Recent trends in the HPC field have introduced new CPU architectures with improved vectorization capabilities that require optimization to achieve peak performance and thus pose challenges for performance portability. The deployment of high-performing scientific applications for CPUs requires adapting the codebase and optimizing for performance. Evaluating these applications provides insights into the complex interactions between code, compilers, and hardware. We evaluate compiler auto-vectorization and explicit vectorization to achieve performance portability across modern CPUs with long vectors. We select a molecular docking application as a case study, as it represents computational patterns commonly found across HPC workloads. We report insights into the technical challenges, architectural trends, and optimization strategies relevant to the future development of scientific applications for HPC. Our results show which code transformations enable portable auto-vectorization, reaching performance similar to explicit vectorization. Experimental data confirms that x86 CPUs typically achieve higher execution performance than ARM CPUs, primarily due to their wider vectorization units. However, ARM architectures demonstrate competitive energy consumption and cost-effectiveness.

High-Performance Computing (HPC) is a fundamental tool for tackling complex scientific and engineering problems. Optimizing applications for the heterogeneous and massively parallel nature of modern HPC hardware is essential for achieving timely and resource-efficient results. This paper introduces ABC-MLCO, a novel Artificial Bee Colony (ABC) algorithm enhanced with machine learning for constrained optimization in discrete configuration spaces, specifically targeting HPC scenarios with black-box performance metrics. ABC-MLCO extends the original ABC algorithm with mechanisms that define the feasible region, promote escape from local optima, prevent redundant evaluations, and intensify exploration and exploitation. We first evaluate the algorithm on benchmark functions, observing improvements in regret, feasibility rate, and convergence speed over the original ABC. Then, targeting a virtual screening application for drug discovery, ABC-MLCO outperforms well-known state-of-the-art methods in terms of final regret, achieving average improvements up to $93 \%$.

Drug discovery is a long and costly process aimed at finding a molecule that yields a therapeutic effect. Virtual screening is one of the initial in-silico steps that aims at estimating how promising a molecule is. This stage needs to solve two well-known domain problems: molecular docking and scoring. While the accuracy of scoring functions is extensively investigated in comparisons, the execution time of their implementation is usually not considered. In virtual screening campaigns, the definition of a fixed time budget for the entire process and the average time required to process each molecule determines the upper limit of the number of molecules that can be evaluated. By reducing the time needed to evaluate a single molecule, we can screen a larger number of molecules, thereby increasing the possibility of finding a promising solution. For extreme-scale virtual screening campaigns, the computational budget is a critical aspect since even utilizing large-scale facilities would make it impractical to complete the screening within a feasible time unless the computational time for a single molecule is significantly reduced. In this paper, we explore optimization and approximation techniques applied to two well-known scoring functions, which we modify to investigate different accuracy-performance trade-offs to support large-scale virtual screening campaigns. Despite the different approaches we considered, experimental results demonstrate that the proposed enhancements achieve better enrichment factors in virtual screening scenarios. Moreover, we port both implementations to CUDA to show that the proposed techniques are GPU-friendly and aligned with modern supercomputing infrastructures. • We optimized two SFs to explore the accuracy-throughput trade-off in extreme-scale VS. • We sped up X-SCORE 13 × via pre-computed approximations with a 10 % accuracy loss. • We sped up DrugScore ∼ 3 × via memoization, using saved time to improve scoring accuracy. • We ported SFs to CUDA to show how our methods are GPU-friendly and faster than on CPU. • We show that balancing accuracy-throughput can boost overall EF in large-scale VS.

