Abstract Analysis on HEP data is an iterative process in which the results of one step often inform the next. In an exploratory analysis, it is common to perform one computation on a collection of events, then view the results (often with histograms) to decide what to try next. Awkward Array is a Scikit-HEP Python package that enables data analysis with array-at-a-time operations to implement cuts as slices, combinatorics as composable functions, etc. However, most C++ HEP libraries, such as FastJet, have an imperative, one-particle-at-a-time interface, which would be inefficient in Python and goes against the grain of the array-at-a-time logic of scientific Python. Therefore, we developed fastjet, a pip-installable Python package that provides FastJet C++ binaries, the classic (particle-at-a-time) Python interface, and the new array-oriented interface for use with Awkward Array. The new interface streamlines interoperability with scientific Python software beyond HEP, such as machine learning. In one case, adopting this library along with other array-oriented tools accelerated HEP analysis code by a factor of 20. It was designed to be easily integrated with libraries in the Scikit-HEP ecosystem, including Uproot (file I/O), hist (histogramming), Vector (Lorentz vectors), and Coffea (high-level glue). We discuss the design of the fastjet Python library, integrating the classic interface with the array oriented interface and with the Vector library for Lorentz vector operations. The new interface was developed as open source.
more »
« less
hep_tables: Heterogeneous Array Programming for HEP
Array operations are one of the most concise ways of expressing common filtering and simple aggregation operations that are the hallmark of a particle physics analysis: selection, filtering, basic vector operations, and filling histograms. The High Luminosity run of the Large Hadron Collider (HL-LHC), scheduled to start in 2026, will require physicists to regularly skim datasets that are over a PB in size, and repeatedly run over datasets that are 100’s of TB’s – too big to fit in memory. Declarative programming techniques are a way of separating the intent of the physicist from the mechanics of finding the data and using distributed computing to process and make histograms. This paper describes a library that implements a declarative distributed framework based on array programming. This prototype library provides a framework for different sub-systems to cooperate in producing plots via plug-in’s. This prototype has a ServiceX data-delivery sub-system and an awkward array sub-system cooperating to generate requested data or plots. The ServiceX system runs against ATLAS xAOD data and flat ROOT TTree’s and awkward on the columnar data produced by ServiceX.
more »
« less
- Award ID(s):
- 1836650
- NSF-PAR ID:
- 10354363
- Editor(s):
- Biscarat, C.; Campana, S.; Hegner, B.; Roiser, S.; Rovelli, C.I.; Stewart, G.A.
- Date Published:
- Journal Name:
- EPJ Web of Conferences
- Volume:
- 251
- ISSN:
- 2100-014X
- Page Range / eLocation ID:
- 03061
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
more » « less
Summary One of the activities of the Pacific Rim Applications and Grid Middleware Assembly (PRAGMA) is fostering Virtual Biodiversity Expeditions by bringing domain scientists and cyber infrastructure specialists together as a team. Over the past few years, PRAGMA members have been collaborating on virtualizing the Lifemapper software. Virtualization and cloud computing have introduced great flexibility and efficiency into IT projects. Virtualization refers to the technologies that provide a layer of abstraction between server hardware system and software that runs on it. This abstraction enables a logical view of computing resources and allows multiple servers to run on the same hardware. With this project, we are virtualizing Lifemapper by enabling its installation and configuration on a virtual cluster. Virtualization provides application scalability, maximizes resources utilization, and creates a more efficient, agile, and automated infrastructure. However, there are downsides to the complexity inherent in these environments, including the need for special techniques to deploy cluster hosts, dependence on virtual environments, and challenging application installation, management, and configuration. In this study, we report on progress of the Lifemapper virtualization framework focused on a reproducible and highly configurable infrastructure capable of fast deployment.
Lifemapper is a distributed software application developed by the Biodiversity Institute at The University of Kansas. Lifemapper creates and maintains a publicly accessible archive of species distribution maps calculated from public specimen data. Lifemapper software also provides a suite of tools for biodiversity researchers that calculate single and multispecies distribution predictions and macroecological analyses through application programming interfaces. Our goal is to create a viable solution that can be easily adopted and reused by scientists from multiple institutions or projects. This solution (1) allows fast deployment of ready‐made cluster images, (2) reproduces the complete Lifemapper processing pipeline on demand at multiple sites and in different hosting environments, and (3) enables scientists to perform Lifemapper‐facilitated data processing on restricted‐use data, very large datasets, or other unique data.
