Compute kernelIn computing, a compute kernel is a routine compiled for high throughput accelerators (such as graphics processing units (GPUs), digital signal processors (DSPs) or field-programmable gate arrays (FPGAs)), separate from but used by a main program (typically running on a central processing unit). They are sometimes called compute shaders, sharing execution units with vertex shaders and pixel shaders on GPUs, but are not limited to execution on one class of device, or graphics APIs.
Intel MICXeon Phi was a series of x86 manycore processors designed and made by Intel. It was intended for use in supercomputers, servers, and high-end workstations. Its architecture allowed use of standard programming languages and application programming interfaces (APIs) such as OpenMP. Xeon Phi launched in 2010. Since it was originally based on an earlier GPU design (codenamed "Larrabee") by Intel that was cancelled in 2009, it shared application areas with GPUs.
Frontier (superordinateur)Frontier, ou OLCF-5, est le premier supercalculateur exaflopique au monde, hébergé à l'Oak Ridge Leadership Computing Facility (OLCF) dans le Tennessee, aux États-Unis. C'est le successeur de Summit (OLCF-4). Frontier devient en juin 2022 le supercalculateur le plus rapide au monde, avec une performance au Rmax de . Frontier utilise une combinaison de processeurs AMD Epyc 7A53s avec 64 cœurs et de GPU Radeon Instinct MI250X, et occupe 74 baies .
OpenCLOpenCL (Open Computing Language) est la combinaison d'une API et d'un langage de programmation dérivé du C, proposé comme un standard ouvert par le Khronos Group. OpenCL est conçu pour programmer des systèmes parallèles hétérogènes comprenant par exemple à la fois un CPU multi-cœur et un GPU. OpenCL propose donc un modèle de programmation se situant à l'intersection naissante entre le monde des CPU et des GPU, les premiers étant de plus en plus parallèles, les seconds étant de plus en plus programmables.
Embarrassingly parallelIn parallel computing, an embarrassingly parallel workload or problem (also called embarrassingly parallelizable, perfectly parallel, delightfully parallel or pleasingly parallel) is one where little or no effort is needed to separate the problem into a number of parallel tasks. This is often the case where there is little or no dependency or need for communication between those parallel tasks, or for results between them. Thus, these are different from distributed computing problems that need communication between tasks, especially communication of intermediate results.
Scratchpad memoryScratchpad memory (SPM), also known as scratchpad, scratchpad RAM or local store in computer terminology, is an internal memory, usually high-speed, used for temporary storage of calculations, data, and other work in progress. In reference to a microprocessor (or CPU), scratchpad refers to a special high-speed memory used to hold small items of data for rapid retrieval. It is similar to the usage and size of a scratchpad in life: a pad of paper for preliminary notes or sketches or writings, etc.