Virtual screening applications are highly parameterized to optimize the balance between quality and execution performance. While output quality is critical, the entire screening process must be completed within a reasonable time. In fact, a slight reduction in output accuracy may be acceptable when dealing with large datasets. Finding the optimal quality-throughput trade-off depends on the specific HPC system used and should be re-evaluated with each new deployment or significant code update. This paper presents two parallel autotuning techniques for constrained optimization in distributed High-Performance Computing (HPC) environments. These techniques extend sequential Bayesian Optimization (BO) with two parallel asynchronous approaches, and they integrate predictions from Machine Learning (ML) models to help comply with constraints. Our target application is LiGen, a real-world virtual screening software for drug discovery. The proposed methods address two relevant challenges: efficient exploration of the parameter space and performance measurement using domain-specific metrics and procedures. We conduct an experimental campaign comparing the two methods with a popular state-of-the-art autotuner. Results show that our methods find configurations that are, on average, up to 35–42% better than the ones found by the autotuner and the default expert-picked LiGen configuration. • We propose two parallel algorithms for black-box constrained optimization. • We integrate Bayesian Optimization and Machine Learning for constraint estimation. • We use a meta-scheduler to hinge on the resources available in HPC settings. • We evaluate the benefits of the proposed approach using a relevant case study.

In recent years, there has been a growing interest in developing high-performance implementations of drug discovery processing software. To target modern GPU architectures, such applications are mostly written in proprietary languages such as CUDA or HIP. However, with the increasing heterogeneity of modern HPC systems and the availability of accelerators from multiple hardware vendors, it has become critical to be able to efficiently execute drug discovery pipelines on multiple large-scale computing systems, with the ultimate goal of working on urgent computing scenarios. This article presents the challenges of migrating LiGen, an industrial drug discovery software pipeline, from CUDA to the SYCL programming model, an industry standard based on C++ that enables heterogeneous computing. We perform a structured analysis of the performance portability of the SYCL LiGen platform, focusing on different aspects of the approach from different perspectives. First, we analyze the performance portability provided by the high-level semantics of SYCL, including the most recent group algorithms and subgroups of SYCL 2020. Second, we analyze how low-level aspects such as kernel occupancy and register pressure affect the performance portability of the overall application. The experimental evaluation is performed on two different versions of LiGen, implementing two different parallelization patterns, by comparing them with a manually optimized CUDA version, and by evaluating performance portability using both known and ad hoc metrics. The results show that, thanks to the combination of high-level SYCL semantics and some manual tuning, LiGen achieves native-comparable performance on NVIDIA, while also running on AMD GPUs.

Finding a novel drug is a very long and complex procedure. Using computer simulations, it is possible to accelerate the preliminary phases by performing a virtual screening that filters a large set of drug candidates to a manageable number. This paper presents the implementations and comparative analysis of two GPU-optimised implementations of a virtual screening algorithm targeting novel GPU architectures. This work focuses on the analysis of parallel computation patterns and their mapping onto the target architecture. The first method adopts a traditional approach that spreads the computation for a single molecule across the entire GPU. The second uses a novel batched approach that exploits the parallel architecture of the GPU to evaluate more molecules in parallel. Experimental results showed a different behavior depending on the size of the database to be screened, either reaching a performance plateau sooner or having a more extended initial transient period to achieve a higher throughput (up to 5x), which is more suitable for extreme-scale virtual screening campaigns.

Abstract Virtual screening is an early stage in the drug discovery process that selects the most promising candidates. In the urgent computing scenario, finding a solution in the shortest time frame is critical. Any improvement in the performance of a virtual screening application translates into an increase in the number of candidates evaluated, thereby raising the probability of finding a drug. In this paper, we show how we can improve application throughput using Out-of-kernel optimizations. They use input features, kernel requirements, and architectural features to rearrange the kernel inputs, executing them out of order, to improve the computation efficiency. These optimizations’ implementations are designed on an extreme-scale virtual screening application, named LiGen, that can hinge on CUDA and SYCL kernels to carry out the computation on modern supercomputer nodes. Even if they are tailored to a single application, they might also be of interest for applications that share a similar design pattern. The experimental results show how these optimizations can increase kernel performance by 2 $$\times$$ × , respectively, up to 2.2 $$\times$$ × in CUDA and up to 1.9 $$\times$$ × , in SYCL. Moreover, the reported speedup can be achieved with the best-proposed parameterization, as shown by the data we collected and reported in this manuscript.

Today digital revolution is having a dramatic impact on the pharmaceutical industry and the entire healthcare system. The implementation of machine learning, extreme-scale computer simulations, and big data analytics in the drug design and development process offers an excellent opportunity to lower the risk of investment and reduce the time to the patient.