A key contribution of this work is describing the practical experience in taking a complex, clustered, domain‐specific, data analysis, and simulation system and enabling its operation on a variety of system configurations. Uses of this portability range from whole cluster replication to teaching and experimentation on a single laptop. System virtualization is used to practically define and make portable the full application stack, including all of its complex set of supporting software and allows Lifemapper deployment in a variety of environments.
-
Access libraries such as ROOT[1] and HDF5[2] allow users to interact with datasets using high level abstractions, like coordinate systems and associated slicing operations. Unfortunately, the implementations of access libraries are based on outdated assumptions about storage systems interfaces and are generally unable to fully benefit from modern fast storage devices. For example, access libraries often implement buffering and data layout that assume that large, single-threaded sequential access patterns are causing less overall latency than small parallel random access: while this is true for spinning media, it is not true for flash media. The situation is getting worse with rapidly evolving storage devices such as non-volatile memory and ever larger datasets. This project explores distributed dataset mapping infrastructures that can integrate and scale out existing access libraries using Ceph’s extensible object model, avoiding re-implementation or even modifications of these access libraries as much as possible. These programmable storage extensions coupled with our distributed dataset mapping techniques enable: 1) access library operations to be offloaded to storage system servers, 2) the independent evolution of access libraries and storage systems and 3) fully leveraging of the existing load balancing, elasticity, and failure management of distributed storage systems like Ceph. They also create more opportunities to conduct storage server-local optimizations specific to storage servers. For example, storage servers might include local key/value stores combined with chunk stores that require different optimizations than a local file system. As storage servers evolve to support new storage devices like non-volatile memory, these server-local optimizations can be implemented while minimizing disruptions to applications. We will report progress on the means by which distributed dataset mapping can be abstracted over particular access libraries, including access libraries for ROOT data, and how we address some of the challenges revolving around data partitioning and composability of access operations.more » « less
-
Experience shows that on today's high performance systems the utilization of different acceleration cards in conjunction with a high utilization of all other parts of the system is difficult. Future architectures, like exascale clusters, are expected to aggravate this issue as the number of cores are expected to increase and memory hierarchies are expected to become deeper. One big aspect for distributed applications is to guarantee high utilization of all available resources, including local or remote acceleration cards on a cluster while fully using all the available CPU resources and the integration of the GPU work into the overall programming model. For the integration of CUDA code we extended HPX, a general purpose C++ run time system for parallel and distributed applications of any scale, and enabled asynchronous data transfers from and to the GPU device and the asynchronous invocation of CUDA kernels on this data. Both operations are well integrated into the general programming model of HPX which allows to seamlessly overlap any GPU operation with work on the main cores. Any user defined CUDA kernel can be launched on any (local or remote) GPU device available to the distributed application. We present asynchronous implementations for the data transfers and kernel launches for CUDA code as part of a HPX asynchronous execution graph. Using this approach we can combine all remotely and locally available acceleration cards on a cluster to utilize its full performance capabilities. Overhead measurements show, that the integration of the asynchronous operations (data transfer + launches of the kernels) as part of the HPX execution graph imposes no additional computational overhead and significantly eases orchestrating coordinated and concurrent work on the main cores and the used GPU devices.more » « less
-
Biscarat, C. ; Campana, S. ; Hegner, B. ; Roiser, S. ; Rovelli, C.I. ; Stewart, G.A. (Ed.)The cabinetry library provides a Python-based solution for building and steering binned template fits. It tightly integrates with the pythonic High Energy Physics ecosystem, and in particular with pyhf for statistical inference. cabinetry uses a declarative approach for building statistical models, with a JSON schema describing possible configuration choices. Model building instructions can additionally be provided via custom code, which is automatically executed when applicable at key steps of the workflow. The library implements interfaces for performing maximum likelihood fitting, upper parameter limit determination, and discovery significance calculation. cabinetry also provides a range of utilities to study and disseminate fit results. These include visualizations of the fit model and data, visualizations of template histograms and fit results, ranking of nuisance parameters by their impact, a goodness-of-fit calculation, and likelihood scans. The library takes a modular approach, allowing users to include some or all of its functionality in their workflow.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1051/epjconf/202125103061
