716 technical terms and definitions
fail to reject vs accept, legal trial analogy, Ronald Fisher p-value, Neyman-Pearson decision framework
The null hypothesis is the default assumption in a hypothesis test — typically that there's no effect, no difference between groups, or no relationship between variables — which…
generative models
Null-Text Inversion is a technique for inverting real images into the latent space of a text-guided diffusion model by optimizing the unconditional (null-text) embedding at each…
multimodal ai
Null-Text Inversion is an inversion method that optimizes unconditional text embeddings to reconstruct a real image in diffusion models - It enables faithful real-image editing…
Null Value Imputation via Inference Introduction & Motivation Null Value Imputation via Inference addresses a central problem in {subject} assurance and {target} management: how…
non uniform memory access, numa aware
NUMA (Non-Uniform Memory Access) — a memory architecture where access time depends on which CPU socket the memory is attached to, critical for multi-socket server performance.…
non uniform memory access, numa aware scheduling, memory affinity numa, socket memory topology
NUMA Architecture and Optimization is the multi-processor memory architecture where each processor socket has locally attached memory that it can access faster (50-100 ns) than…
numa node affinity, libnuma binding, first touch policy numa, remote numa penalty
NUMA Architecture and Memory Affinity enable explicit placement of data and threads on multi-socket systems to exploit local memory bandwidth and latency, critical for HPC…
non uniform memory access, numa node affinity binding, numa memory placement policy, numa interleave first touch
NUMA-Aware Memory Allocation is the practice of placing memory pages on the NUMA (Non-Uniform Memory Access) node closest to the processor that will most frequently access them…
non-uniform memory access, memory affinity binding, numa node topology, local memory bandwidth optimization
NUMA-Aware Memory Allocation — Optimizing memory placement and access patterns on Non-Uniform Memory Access architectures where memory latency and bandwidth depend on the physical…
non uniform memory access, numa affinity, memory locality parallel
NUMA-Aware Optimization is the set of programming and system configuration techniques that account for Non-Uniform Memory Access (NUMA) architecture in multi-socket and modern…
memory binding, libnuma, numa topology, numa optimization
NUMA-Aware Programming is the practice of allocating and accessing memory in ways that minimize cross-NUMA-node memory accesses — exploiting the topology of Non-Uniform Memory…
numa memory allocation, numa topology, numa binding, non uniform memory access
NUMA-Aware Programming is the performance optimization discipline for multi-socket and chiplet-based systems where memory access latency and bandwidth depend on the physical…
non uniform memory access, numa topology scheduling, numa memory allocation policy, numa balancing linux
NUMA-Aware Programming is the practice of structuring parallel applications to account for the non-uniform memory access costs of modern multi-socket systems — placing data…
numa memory allocation policy, numa thread affinity binding, numa topology detection, numa performance penalty
NUMA-Aware Programming is the practice of structuring parallel applications to account for Non-Uniform Memory Access architecture — where memory access latency and bandwidth…
numa placement policy, memory locality scheduler, socket affinity control, numa runtime tuning
NUMA-Aware Scheduling is the placement strategy that aligns threads and memory to socket locality on multisocket servers. What It Covers Core concept: reduces remote memory…
numa node, memory controller cpu, numa locality, smp symmetric multiprocessing
Non-Uniform Memory Access (NUMA) is the dominant memory architecture in massive modern servers and supercomputers where memory banks are physically divided into localized "nodes"…
numa non uniform memory, numa non uniform, non uniform memory, numa non, non uniform, uniform memory
NUMA Non Uniform Memory Introduction & Motivation NUMA Non Uniform Memory addresses a central problem in systems process control and remote-access-latency-penalty prediction: how…
generative models
Number of diffusion steps is the count of reverse denoising iterations executed during sampling to transform noise into a final image - it is the main quality-latency control knob…
Number Theory: Primes, Congruences, and Diophantine Equations Number theory — the study of integers and their properties — is one of the oldest branches of mathematics yet remains…
evaluation
Numeracy Analysis in NLP is the systematic study and evaluation of how well language models understand, represent, and generate numerical information — covering magnitude…
Numeric Data Validation and Correction Introduction & Motivation Numeric Data Validation and Correction addresses a central problem in {subject} assurance and {target} management…
computational mathematics, numerical computation, numerical error analysis, scientific computing foundations, approximation algorithms, roundoff and truncation error
Numerical analysis studies how continuous and discrete mathematical problems become reliable finite computations. It asks whether a problem is sensitive to its data, whether…
numerical aperture, na, lithography
Numerical Aperture (NA) is the fundamental optical parameter that determines a lithography lens's ability to resolve fine features — defined as NA = n × sin(θ) where n is…
discrete vs continuous numerical data, interval vs ratio scale, quantitative data
Numerical data is data expressed as numbers that measure a quantity — height, revenue, temperature, count of items sold — and the defining property that separates it…
Numerical Linear Algebra: Decompositions and Eigenvalue Computation 1. Why numerical linear algebra matters Scientific computing depends on solving linear systems and computing…
