Our services

At Exeliz Solutions we have been advising supercomputing centers, universities, public institutions and industrial companies since 2011 on the optimization of their HPC infrastructure. Our consulting approach combines a deep understanding of high-performance hardware with comprehensive knowledge of the real needs of end users.

We carry out complete audits of your current infrastructure, identifying bottlenecks, operational inefficiencies and improvement opportunities. We analyze your workloads to understand usage patterns, software dependencies, memory requirements and I/O patterns. From this data, we formulate architecture recommendations aligned with your technical and budgetary objectives.

Our team offers ongoing consulting for operational HPC systems, including optimization sessions, resolution of inefficiencies and strategic advice for planned upgrades. We also advise our clients on the best deployment options, whether in Cloud or Bare Metal environments, always with an agnostic approach that prioritizes customer value.

With over 100 clusters installed and 4,000 users trained, our team brings unique experience spanning from initial audit through implementation and ongoing support. This dual expertise in IT infrastructure and end-user needs allows us to deliver a truly comprehensive consulting service.

• › Full audit of current HPC infrastructure
• › Detailed workload analysis
• › Architecture recommendations aligned with objectives
• › Consulting for operational HPC systems (optimization, inefficiency resolution)
• › Guidance on execution in remote environments (Cloud or Bare Metal)
• › Dual expertise in IT infrastructure and end-user needs

Every project we tackle at Exeliz Solutions begins with a rigorous analysis of what your team truly needs. Before proposing any architecture, we study the code to be executed, run benchmarks with representative workloads, and conduct a market study to identify the most cost-effective solution for each specific case.

Our approach is completely agnostic regarding manufacturers and technologies. We are not aligned with any vendor, which allows us to act as impartial consultants in both public tenders and private procurement. This independence ensures that every recommendation is based solely on the technical and economic requirements of your project.

We design HPC infrastructure from scratch for new installations and offer improvement consulting for systems already in operation. Each essential component is meticulously arranged to maximize system efficiency: from hardware integration to software optimizations. Precision and attention to detail accompany every phase of the process.

This rigorous methodology has enabled us to participate in TOP500 projects, including configuring supercomputers with over 16,000 cores and recording the highest Lustre storage performance benchmarks in Spain. Our experience translates into solutions that meet and exceed expectations.

• › Hardware and software agnostic approach — no manufacturer alignment
• › Impartial consultant for public tenders and private procurement
• › Meticulous code analysis prior to hardware solution design
• › Representative benchmarks across multiple configurations
• › Market study to present the most cost-effective solution
• › Comprehensive requirements matching for optimal value

At Exeliz Solutions, our comprehensive "Installation" service embodies an exhaustive approach towards overseeing the entire lifecycle of High-Performance Computing (HPC) systems. By seamlessly orchestrating hardware assembly, software integration, and the finalization of applications, this service ensures a holistic installation process. Our dedication goes beyond just the installation phase. We curate personalized training sessions to equip your team with the necessary skills in system administration and optimal resource utilization. These tailored training sessions are designed to empower your team, fostering efficient operation and utilization of your HPC infrastructure, thereby maximizing its potential. Why Is Installation Important? Installation serves as the fundamental cornerstone in the realm of High-Performance Computing (HPC), playing a pivotal role in the functionality and performance of these intricate systems. A meticulously orchestrated installation process ensures the seamless integration of hardware, software, and applications, creating a robust infrastructure that fosters peak efficiency. The process goes beyond mere setup, encompassing detailed configurations, thorough checks, and optimized utilization of resources. Each phase, from hardware assembly to software integration, demands precision to ensure the system's readiness for execution. The success of an HPC system heavily relies on the effectiveness of its installation. A well-executed installation not only sets the stage for future operations but also paves the way for optimal performance, minimized errors, and the longevity of the system's lifecycle. In essence, the significance of installation lies in its ability to lay a solid foundation, ensuring smooth, effective, and future-ready operations for the entire system.

