preloader

Explore our High Performance Computing (HPC) glossary. A resource designed to provide short but clear explanations of the specialized terms and concepts within the field.

 

TermTranslation (eng-ger)AbbreviationExplanation
AlgorithmAlgorithmus-A sequence of steps to solve a problem or perform a task in computing.
Application Programming InterfaceProgrammierschnittstelleAPIA set of functions allowing applications to access the features or data of something offering the API, like an operating system, application, or other service.
Artificial IntelligenceKünstliche IntelligenzAIVaguely defined term, ususally refering to a set of techniques making software more similar to human intelligence, e.g. by learning from large amounts of data.
BandwidthBandbreite-Rate of data transfer in Bytes/second.
ByteByteBA data unit containing 8 bits, often used to represent a single character of data.
Central Processing UnitProzessorCPUThe primary component of a computer that performs most of the processing.
ClusterCluster-An accumulation, e.g. of points in a data set, or of computers connected together to work as a single system.
DatabaseDatenbankDBA structured collection of data, efficiently organized for secure and safe permanent storage and management of large amounts of data.
EncryptionVerschlüsselung-The process of converting readable plaintext information into a ciphertext to prevent unauthorized access to the plaintext.
FirewallFirewall-A system designed to block unwanted network connections. For example, a firewall on a WiFi router at your home might reject an incoming external connection to your printer.
Graphics Processing UnitGrafikkarte / RechenbeschleunigerGPU
Grid ComputingLoses Cluster zum verteilten Rechnen-A form of distributed computing where loosely coupled, unused computer resources are harnessed to solve complex problems. For example, a collection of laptops at home in many different countries.
High Performance ComputingHochleistungsrechnenHPCUse of supercomputers and parallel processing techniques for solving advanced large-scale computational problems.
InfiniBandInfiniBandIBA low latency and high bandwith communication network used in HPC environments.
input/outputEingabe und AusgabeI/OCommunication between a system or component and its surroundings.
Internet ProtocolInternetprotokollIPA communications protocol for sending data packets. Can be used between different processes, computers or networks.
JavaJava-A popular, high-level, object-oriented programming language.
KernelKernel-The core component of an operating system that manages system operations. Sometimes also used to refer to the core component of a convolutional neural network, filtering algorithm, or other computing method.
LatencyLatenz-One-way network latency is the time delay between the sender sending a message and the receiver receiving it. Round-trip network latency is the time delay between sending a message and receiving a response.
Learning NuggetsLernhäppchen-Small, focused units or chunks of learning, usually designed for easy consumption
LibraryProgrammbibliothek-A collection of pieces of code that can be reused in a program. The library is often pre-compiled to a shared object.
Linear AlgebraLineare Algebra-A branch of mathematics dealing with vectors, matrices, and linear transformations.
Linpack BenchmarkLinpack Benchmark-A measure of a computer's maximum floating-point rate of execution, often used in the HPC world.
LinuxLinux-An open-source Unix-like operating system, based on the Linux kernel.
Liquid CoolingFlüssigkühlung-A method to dissipate heat in computers using liquid coolants rather than air.
Load BalancingLastausgleich-Distributing computational workloads uniformly among parallel processes or nodes of a cluster. This is done to avoid overloading and to ensure optimal resource utilization.
LustreLustre-A type of distributed parallel file system commonly used in HPC environments.
MemoryArbeitspeicher-Physical devices used to store applications and data in a form that can be immediately used by a computer. Often synonymous to RAM.
Message Passing Interface(Eine Kommunikationsbibliothek)MPIA standardized communication protocol for programming parallel computers.
Multi-core ProcessorMehrkernprozessor-A processor with multiple cores which can run multiple tasks simultaneously.
MultithreadingNebenläufigkeit-A technique where parts of a process, called 'threads', can be executed in parallel.
NetworkNetzwerk-Set of connected computers.
