| CPU |
Central Processing Unit - the primary component of a computer that performs most processing inside a computer. |
| CUDA |
Compute Unified Device Architecture - a parallel computing platform and application programming interface model created by NVIDIA. |
| IPU |
Intelligence Processing Unit - a type of microprocessor designed specifically for artificial intelligence and machine learning tasks. |
| XPU |
Generic term, often referring to a heterogeneous compute platform encompassing CPUs, GPUs, and other acceleration devices. |
| MKLDNN |
Math Kernel Library for Deep Neural Networks, now Intel DNNL (Deep Neural Network Library) - an open-source performance library for deep learning applications. |
| OpenGL |
Open Graphics Library - a cross-language, cross-platform specification for rendering 2D and 3D vector graphics. |
| OpenCL |
Open Computing Language - a framework for writing programs that execute across heterogeneous platforms consisting of CPUs, GPUs, and other processors. |
| iDeep |
Intel Deep Learning Enhancement Primitives - provides deep learning executions in C++. |
| HIP |
Heterogeneous-compute Interface for Portability - an API developed by AMD for writing portable applications across GPU platforms. |
| VE |
Vector Engine - might refer to vector processing units used for accelerating data-parallel workloads. |
| FPGA |
Field-Programmable Gate Array - an integrated circuit designed to be configured by the customer or designer after manufacturing. |
| MAIA |
Specific reference unknown - the name could be associated with a specific project or technology initiative. |
| XLA |
Accelerated Linear Algebra - a domain-specific compiler for linear algebra that can speed up TensorFlow computations. |
| Lazy |
Lazy execution strategy - deferring computations until necessary, typically for optimization reasons. |
| Vulkan |
A low-overhead, cross-platform 3D graphics and computing API. |
| MPS |
Metal Performance Shaders - a framework that provides a highly optimized set of graphics and compute shaders for Apple platforms. |
| Meta |
Could refer to the use of metaprogramming techniques or a project within the broader computing context. |
| HPU |
Holographic Processing Unit - found in devices like Microsoft's HoloLens for processing spatial input and environment mapping. |
| MTIA |
Meta's custom AI inference chip, targeted at accelerating machine learning workloads. |
| PrivateUseOne |
Typically refers to user-defined or proprietary hardware units within a system. |