Related papers: GPU performance in Run3 ALICE online/offline recon…

The Time Projection Chamber is the main tracking and particle identification detector of the ALICE experiment. The high luminosities delivered by the CERN LHC in Run 2 (2015-2018) posed new challenges in terms of detector performance and…

Efficient Graph processing is challenging because of the irregularity of graph algorithms. Using GPUs to accelerate irregular graph algorithms is even more difficult to be efficient, since GPU's highly structured SIMT architecture is not a…

At the Large Hadron Collider at CERN in Geneva, Switzerland, atomic nuclei are collided at ultra-relativistic energies. Many final-state particles are produced in each collision and their properties are measured by the ALICE detector. The…

Real-time data processing is one of the central processes of particle physics experiments which require large computing resources. The LHCb (Large Hadron Collider beauty) experiment will be upgraded to cope with a particle bunch collision…

Charged-particle reconstruction is a fundamental part of the event reconstruction in modern multi-purpose high-energy physics detectors. This paper describes the algorithms used to reconstruct charged particles and primary vertices with the…

A Large Ion Collider Experiment (ALICE) has been conceived and constructed as a heavy-ion experiment at the LHC. During LHC Runs 1 and 2, it has produced a wide range of physics results using all collision systems available at the LHC. In…

We present robust high-performance implementations of signal-processing tasks performed by a high-throughput wildlife tracking system called ATLAS. The system tracks radio transmitters attached to wild animals by estimating the time of…

Measurements of the suppression of high-pT particles and the away-side jets from heavy-ion collisions at RHIC have shown that medium-induced energy loss affects partons produced in the early stage of a heavy-ion collision. At LHC energies…

It is important to scale out deep neural network (DNN) training for reducing model training time. The high communication overhead is one of the major performance bottlenecks for distributed DNN training across multiple GPUs. Our…

The LHC Run 5 and 6 data taking phases will bring unprecedented luminosity in high-energy proton-proton and in heavy-ion collisions. The ALICE Collaboration proposes a next-generation experiment, ALICE 3, specifically designed to operate…

Learned image compression allows achieving state-of-the-art accuracy and compression ratios, but their relatively slow runtime performance limits their usage. While previous attempts on optimizing learned image codecs focused more on the…

Analog in-memory computing (AIMC) cores offers significant performance and energy benefits for neural network inference with respect to digital logic (e.g., CPUs). AIMCs accelerate matrix-vector multiplications, which dominate these…

We study the performance of a cloud-based GPU-accelerated inference server to speed up event reconstruction in neutrino data batch jobs. Using detector data from the ProtoDUNE experiment and employing the standard DUNE grid job submission…

The era of GPU-powered data analytics has arrived. In this paper, we argue that recent advances in hardware (e.g., larger GPU memory, faster interconnect and IO, and declining cost) and software (e.g., composable data systems and mature…

Tensor decompositions, such as CANDECOMP/PARAFAC (CP), are widely used in a variety of applications, such as chemometrics, signal processing, and machine learning. A broadly used method for computing such decompositions relies on the…

Analysis of processing time and similarity of images generated between CPU and GPU architectures and sequential and parallel programming. For image processing a computer with AMD FX-8350 processor and an Nvidia GTX 960 Maxwell GPU was used,…

While GPUs dominate massively parallel computing through the single-instruction, multiple-thread (SIMT) programming model, their underlying single-instruction, multiple-data (SIMD) execution incurs substantial energy overhead from frequent…

Graph neural networks (GNN) analysis engines are vital for real-world problems that use large graph models. Challenges for a GNN hardware platform include the ability to (a) host a variety of GNNs, (b) handle high sparsity in input vertex…

Transformers and LLMs have seen rapid adoption in all domains. Their sizes have exploded to hundreds of billions of parameters and keep increasing. Under these circumstances, the training of transformers is slow and often takes in the order…

Counting (p,q)-bicliques in bipartite graphs poses a foundational challenge with broad applications, from densest subgraph discovery in algorithmic research to personalized content recommendation in practical scenarios. Despite its…