Numerical Linear Algebra: Matrix Factorizations, SVD, QR Decomposition, and Iterative Eigensolvers Numerical linear algebra converts the abstract theorems of matrix theory into…
FEM FDM FVM, finite element, finite difference, conjugate gradient, monte carlo, level set, TCAD simulation, computational methods
Every step in a semiconductor process is, underneath, a partial differential equation nobody can solve with pencil and paper. Dopants diffuse, fields settle, plasma etches, light…
Numerical Methods for ODEs: Runge-Kutta, Multistep, and Stiff Solvers Ordinary differential equations (ODEs) model dynamical systems across all scientific disciplines — orbital…
quantitative variable, numeric feature, feature scaling, binning, discretization, measurement variable
A numerical variable is a variable whose values are numbers on which arithmetic is meaningful — height, income, number of children, reaction time, temperature — so differences…
evaluation
NumGLUE is the multi-task benchmark specifically targeting the numerical reasoning capabilities of NLP models — aggregating 8 distinct datasets that require quantitative…
ndarray, numpy array, numpy broadcasting, numpy vectorization, numpy linear algebra, numpy indexing, numpy strides, numpy ufunc, numpy performance, numerical python, numpy vs python list
NumPy is a Python library that stores numerical data in contiguous, typed memory buffers (the ndarray) and dispatches arithmetic over those buffers to optimized C and Fortran…
Nutrition Science: Metabolism, Diet, and Health Nutrition science studies how foods and nutrients support energy balance, growth, repair, and health across the life course. It…
NVIDIA accelerator, H100 GPU, B200 GPU, Blackwell GPU, CUDA GPU, nvidia, nvidia corporation, jensen huang, nvidia ai
NVIDIA GPU. is a massively parallel processor and surrounding platform used for graphics, AI training and inference, scientific computing, simulation, media, and data analytics.…
nvidia a100, what is nvidia a100
An Ampere-generation data-center GPU. NVIDIA A100 is a data-center accelerator introduced in May 2020 using NVIDIA’s Ampere architecture. It targets AI training, inference…
nvidia geforce 256, what is nvidia geforce 256
GeForce 256 was NVIDIA’s 1999 graphics processor associated with integrated hardware transform and lighting. Its historical importance is the movement of more geometry work…
nsight compute, nsight systems, gpu profiling nvidia
NVIDIA Nsight is the NVIDIA profiling suite for detailed analysis of GPU kernels, memory behavior, and system-level execution timelines - it enables deep diagnosis of performance…
nvidia rubin, what is nvidia rubin
A GPU architecture within the Vera Rubin platform. NVIDIA Rubin identifies a data-center GPU generation that follows Blackwell in NVIDIA’s AI computing roadmap. It also appears…
nvidia tesla, what is nvidia tesla
NVIDIA Tesla was a GPU-computing product line introduced in 2007. The company’s June 2007 product literature positioned it for scientific and engineering computation…
nvidia volta, what is nvidia volta
Volta is an NVIDIA GPU architecture introduced commercially with the Tesla V100 in 2017. The architecture is a design generation rather than a single board configuration or one…
gb200 nvl72, rack scale, rack scale computing, nvlink fabric, nvlink switch fabric
Rack-scale computing is the recognition that a single AI training or inference job has outgrown any one server, so the unit of design has moved from the chip to the entire rack.…
pcie, interconnect, bandwidth, gpu, nvswitch, nccl
NVLink is NVIDIA's high-bandwidth interconnect for GPU-to-GPU and GPU-to-CPU communication — providing 600-900 GB/s bidirectional bandwidth compared to PCIe's 64 GB/s, enabling…
infrastructure
NVLink is NVIDIA's high-bandwidth interconnect that wires GPUs directly to one another so they can read and write each other's memory far faster than over the standard PCIe bus.…
gpu interconnect, peer to peer gpu, p2p access, multi-gpu communication
NVLink is NVIDIA's high-bandwidth GPU-to-GPU interconnect — providing substantially higher bandwidth and lower latency than PCIe for multi-GPU systems, enabling efficient…
nvlink bandwidth topology, nvswitch fabric architecture, nvlink vs pcie performance, multi gpu nvlink
NVLink Interconnect is NVIDIA's proprietary high-bandwidth, low-latency GPU-to-GPU interconnect that provides 10-15× higher bandwidth than PCIe — enabling direct GPU memory access…
gpu interconnect comparison, pcie gpu, nvlink bandwidth, gpu to gpu communication
GPU Interconnect Technologies (NVLink vs. PCIe vs. NVSwitch) are the communication fabrics that connect GPUs to each other and to CPUs — where the bandwidth, latency, and topology…
gpu interconnect nvlink, nvlink bandwidth, nvswitch all to all, multi gpu communication
NVLink and NVSwitch are NVIDIA's proprietary high-bandwidth, low-latency interconnect technologies that connect GPUs within a server at bandwidths far exceeding PCIe — where…
nvm write disturb prediction, nvm write disturb, write disturb prediction, nvm write, write disturb, disturb prediction
NVM Write Disturb Prediction with Machine Learning Introduction & Motivation NVM Write Disturb Prediction with Machine Learning addresses a central problem in non-volatile memory…
infrastructure
NVSwitch is the switching fabric that interconnects multiple GPUs with high-bandwidth non-blocking communication inside accelerated systems - it provides uniform, scalable…
nvswitch topology design, gpu fabric nvswitch, nvswitch routing protocol, multi nvswitch configuration
NVSwitch Fabric Architecture is the switched interconnect topology that provides full non-blocking, all-to-all connectivity among GPUs using dedicated NVSwitch chips — each switch…
llm architecture
Nyströmformer is an efficient Transformer architecture that approximates the full softmax attention matrix using the Nyström method—a classical technique for approximating large…