Enterprise Technology Glossary

6,173 results · page 123 of 309

• GPU-CPU Coherence

  GPU-CPU coherence is the property of a system in which GPUs and CPUs maintain a consistent view of shared memory through cache-coherence mechanisms, which matters in enterprises because it affects performance, data-movement overhead, and software design in GPU-accelerated workloads.

• GPU–CPU Coherence

  GPU–CPU coherence is a computer architecture capability that maintains a consistent shared-memory view between GPUs and CPUs, which matters in enterprise environments because it reduces data copying, simplifies heterogeneous programming, and affects performance and design choices for AI, analytics, and HPC platforms.

• GPU Health Monitor

  GPU Health Monitor is an automated capability that tracks GPU status, performance, and reliability metrics across servers and clusters, enabling enterprises to maintain availability, plan capacity, and coordinate operations for accelerator-dependent workloads in data center and cloud environments.

• GPU Scheduling Framework

  GPU scheduling framework is a software or system-level mechanism that manages how multiple workloads share GPU resources in enterprise and cloud environments, enabling controlled allocation, prioritization, and isolation of GPU capacity for AI, analytics, and high-performance computing workloads.

• GPU Virtualization

  GPU virtualization is a hardware and software capability that lets multiple virtual machines, containers, or processes share physical GPUs with controlled allocation and isolation, enabling centralized, policy-based use of GPU compute in virtual desktop, cloud, and data center environments.

• Gradient Accumulation

  Gradient accumulation is a deep learning training technique that aggregates gradients across multiple mini-batches before updating model parameters, enabling the use of larger effective batch sizes on constrained hardware and supporting enterprise training efficiency, scalability planning, and resource utilization control.

• Gradient Compression

  Gradient compression is a collection of methods that reduce the size of gradient data exchanged during distributed or federated training of machine learning models, which helps manage network bandwidth, training time, and operational cost in enterprise-scale AI workloads.

• Gradient Descent

  Gradient descent is an iterative optimization algorithm that updates model parameters along the negative gradient of an objective function during training, which matters in enterprise contexts because it determines training efficiency, resource consumption, and model quality for machine learning and deep learning workloads.

• Gramm–Leach–Bliley Act

  Gramm–Leach–Bliley Act is a U.S. federal law that imposes privacy, safeguards, and anti-pretexting requirements on financial institutions handling consumers’ nonpublic personal information, shaping data security programs, risk management, and regulatory compliance activities across banking, insurance, and other financial services enterprises.

• Graph Analytics Engine

  Graph analytics engine is a software platform or component that processes graph-structured data to compute relationships, paths, and structural patterns using graph algorithms, relevant for enterprises that need to analyze connected data such as users, systems, networks, and business entities.

• Graph Database

  Graph database is a type of NoSQL database that models data as nodes, relationships, and properties, enabling enterprises to query connected data for uses such as security analysis, operations, and customer analytics within broader data and application architectures.

• Graph Embedding

  Graph embedding is a machine learning technique that converts nodes, edges, or graphs into low-dimensional numeric vectors that preserve structural relationships, enabling enterprises to apply standard analytics and machine learning workflows to graph-structured data for classification, prediction, and similarity tasks.

• Graphics Processing Unit

  Graphics processing unit (GPU) is a specialized processor that executes parallel arithmetic and logic operations to accelerate graphics, artificial intelligence, and data-intensive workloads in enterprise environments, supporting higher throughput and efficiency for tasks such as model training, analytics, and visualization.

• Graph Inference Model

  Graph inference model is a machine learning approach that operates on graph-structured data to infer missing links, labels, or embeddings from node and edge relationships, which enterprises use in areas such as fraud detection, cybersecurity analytics, recommendations, and complex relational analysis.

• Graph Knowledge Store

  Graph knowledge store is a data management layer that represents and queries entities and relationships as a graph, enabling knowledge-centric retrieval, reasoning, and analytics across heterogeneous enterprise data sources for use cases such as governance, security analysis, and complex dependency queries.

• Graph Neural Network

  Graph Neural Network is a neural network architecture for learning on graph-structured data, allowing enterprises to model entities and relationships as nodes and edges and to perform prediction, classification, and risk analysis directly on relational data structures.

• Graph Optimization

  Graph optimization is the discipline of computing optimal or near-optimal solutions to problems expressed on graph models of nodes and edges, used in enterprises to support routing, allocation, capacity planning, and risk-aware decisions across interconnected systems and networks.

• Graph Optimization Pass

  Graph optimization pass is a compiler or runtime stage that rewrites and simplifies a computational graph to improve performance and resource efficiency while preserving semantics, which matters to enterprises running machine learning, dataflow, and analytics workloads at scale across diverse hardware.

• GraphQL

  GraphQL is a typed query language and runtime for APIs that exposes data through a schema-based contract, enabling clients in enterprise environments to request only required fields while centralizing API governance across services, domains, and digital products.

• Graph Query Language

  Graph query language is a formal query syntax for expressing operations over graph-structured data in graph databases and platforms. It matters in enterprise environments because it enables direct, declarative access to connected data for analytics, applications, and governance workflows.