The Elastic Analysis Facility (EAF) hosted at Fermi National Accelerator Laboratory (Fermilab) is a platform being developed with the goal of providing a fast and efficient facility for physics analysis. As high-energy physics moves towards collecting larger datasets, such as those from the High-Luminosity LHC, the EAF strives to provide a powerful and adaptable framework for future colliders and multi-experiment institutions. Currently, the EAF supports several experiments including CMS, NOvA, and DUNE as well as serving accelerator physicists and beam line operations through integrated software and secure connections to Fermilab's computing resources. In addition, the EAF was designed with a user-friendly interface, intended to be more intuitive for emerging generations of physicists, that is still accessible for established styles of analysis. The EAF can also achieve better analysis efficiency due to the modernization of software and tools that can better utilize Fermilab's computing power. Furthermore, its design incorporates industry standards whenever possible, enhancing its sustainability and making it a possible template for other national or international laboratories and research facilities. Overall, the EAF is a forward-looking solution that will meet the evolving needs of particle physics, ensuring readiness for future colliders and multi-experiment research institutions.

As the Large Hadron Collider (LHC) transitions into the High-Luminosity LHC (HL-LHC) era, the volume of data to be processed is expected to increase significantly. The CMS Experiment currently utilizes various data formats, including AOD, MiniAOD, and NanoAOD, each with different levels of detail and storage requirements. This paper addresses the challenges of data duplication and storage inefficiencies in high-energy physics (HEP) analyses by proposing an on-demand column-joining solution. This approach aims to reduce data duplication by enabling the dynamic combination of NanoAOD data with auxiliary information from larger data tiers, such as MiniAOD. The proposed solution leverages Trino, a high-performance distributed SQL query engine, to perform efficient and scalable data joins. Benchmarks using CMS OpenData demonstrate the feasibility of this approach, showing that it can handle large datasets with low latency. Integration with the scikit-hep ecosystem and the coffea analysis framework is also discussed, highlighting the potential for seamless end-to-end data processing and analysis. Ongoing and future work focuses on expanding benchmarks, integrating ServiceX for data transformation, and exploring the use of native object storage solutions.

With machine learning applications now spanning a variety of computational tasks, multi-user shared computing facilities are devoting a rapidly increasing proportion of their resources to such algorithms. Graph neural networks (GNNs), for example, have provided astounding improvements in extracting complex signatures from data and are now widely used in a variety of applications, such as particle jet classification in high energy physics (HEP). However, GNNs also come with an enormous computational penalty that requires the use of GPUs to maintain reasonable throughput. At shared computing facilities, such as those used by physicists at Fermi National Accelerator Laboratory (Fermilab), methodical resource allocation and high throughput at the many-user scale are key to ensuring that resources are being used as efficiently as possible. These facilities, however, primarily provide CPU-only nodes, which proves detrimental to time-to-insight and computational throughput for workflows that include machine learning inference. In this work, we describe how a shared computing facility can use the NVIDIA Triton Inference Server to optimize its resource allocation and computing structure, recovering high throughput while scaling out to multiple users by massively parallelizing their machine learning inference. To demonstrate the effectiveness of this system in a realistic multi-user environment, we use the Fermilab Elastic Analysis Facility augmented with the Triton Inference Server to provide scalable and high throughput access to a HEP-specific GNN and report on the outcome.

This white paper presents the current status of the R&D for Analysis Facilities (AFs) and attempts to summarize the views on the future direction of these facilities. These views have been collected through the High Energy Physics (HEP) Software Foundation's (HSF) Analysis Facilities forum, established in March 2022, the Analysis Ecosystems II workshop, that took place in May 2022, and the WLCG/HSF pre-CHEP workshop, that took place in May 2023. The paper attempts to cover all the aspects of an analysis facility.

The HL-LHC run is anticipated to start at the end of this decade and will pose a significant challenge for the scale of the HEP software and computing infrastructure. The mission of the U.S. CMS Software & Computing Operations Program is to develop and operate the software and computing resources necessary to process CMS data expeditiously and to enable U.S. physicists to fully participate in the physics of CMS. We have developed a strategic plan to prioritize R&D efforts to reach this goal for the HL-LHC. This plan includes four grand challenges: modernizing physics software and improving algorithms, building infrastructure for exabyte-scale datasets, transforming the scientific data analysis process and transitioning from R&D to operations. We are involved in a variety of R&D projects that fall within these grand challenges. In this talk, we will introduce our four grand challenges and outline the R&D program of the U.S. CMS Software & Computing Operations Program.

