2,700 technical terms and definitions
copper to copper direct bonding, dbi direct bond interconnect, bumpless 3d packaging, wafer to wafer bonding alignment, hybrid bonding overlay accuracy, hbm4 hybrid bonding, hbm4 die to die bonding
Hybrid bonding joins two wafers or dies by bonding copper pads and the surrounding dielectric at the same time, producing a metallurgical copper-to-copper connection with no…
cubesat component qualification drift prediction, cubesat component qualification, component qualification drift, qualification drift prediction, cubesat component
CubeSat Component Qualification Drift Prediction with Machine Learning Introduction & Motivation CubeSat Component Qualification Drift Prediction with Machine Learning addresses…
infrastructure
cuBLAS is the NVIDIA GPU implementation of BLAS routines for dense linear algebra, especially matrix multiplication - it powers the GEMM-heavy compute core of modern neural…
gpu math libraries, tensor operations gpu, cublas performance tuning, cudnn convolution optimization
cuBLAS and cuDNN Optimization is the systematic tuning of NVIDIA's highly-optimized math libraries to achieve 80-95% of theoretical peak performance — where cuBLAS (CUDA Basic…
cuckoo hashing collision, cuckoo hashing, hashing collision
Cuckoo Hashing Collision Introduction & Motivation Cuckoo Hashing Collision addresses a central problem in data-structures process control and load-factor-high-threshold…
cuckoo search levy flight, cuckoo search levy, search levy flight, cuckoo search, search levy, levy flight
Cuckoo Search Levy Flight Introduction & Motivation Cuckoo Search Levy Flight addresses a central problem in ml process control and convergence-speed-generations prediction: how…
compute capability, nvidia
CUDA and Compute Capability What is CUDA? CUDA (Compute Unified Device Architecture) is NVIDIA's parallel computing platform and API that enables GPUs to be used…
nvidia, programming, gpu, kernel, parallel, cudnn
CUDA (Compute Unified Device Architecture) is NVIDIA's parallel computing platform and API for GPU programming — enabling developers to leverage GPU hardware for general-purpose…
hardware
CUDA (Compute Unified Device Architecture) is NVIDIA's parallel computing platform and programming model that enables developers to use NVIDIA GPUs for general-purpose…
thread block cluster, cooperative launch, grid sync, gpu thread cooperation
CUDA Cooperative Groups is the programming model extension that provides flexible, hierarchical thread synchronization and communication beyond the traditional warp and thread…
shader, gpu core
CUDA is NVIDIA's platform for general-purpose computing on GPUs: a programming model, a compiler and runtime, and a deep stack of libraries that together let software harness…
cuda, hardware
CUDA is NVIDIA's platform for general-purpose computing on GPUs: a programming model, a compiler and runtime, and a deep stack of libraries that together let software harness…
Parallelism, nested kernels, recursion
CUDA Dynamic Parallelism is a CUDA feature enabling kernels executing on the GPU to dynamically launch additional kernels without returning control to the CPU — enabling recursive…
kernel launch kernel, device launch, nested kernels, gpu recursion
CUDA Dynamic Parallelism is the ability for GPU kernels to launch other GPU kernels directly from the device — eliminating round-trips to the CPU for recursive or adaptive…
device side kernel launch, nested parallelism cuda, gpu dynamic scheduling, cuda cdp
CUDA Dynamic Parallelism is the capability for GPU kernels to launch other kernels directly from device code without CPU involvement — enabling recursive algorithms, adaptive…
cuda graphs, cudagraph, cuda graph replay, cuda graph launch, cuda, optimization
CUDA Graphs is a CUDA runtime feature that records a sequence of GPU kernel launches and memory operations into a reusable graph object, then replays the entire sequence with…
execution graph, operator fusion, kernel fusion, cuda graph optimization
CUDA Graphs are a mechanism to capture a sequence of GPU operations as a graph and replay them with minimal CPU overhead — eliminating the per-kernel launch overhead that limits…
cuda graph capture, gpu graph execution, stream capture, cuda graph optimization
CUDA graphs are a mechanism for capturing a whole sequence of GPU operations once and then replaying it with a single launch, so that the per-kernel cost of telling the GPU what…
execution, optimization, launch overhead
CUDA Graphs is a CUDA runtime feature that records a sequence of GPU kernel launches and memory operations into a reusable graph object, then replays the entire sequence with…
cuda graph optimization, graph capture cuda, cuda graph launch, kernel launch overhead reduction
CUDA Graphs is a CUDA runtime feature that records a sequence of GPU kernel launches and memory operations into a reusable graph object, then replays the entire sequence with…
cuda graph capture, graph instantiation launch, graph node kernel memcpy, cuda graph optimization
CUDA Graph API: Fixed-Topology Amortized Launching — reducing kernel launch overhead for inference and fixed-pattern workloads CUDA Graphs capture sequences of kernels and memory…
gpu kernel tuning, cuda performance optimization, warp efficiency optimization, cuda memory coalescing
CUDA Kernel Optimization is the systematic tuning of GPU kernels to maximize throughput, minimize latency, and achieve peak hardware utilization — where optimizations like memory…
optimization, latency hiding, throughput
CUDA Occupancy Optimization is a critical CUDA optimization methodology balancing the number of active warps on GPU against available GPU resources (register files, shared memory)…
cuda, cuda c++, nvidia gpu programming, cuda kernel
CUDA programming definition and practical boundary. uses NVIDIA’s parallel-computing platform and language extensions to launch C++, Python, or library work on NVIDIA GPUs.…
