Physics Data Processing Research Articles

Quantum speedup for track reconstruction in particle accelerators

To investigate the fundamental nature of matter and its interactions, particles are accelerated to very high energies and collided inside detectors, producing a multitude of other particles that are scattered in all directions. As charged particles traverse the detector, they leave signals of their passage. The problem of track reconstruction is to recover the original trajectories from these signals. This challenging data analysis task will become even more demanding as the luminosity of future accelerators increases, leading to collision events with a more complex structure. We identify four fundamental routines present in every local tracking method and analyse how they scale in the context of a standard tracking algorithm. We show that for some of these routines we can reach a lower computational complexity with quantum search algorithms. Although the found quantum speedups are mild, this constitutes, to the best of our knowledge, the first rigorous evidence of a quantum advantage for a high-energy physics data processing task.

Increasing Parallelism in the ROOT I/O Subsystem

When processing large amounts of data, the rate at which reading and writing can take place is a critical factor. High energy physics data processing relying on ROOT is no exception. The recent parallelisation of LHC experiments’ software frameworks and the analysis of the ever increasing amount of collision data collected by experiments further emphasised this issue underlying the need of increasing the implicit parallelism expressed within the ROOT I/O.In this contribution we highlight the improvements of the ROOT I/O subsystem which targeted a satisfactory scaling behaviour in a multithreaded context. The effect of parallelism on the individual steps which are chained by ROOT to read and write data, namely (de)compression, (de)serialisation, access to storage backend, are discussed. Performance measurements are discussed through real life examples coming from CMS production workflows on traditional server platforms and highly parallel architectures such as Intel Xeon Phi.

BESIU Physical Analysis on Hadoop Platform

In the past 20 years, computing cluster has been widely used for High Energy Physics data processing. The jobs running on the traditional cluster with a Data-to-Computing structure, have to read large volumes of data via the network to the computing nodes for analysis, thereby making the I/O latency become a bottleneck of the whole system. The new distributed computing technology based on the MapReduce programming model has many advantages, such as high concurrency, high scalability and high fault tolerance, and it can benefit us in dealing with Big Data. This paper brings the idea of using MapReduce model to do BESIII physical analysis, and presents a new data analysis system structure based on Hadoop platform, which not only greatly improve the efficiency of data analysis, but also reduces the cost of system building. Moreover, this paper establishes an event pre-selection system based on the event level metadata(TAGs) database to optimize the data analyzing procedure.

CLARA: A Contemporary Approach to Physics Data Processing

CLARA (CLAS12 Reconstruction and Analysis framework) is CLAS12 physics data processing (PDP) application development framework based on a service oriented architecture (SOA). This framework allows users to design and deploy data processing services as well as dynamically compose PDP applications using available services. Services can be written in Java, C++, and Python languages. The PDP service bus provides a layer on top of a distributed pub-sub middleware implementation. This allows complex service composition and integration without writing a code. We believe that by deviating from the traditional self contained, monolithic PDP application models we can improve maintenance, scalability and quality of physics data analysis. The SOA approach also helps us to separate a specific service programmer from a PDP application designer. Examples of service creation and deployment, along with the CLAS12 track reconstruction application design are presented.

Data access performance through parallelization and vectored access. Some results.

High Energy Physics data processing and analysis applications typically deal with the problem of accessing and processing data at high speed. Recent studies, development and test work have shown that the latencies due to data access can often be hidden by parallelizing them with the data processing, thus giving the ability to have applications which process remote data with a high level of efficiency.Techniques and algorithms able to reach this result have been implemented in the client side of the Scalla/xrootd system, and in this contribution we describe the results of some tests done in order to compare their performance and characteristics. These techniques, if used together with multiple streams data access, can also be effective in allowing to efficiently and transparently deal with data repositories accessible via a Wide Area Network.

A Comparison of Data-Access Platforms for the Computing of Large Hadron Collider Experiments

Performance, reliability and scalability in data-access are key issues in the context of the computing Grid and High Energy Physics data processing and analysis applications, in particular considering the large data size and I/O load that a Large Hadron Collider data centre has to support. In this paper we present the technical details and the results of a large scale validation and performance measurement employing different data-access platforms-namely CASTOR, dCache, GPFS and Scalla/Xrootd. The tests have been performed at the CNAF Tier-1, the central computing facility of the Italian National Institute for Nuclear Research (INFN). Our storage back-end was based on Fibre Channel disk-servers organized in a Storage Area Network, being the disk-servers connected to the computing farm via Gigabit LAN. We used 24 disk-servers, 260 TB of raw-disk space and 280 worker nodes as computing clients, able to run concurrently up to about 1100 jobs. The aim of the test was to perform sequential and random read/write accesses to the data, as well as more realistic access patterns, in order to evaluate efficiency, availability, robustness and performance of the various data-access solutions.

Particle detection optimization by means of a distributed evolutionary algorithm

This paper presents an approach to optimize the reconstruction of particle tracks in a large area spectrometer using an evolutionary algorithm. First, the basic concepts of the measurement of the momentum are presented, including some details on the track reconstruction. An objective function is formulated to measure the efficiency of the reconstruction. Then, the optimizer is described in two versions: sequential for a single CPU and distributed for a cluster of PCs. The architecture of both of them is presented in detail, including the description of the operators used plus some words on the infrastructure required for the distributed version. Finally, the results from the tests of the program are presented, from which the conclusion is drawn that the distributed genetic algorithm, easy to implement on high energy physics data processing clusters, may be helpful in improving the efficiency of the reconstruction.

A framework for modelling and simulating data flows in distributed computing systems

This paper introduces a framework for modelling and simulating the behaviour of distributed computing systems. The framework comprises both means for abstract modelling and an implementation that allows the evaluation of abstract models by means of discrete event simulation. The basic characteristics of distributed computing systems used for physics data processing at CERN are presented, motivating the purpose and need for modelling and simulation. As an example, the central data recording system of a current High-Energy Physics experiment is introduced. The modelling framework provides concepts to structure relevant state information of a system; The most important factor to set up the structure of components encapsulating state information is thereby given by dynamic aspects of the system, i.e. by identified data flows. Traditional approaches to modelling often tend to focus purely on static, architectural system aspects. All abstract concepts are motivated and demonstrated using an example. Results of the simulation are compared to figures measured during real system operation, rating the proposed framework and its underlying modelling approach.