Virtual screening is one of the early stages that aims to select a set of promising ligands from a vast chemical library. Molecular Docking is a crucial task in the process of drug discovery and it consists of the estimation of the position of a molecule inside the docking site. In the contest of urgent computing, we designed from scratch the EXSCALATE molecular docking platform to benefit from heterogeneous computation nodes and to avoid scaling issues. This poster presents the achievements and ongoing development of the EXSCALATE platform, together with an example of usage in the context of the COVID-19 pandemic.

Abstract Due to the increasing size of HPC machines, dealing with faults is becoming mandatory due to their high frequency. Natively, MPI cannot handle faults and it stops the execution prematurely when it finds one. With the introduction of ULFM, it is possible to continue the execution, but it requires complex integration with the application. In this paper we propose Legio, a framework that introduces fault resiliency in embarrassingly parallel MPI applications. Legio exposes its features to the application transparently, removing any integration difficulty. After a fault, the execution continues only with the non-failed processes. We also propose a hierarchical alternative, which features lower repair costs on large communicators. We evaluated our solutions on the Marconi100 cluster at CINECA with benchmarks and real-world applications, showing that the overhead introduced by the library is negligible and it does not limit the scalability properties of MPI.

Developing and optimizing software applications for high performance and energy efficiency is a very challenging task, even when considering a single target machine. For instance, optimizing for multicore-based computing systems requires in-depth knowledge about programming languages, application programming interfaces (APIs), compilers, performance tuning tools, and computer architecture and organization. Many of the tasks of performance engineering methodologies require manual efforts and the use of different tools not always part of an integrated toolchain. This paper presents Pegasus, a performance engineering approach supported by a framework that consists of a source-to-source compiler, controlled and guided by strategies programmed in a Domain-Specific Language, and an autotuner. Pegasus is a holistic and versatile approach spanning various decision layers composing the software stack, and exploiting the system capabilities and workloads effectively through the use of runtime autotuning. The Pegasus approach helps developers by automating tasks regarding the efficient implementation of software applications in multicore computing systems. These tasks focus on application analysis, profiling, code transformations, and the integration of runtime autotuning. Pegasus allows developers to program their strategies or to automatically apply existing strategies to software applications in order to ensure the compliance of non-functional requirements, such as performance and energy efficiency. We show how to apply Pegasus and demonstrate its applicability and effectiveness in a complex case study, which includes tasks from a smart navigation system.

The multiplication of large integers represents a significant computation effort in some cryptographic techniques. The use of dedicated hardware is an appealing solution to improve performances or efficiency. We propose a methodology to generate throughput oriented hardware accelerators for large integers multiplication leveraging High-Level Synthesis. The proposed micro-architectural template combines Karatsuba and Comba algorithms to control the extra-functional properties of the generated multiplier. The goal is to enable the end user to explore a wide range of possibilities, in terms of performance and resource utilization, without requiring them to know implementation and synthesis details. Experimental results show the large flexibility of the generated architectures and that the generated Pareto-set of multipliers can outperform some state-of-the-art RTL design.

Molecular Docking is a crucial task in the process of Drug Discovery. This task consists in the estimation of the position of a molecule inside the docking site. It is used in the early stages of the drug discovery process to perform a virtual screening of a large library of molecule candidates. This task is usually performed using High Performance Computing platforms, due to sheer number of candidates and due to complexity of the docking problem. In this work we ported and optimized a Molecular Docking Module to an heterogeneous system with one or more GPGPU accelerators, leveraging the directive languages OpenMP and OpenACC. We show that with the proposed approach, we are able to reach a better utilization of the available resources compared to the usual CPU/GPU data splitting, reaching a 25% throughput improvement within the single node.

The ANTAREX project developed an approach to the performance tuning of High Performance applications based on an Aspect-oriented Domain Specific Language (DSL), with the goal to simplify the enforcement of extra-functional properties in large scale applications. The project aims at demonstrating its tools and techniques on two relevant use cases, one in the domain of computational drug discovery, the other in the domain of online vehicle navigation. In this paper, we present an overview of the project and of its main achievements, as well as of the large scale experiments that have been planned to validate the approach.