gpu thread, cuda kernel
CUDA Programming — NVIDIA's parallel computing platform that enables general-purpose computation on GPUs, leveraging thousands of cores for massive parallelism. Execution…
shared memory bank conflict, shared memory tiling technique, shared memory reduction, dynamic shared memory allocation
CUDA Shared Memory Optimization is the technique of utilizing the fast, programmer-managed on-chip memory (shared memory) within each GPU streaming multiprocessor to cache…
concurrency, overlap, asynchronous
CUDA Stream Concurrency and Overlap is an advanced CUDA programming technique employing independent streams to enable concurrent execution of kernels, data transfers, and host…
gpu stream overlap, cuda async execution, stream synchronization, multi stream gpu
CUDA Stream Concurrency is the GPU programming technique that uses multiple independent execution streams to overlap kernel execution, memory transfers, and host computation…
gpu kernel overlap, async execution gpu, multi stream gpu, concurrent kernel execution
CUDA Streams and Asynchronous GPU Execution are the programming mechanisms that enable overlapping of multiple GPU operations — kernel execution, memory transfers, and host…
cuda, infrastructure
CUDA Stream Concurrency and Overlap is an advanced CUDA programming technique employing independent streams to enable concurrent execution of kernels, data transfers, and host…
async cuda, stream synchronization, multi-stream, cuda concurrent execution
CUDA Streams are queues of GPU operations that execute in order within the stream but potentially overlap with operations in other streams — enabling concurrent execution…
cuda event synchronization, multi stream overlap, async memcpy compute overlap, stream priority cuda
CUDA Streams and Asynchronous Execution enable concurrent kernel launches, memory transfers, and host-device synchronization, hiding latencies and improving GPU utilization…
asynchronous execution gpu, stream synchronization, concurrent kernel execution, multi stream programming
CUDA Streams and Concurrency are the mechanisms for overlapping independent GPU operations — enabling simultaneous execution of multiple kernels, concurrent data transfers…
gpu stream programming, cuda concurrent execution, asynchronous cuda operations, cuda stream synchronization
CUDA Streams and Concurrency is the programming model that enables overlapping of kernel execution, memory transfers, and host operations through asynchronous task queues — where…
cublas, cublaslt, matrix multiply cuda, gemm gpu, cuda linear algebra
CUDA Tensor Operations and cuBLAS is the NVIDIA GPU library ecosystem for high-performance linear algebra and tensor computation — providing highly optimized implementations…
cuda grid block thread, gpu multiprocessing, sm streaming multiprocessor, cuda programming model
CUDA Thread Hierarchy is the elegant software abstraction introduced by NVIDIA that perfectly maps massive amounts of parallel software work (millions of threads) onto…
thrust parallel algorithms, thrust stl gpu, cuda high level library, thrust performance
CUDA Thrust Library is the high-level C++ template library providing STL-like parallel algorithms and data structures for CUDA — offering 40+ algorithms including sort (100-300…
managed memory, cuda uvm, page migration gpu, memory oversubscription gpu
CUDA Unified Memory is the programming model that provides a single, coherent address space accessible from both CPU and GPU — automatically migrating pages between CPU and GPU…
advanced patterns, oversubscription
CUDA Unified Memory Advanced Patterns is an advanced GPU memory management feature enabling transparent migration of data between host and GPU memories through unified virtual…
unified virtual addressing gpu, managed memory cuda malloc, page migration gpu cpu, cuda memory prefetch hints
CUDA Unified Memory Management is a memory architecture feature that creates a single coherent virtual address space accessible by both CPU and GPU, with the CUDA runtime…
warp intrinsics, warp cooperative, warp synchronous programming
CUDA Warp-Level Programming is the exploitation of the GPU's SIMT execution model at the warp granularity (32 threads) using warp-synchronous primitives, shuffle instructions…
hardware
cuDNN (CUDA Deep Neural Network Library) is NVIDIA's GPU-accelerated library providing highly optimized implementations of deep learning primitives — delivering the hand-tuned…
infrastructure
cuDNN is the NVIDIA deep neural network primitives library that provides optimized kernels for core DL operations - it is the standard acceleration layer behind convolution…
packaging
Cull is the residual molding compound left in the pot and transfer channels after cavity filling in transfer molding - it is non-product material that affects both process…
Cultural Anthropology: Kinship, Culture, and Globalization Cultural anthropology studies how people create meaning, organize relationships, exchange resources, and adapt…
reliability
Cumulative failure distribution is the probability curve that shows what fraction of a population has failed by a given time - it is the direct view of accumulated reliability…
production
Cumulative Yield is the total yield considering all yield loss mechanisms across the entire manufacturing flow — calculated as the product of individual yields at each stage…
apple, apple park
Cupertino is location intent associated with Cupertino city context and major technology-campus references - It is a core method in modern semiconductor AI, geographic-intent…
packaging
Cure time is the duration required for molding compound to achieve sufficient crosslinking and mechanical integrity in the mold - it governs package strength, residual stress…
learning, growth mindset
Cultivating curiosity and a growth mindset Cultivating curiosity and a growth mindset is essential for AI practitioners as the field evolves rapidly, requiring continuous…