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 tools, we attempt to reprocess the data by running several thousand concurrent grid jobs, a rate we expect to be typical of current and future neutrino physics experiments. We process most of the dataset with the GPU version of our processing algorithm and the remainder with the CPU version for timing comparisons. We find that a 100-GPU cloud-based server is able to easily meet the processing demand, and that using the GPU version of the event processing algorithm is two times faster than processing these data with the CPU version when comparing to the newest CPUs in our sample. The amount of data transferred to the inference server during the GPU runs can overwhelm even the highest-bandwidth network switches, however, unless care is taken to observe network facility limits or otherwise distribute the jobs to multiple sites. We discuss the lessons learned from this processing campaign and several avenues for future improvements.

The field of transient astronomy has seen a revolution with the first gravitational-wave detections and the arrival of multi-messenger observations they enabled. Transformed by the first detection of binary black hole and binary neutron star mergers, computational demands in gravitational-wave astronomy are expected to grow by at least a factor of two over the next five years as the global network of kilometer-scale interferometers are brought to design sensitivity. With the increase in detector sensitivity, real-time delivery of gravitational-wave alerts will become increasingly important as an enabler of multi-messenger followup. In this work, we report a novel implementation and deployment of deep learning inference for real-time gravitational-wave data denoising and astrophysical source identification. This is accomplished using a generic Inference-as-a-Service model that is capable of adapting to the future needs of gravitational-wave data analysis. Our implementation allows seamless incorporation of hardware accelerators and also enables the use of commercial or private (dedicated) as-a-service computing. Based on our results, we propose a paradigm shift in low-latency and offline computing in gravitational-wave astronomy. Such a shift can address key challenges in peak-usage, scalability and reliability, and provide a data analysis platform particularly optimized for deep learning applications. The achieved sub-millisecond scale latency will also be relevant for any machine learning-based real-time control systems that may be invoked in the operation of near-future and next generation ground-based laser interferometers, as well as the front-end collection, distribution and processing of data from such instruments.

Machine learning (ML) is becoming an increasingly important component of cutting-edge physics research, but its computational requirements present significant challenges. In this white paper, we discuss the needs of the physics community regarding ML across latency and throughput regimes, the tools and resources that offer the possibility of addressing these needs, and how these can be best utilized and accessed in the coming years.

Prior to the public release of Kubernetes it was difficult to conduct joint development of elaborate analysis facilities due to the highly non-homogeneous nature of hardware and network topology across compute facilities. However, since the advent of systems like Kubernetes and OpenShift, which provide declarative interfaces for building fault-tolerant and self-healing deployments of networked software, it is possible for multiple institutes to collaborate more effectively since resource details are abstracted away through various forms of hardware and software virtualization. In this whitepaper we will outline the development of two analysis facilities: "Coffea-casa" at University of Nebraska Lincoln and the "Elastic Analysis Facility" at Fermilab, and how utilizing platform abstraction has improved the development of common software for each of these facilities, and future development plans made possible by this methodology.

The HL-LHC presents significant challenges for the HEP analysis community. The number of events in each analysis is expected to increase by an order of magnitude and new techniques are expected to be required; both challenges necessitate new services and approaches for analysis facilities. These services are expected to provide new capabilities, a larger scale, and different access modalities (complementing -- but distinct from -- traditional batch-oriented approaches). To facilitate this transition, the US-LHC community is actively investing in analysis facilities to provide a testbed for those developing new analysis systems and to demonstrate new techniques for service delivery. This whitepaper outlines the existing activities within the US LHC community in this R&D area, the short- to medium-term goals, and the outline of common goals and milestones.

The File Transfer Service (FTS3) is a data movement service developed at CERN which is used to distribute the majority of the Large Hadron Collider’s data across the Worldwide LHC Computing Grid (WLCG) infrastructure. At Fermilab, we have deployed FTS3 instances for Intensity Frontier experiments (e.g. DUNE) to transfer data in America and Europe, using a container-based strategy. In this article we summarize our experience building docker images based on work from the SLATE project (slateci.io) and deployed in OKD, the community distribution of Red Hat OpenShift. Additionally, we discuss our method of certificate management and maintenance utilizing Kubernetes CronJobs. Finally, we also report on the configuration currently running at Fermilab.