In the autonomic computing context, the system is perceived as a set of autonomous elements capable of self-management, where end-users define high-level goals and the system shall adapt to achieve the desired behaviour. Runtime adaptation creates several optimization opportunities, especially if we consider approximate computing applications, where it is possible to trade off the accuracy of the result and the performance. Given that modern systems are limited by the power dissipated, autonomic computing is an appealing approach to increase the computation efficiency. In this paper, we introduce mARGOt, a dynamic autotuning framework to enhance the target application with an adaptation layer to provide self-optimization capabilities. The framework is implemented as a C++ library that works at function-level and provides to the application a mechanism to adapt in a reactive and a proactive way. Moreover, the application is capable to change dynamically its requirements and to learn online the underlying application-knowledge. We evaluated the proposed framework in three real-life scenarios, ranging from embedded to HPC applications. In the three use cases, experimental results demonstrate how, thanks to mARGOt, it is possible to increase the computation efficiency by adapting the application at runtime with a limited overhead.

Designing and optimizing applications for energy-efficient High Performance Computing systems up to the Exascale era is an extremely challenging problem. This paper presents the toolbox developed in the ANTAREX European project for autotuning and adaptivity in energy efficient HPC systems. In particular, the modules of the ANTAREX toolbox are described as well as some preliminary results of the application to two target use cases. 1

Configuring program parallelism and selecting optimal compiler options according to the underlying platform architecture is a difficult task. Tipically, this task is either assigned to the programmer or done by a standard one-fits-all policy generated by the compiler or runtime system. A runtime selection of the best configuration requires the insertion of a lot of glue code for profiling and runtime selection. This represents a programming wall for application developers. This paper presents a structured approach, called SOCRATES, based on an aspect-oriented language (LARA) and a runtime autotuner (mARGOt) to mitigate this problem. LARA has been used to hide the glue code insertion, thus separating the pure functional application description from extra-functional requirements. mARGOT has been used for the automatic selection of the best configuration according to the runtime evolution of the application.1

The ANTAREX project relies on a Domain Specific Language (DSL) based on Aspect Oriented Programming (AOP) concepts to allow applications to enforce extra functional properties such as energy-efficiency and performance and to optimize Quality of Service (QoS) in an adaptive way. The DSL approach allows the definition of energy-efficiency, performance, and adaptivity strategies as well as their enforcement at runtime through application autotuning and resource and power management. In this paper, we present an overview of the ANTAREX DSL and some of its capabilities through a number of examples, including how the DSL is applied in the context of one of the project use cases.

The increasing processing power of today's HW/SW platforms leads to the integration of more and more functions in a single device. Additional design challenges arise when these functions share computing resources and belong to different criticality levels. The paper presents the CONTREX European project and its preliminary results. CONTREX complements current activities in the area of predictable computing platforms and segregation mechanisms with techniques to consider the extra-functional properties, i.e., timing constraints, power, and temperature. CONTREX enables energy efficient and cost aware design through analysis and optimization of these properties with regard to application demands at different criticality levels.

The ever increasing number of processing units integrated on the same many-core chip delivers computational power that can exceed the performance requirements of a single application. The number of chips (and related power consumption) can thus be reduced to serve multiple applications - a practice which is called resource consolidation. However, this solution requires techniques to partition and assign resources among the applications and to manage unpredictable dynamic workloads.

To support adaptivity of data parallel applications on multi-core platforms, we propose a framework based on the combination of OpenCL application auto-tuning and run-time resource management. The framework addresses computationally intensive multimedia OpenCL applications. For these target applications, we show that application auto-tuning, based on design-time analysis, can become synergistic with run-time resource management. In the proposed framework, run-time decisions are taken by each application, autonomously, to achieve system adaptivity. This paper describes the methodology and related toolchain, defined during the 2PARMA European project, based on the integration of independent tools to provide effective compilation of OpenCL code, multi-objective design space exploration, application monitoring and tuning and system-wide run-time resource management. Experimental results are reported for design optimization of an OpenCL stereo-matching application and then for a resource contention scenario where multiple stereo-matching applications are executed on the same platform with different run-time requirements.