• › Complete HPC cluster installation
• › Hardware provisioning, racking and cabling
• › Network infrastructure configuration (Infiniband, Ethernet)
• › Batch scheduler configuration (Slurm, LSF, PBS, Torque, SGE)
• › Integration of hybrid architectures with GPUs and coprocessors
• › Future-ready computing environments

At Exeliz Solutions we cover the entire HPC software landscape, from the basic operating system deployment on nodes to the production release of complex environments with dozens of scientific applications and engineering tools. Our track record includes more than 100 scientific and engineering applications already successfully deployed across various HPC architectures.

We specialize in installing and optimizing essential middleware: parallel file systems such as BeeGFS, Lustre and GPFS; compilers from the major manufacturers (GNU, Intel, PGI, IBM XL); development and analysis tools such as V-Tune, DDT and GDB; and module and environment management systems such as EasyBuild and Spack. Each component is compiled and configured with optimal parameters for the target architecture.

We work closely with certified integrators like Bright Computing to deliver complete and coherent cluster management solutions. Our deployment process includes comprehensive compatibility verification, compilation parameter optimization and advanced configuration of profiling and debugging tools.

The result is a fully integrated, optimized and documented software environment, where every tool is perfectly aligned with the specific workloads of your researchers and engineers. Our users can focus on their scientific work without worrying about the infrastructure that supports it.

• › Operating system installation on node(s)
• › End-user application deployment
• › Over 100 scientific and engineering applications available
• › Middleware: parallel file systems (BeeGFS, Lustre, GPFS)
• › Development, debugging and profiling tools
• › Compilers: GNU, Intel, PGI, IBM XL

Our training courses are meticulously designed to encompass all facets required for effective administration and optimal utilization of HPC resources. The courses are structured into various modules, offering flexible configurations based on individual client needs. This adaptability allows our clients to tailor a custom training course in collaboration with our team. These training sessions are available both online and in-person, providing the flexibility to accommodate diverse preferences and logistical arrangements. Furthermore, upon completion of the training, we provide access to a complete test virtual environment, allowing practical application and further learning. Our aim is to provide an all-encompassing training experience, ensuring participants gain comprehensive expertise in HPC resource management and administration.

• › Modular online and in-person courses
• › HPC system administration training
• › Optimal resource utilization and exploitation training
• › Full virtual test environment access after training
• › Training for fully self-sufficient internal teams

Maintenance and technical support for computing systems are foundational throughout their lifecycle. At our organization, we provide robust and tailored technical support services specifically designed to manage diverse HPC architectures, spanning from SMP systems to Infiniband clusters, and hybrid architectures encompassing GPUs or co-processors. Our adept team collaborates closely with system administrators, offering optimal configuration guidance for all cluster management tools. In addition, we promptly address and resolve technical issues on-site with precision and efficiency.

• › Support hour packs (minimum 10h, 3-month validity)
• › Service level agreements (SLAs)
• › Support for SMP architectures, Infiniband clusters and hybrid GPU setups
• › Parallel file system management (Lustre, BeeGFS, GPFS)
• › Direct collaboration with system administrators
• › Support for users with scientific backgrounds in simulation

To effectively oversee and manage HPC environments, we offer remote administration managed services, ensuring secure and auditable practices. Our suite of services includes the complete or partial outsourcing of a company's infrastructure and personnel when necessary. For instance, in scenarios demanding intensive infrastructure usage to allocate necessary resources across various business areas. Our HPC managed services seamlessly integrate personnel, operational processes, and essential technology for efficient operation. This integration is facilitated remotely using monitoring software that manages functions, connects with requisite servers, and services.

• › Full or partial remote administration
• › Hardware monitoring and security
• › Backups and user/group management
• › Software installations and updates
• › Proactive maintenance throughout the system lifecycle
• › Rapid incident resolution