NodeKnoten-A single element of a cluster, this means a computer in a larger HPC system.
Non-uniform Memory AccessNicht-uniformer SpeicherzugriffNUMAA memory layout in a multiprocessor system. Due to the non-uniform layout, each processor can access some parts of the memory at lower latency and/or higher bandwith.
NVIDIA Data Center GPU ManagerNVIDIA Data Center GPU ManagerDCGMNVIDIA Data Center GPU Manager (DCGM) is a suite of tools for managing and monitoring NVIDIA datacenter GPUs in cluster environments.
Object-Oriented ProgrammingObjektorientierte ProgrammierungOOPA programming paradigm used to structure both code and data. It based on instances of classes called 'objects'.
Open Multi-Processing(Eine Shared-Memory-API)OpenMPAn API that supports shared memory parallel programming.
Parallel ProcessingParallele Verabreitung-Using multithreading or multiple processors to execute tasks simultaneously.
Peta Floating Point Operations per SecondPeta Fließkommaoperationen pro SekundePetaflop/sA measure of computing speed, equivalent to one thousand trillion (10^15) floating-point operations per second.
ProcessorProzessor-A component in a computer that executes instructions. This includes performing arithmetic, logical, controlling and I/O operations.
PythonPython-A popular, multi-purpose, programming language.
Quantum ComputingQuantencomputing-A type of computing based on quantum-mechanical phenomena like superposition and entanglement.
QueryAbfrage-A request for data retrieval from a database.
Random Access MemoryDirektzugriffsspeicherRAMA form of computer memory that can be accessed in any order. This means that read and write access times do not depend on the location of data in the memory.
Redundant Array of Independent DisksRedundante Anordnung unabhängiger FestplattenRAIDA data storage technology that uses multiple drives for redundancy or increased performance.
Running Average Power LimitLaufende Durchschnittliche LeistungsbegrenzungRAPLA power management and monitoring interface on Intel and AMD CPUs.
ScalabilitySkalierbarkeit-Ability of a system to handle increased load by adding more resources. Typically limited, e.g. due to communication overheads. For example, two identical processors might be just 1.5 times faster than a single processor.
SchedulerSteuerprogramm-A tool or system that manages execution of tasks, such as processes on an operating system or compute jobs on HPC clusters. Typically allocates resources in HPC systems, especially for job queues.
Self-Organized LearningSelbstorganisiertes LernenSOLA process where individuals or groups take initiative in diagnosing their learning needs, setting goals, identifying resources, and evaluating their own learning.
Self-Regulated LearningSelbstreguliertes LernenSRLAn approach where learners are proactively involved in their learning processes, managing and regulating their own understanding.
ServerServer-
Single Instruction, Multiple DataEinzelanweisung, MehrfachdatenSIMDA type of parallel processing. Performs the same instruction on multiple data points simultaneously, such as vector operations like AVX-512.
SupercomputerSuperrechner-Vaguely defined term, ususally refering to any high performance computing cluster.
Target-GroupsZielgruppen-Specific groups of people for which something, such as a product or an educational program, is intended.
Tensor Processing UnitTensor-RecheneinheitTPUA type of application-specific integrated circuit specifically for neural network machine learning.
Thread(Ausführungsstrang)-A smallest sequence of instructions that can be managed independently by an operating system scheduler.
Transformative LearningTransformatives Lernen-Learning that induces a significant shift or transformation in the way an individual perceives, understands, or views the world.
TutorialsTutorials-Detailed instructions or lessons aimed at teaching specific topics or skills.
UnixUnix-A family of multitasking, multi-user computer operating systems.
Virtual MachineVirtuelle MaschineVMVirtualization of an isolated guest computer system (VM) can be used to run an OS independently of a host computer system. VMs typically do not emulate the guest computer, but use native execution to run the guest instructions directly on the physical processor of the host system.
Virtual RealityVirtuelle RealitätVRA simulated experience that can be similar or completely different from the real world.
Web BrowserWebbrowser-A software application for accessing information on the World Wide Web.