Simulating quantum field theories on a quantum computer is one of the most exciting fundamental physics applications of quantum information science. Dynamical time evolution of quantum fields is a challenge that is beyond the capabilities of classical computing, but it can teach us important lessons about the fundamental fabric of space and time. Whether we may answer scientific questions of interest using near-term quantum computing hardware is an open question that requires a detailed simulation study of quantum noise. Here we present a large scale simulation study powered by a multi-node implementation of qsim using the Google Cloud Platform. We additionally employ newly-developed GPU capabilities in qsim and show how Tensor Processing Units — Application-specific Integrated Circuits (ASICs) specialized for Machine Learning — may be used to dramatically speed up the simulation of large quantum circuits. We demonstrate the use of high performance cloud computing for simulating ℤ 2 quantum field theories on system sizes up to 36 qubits. We find this lattice size is not able to simulate our problem and observable combination with sufficient accuracy, implying more challenging observables of interest for this theory are likely beyond the reach of classical computation using exact circuit simulation.

In the next decade, the demands for computing in large scientific experiments are expected to grow tremendously. During the same time period, CPU performance increases will be limited. At the CERN Large Hadron Collider (LHC), these two issues will confront one another as the collider is upgraded for high luminosity running. Alternative processors such as graphics processing units (GPUs) can resolve this confrontation provided that algorithms can be sufficiently accelerated. In many cases, algorithmic speedups are found to be largest through the adoption of deep learning algorithms. We present a comprehensive exploration of the use of GPU-based hardware acceleration for deep learning inference within the data reconstruction workflow of high energy physics. We present several realistic examples and discuss a strategy for the seamless integration of coprocessors so that the LHC can maintain, if not exceed, its current performance throughout its running.

Machine learning algorithms are becoming increasingly prevalent and performant in the reconstruction of events in accelerator-based neutrino experiments. These sophisticated algorithms can be computationally expensive. At the same time, the data volumes of such experiments are rapidly increasing. The demand to process billions of neutrino events with many machine learning algorithm inferences creates a computing challenge. We explore a computing model in which heterogeneous computing with GPU coprocessors is made available as a web service. The coprocessors can be efficiently and elastically deployed to provide the right amount of computing for a given processing task. With our approach, Services for Optimized Network Inference on Coprocessors (SONIC), we integrate GPU acceleration specifically for the ProtoDUNE-SP reconstruction chain without disrupting the native computing workflow. With our integrated framework, we accelerate the most time-consuming task, track and particle shower hit identification, by a factor of 17. This results in a factor of 2.7 reduction in the total processing time when compared with CPU-only production. For this particular task, only 1 GPU is required for every 68 CPU threads, providing a cost-effective solution.

Computing needs for high energy physics are already intensive and are expected to increase drastically in the coming years. In this context, heterogeneous computing, specifically as-a-service computing, has the potential for significant gains over traditional computing models. Although previous studies and packages in the field of heterogeneous computing have focused on GPUs as accelerators, FPGAs are an extremely promising option as well. A series of workflows are developed to establish the performance capabilities of FPGAs as a service. Multiple different devices and a range of algorithms for use in high energy physics are studied. For a small, dense network, the throughput can be improved by an order of magnitude with respect to GPUs as a service. For large convolutional networks, the throughput is found to be comparable to GPUs as a service. This work represents the first open-source FPGAs-as-a-service toolkit.

In this letter, we comment on the “last mile” for computing in support of physics analyses; this does not cover simulation/acquisition, batch processing/production/reconstruction, and data reduction. Specifically, we focus on analysis facilities, which provide the foundation for the individual researcher to experiment with and understand their data.

The world currently spends more than $30B per quarter on the consumption of Cloud Computing services. This is 17 times the size of the entire FY20 budget for the Office of Science at the Department of Energy. These resources have been successfully used for scientific computing in HEP and elsewhere under a pay-as-you-go model where users are billed monthly based on the resources they have consumed. There are a wide range of Cloud services, but we categorize them into “capability” and “capacity”. Capability services represent a unique set of features that we have not provisioned on-premises for a variety of reasons (cost-effectiveness, power consumption, proprietary solutions, etc.) Capacity services are services that allow us to scale out commodity services; historically we have focused on high-throughput (batch) computing.

Machine learning (ML) methods have been actively explored as novel solutions for simulation, reconstruction and analysis in fundamental particle physics research. Accelerated ML inference offers numerous opportunities to leverage these advanced algorithms to enhance the instrumentation and computing capabilities of particle physics experiments, and provides a computing ethos for diverse fields where low-latency inference is required. Low-latency applications realized on FPGAs, GPUs, and ASICs can improve trigger performance and minimize data loss. In bringing fast ML as close as possible to sensor data, we can drastically improve the data rate which can be streamed off the detector. In addition, the use of heterogeneous computing resources to accelerate offline reconstruction and simulation may serve as an elegant solution to computing challenges faced by many particle physics experiments. Furthermore, fast ML may improve experimental operation and accelerator and detector control without drastic overhauls in the design. In this letter of interest, we summarize explorations by the community, and highlight prospects in future developments and applications.