIS, OpenBLAS, or Intel OneAPI because of its widespread use in a ...

Graph processing has been widely used in many scenarios, from scientific computing to artificial intelligence. Graph processing exhibits irregular computational parallelism and random memory accesses, unlike traditional workloads. Therefore, ...

Malleability is defined as the ability to vary the degree of parallelism at runtime, and is regarded as a means to improve core occupation on state-of-the-art multicore processors tshat contain tens of computational cores per socket. This property is ...

We investigate the solution of low-rank matrix approximation problems using the truncated singular value decomposition (SVD). For this purpose, we develop and optimize graphics processing unit (GPU) implementations for the randomized SVD and a blocked ...

Traditional data centers include monolithic servers that tightly integrate CPU, memory and disk (Figure 1a). Instead, Disaggregated Systems (DSs) [8, 13, 18, 27] organize multiple compute (CC), memory (MC) and storage devices as independent, failure-...

Following the recent advances in open hardware generally, and RISC-V architectures particularly, we analyse the performance of transformer encoder inference on three low-power platforms with this type of architecture. For this purpose, we conduct ...

Realizing a high-performance and energy-efficient circuit system is one of the critical tasks for circuit designers. Conventional researchers always concentrated on the tradeoffs between the energy and the performance in circuit and system design ...

Resource disaggregation offers a cost effective solution to resource scaling, utilization, and failure-handling in data centers by physically separating hardware devices in a server. Servers are architected as pools of processor, memory, and storage ...

The research paper presents the design methodology with novel task distribution technique on multi-processor system on chip (MPSoC) for speeding up the execution of arithmetic application. Utilisation of multiple soft core processors on field programmable ...

The modern in-memory database (IMDB) can support highly concurrent on-line transaction processing (OLTP) workloads and generate massive transactional logs per second. Quorum-based replication protocols such as Paxos or Raft have been widely used ...

We present a new parallel task-based algorithm for randomized low-rank factorizations of a matrix and its application to fast hierarchical solvers. TaskTorrent is a lightweight, distributed, task-based runtime in C++. We explain how the randomized ...

Randomized singular value decomposition (RSVD) is by now a well-established technique for efficiently computing an approximate singular value decomposition of a matrix. Building on the ideas that underpin RSVD, the recently proposed algorithm “randUTV” ...