Скачать
Programming GPUs in Python with PyCUDA and with Julia with CUDA.jl
Автор: Jan Verschelde
Загружено: 2023-02-18
Просмотров: 980
Описание:
This lecture introduces the data parallelism of the multiplication of two matrices, which is suitable for acceleration on a Graphics Processing Unit (GPU). The triple loop of a matrix-matrix multiplication is replaced by one single loop in a kernel launched by a Python script, using PyCUDA. In Julia, the package CUDA.jl allows to define kernels in the language Julia. All examples in this lecture use the Compute Unified Device Architecture (CUDA) and require a GPU by NVIDIA.
Доступные форматы для скачивания:
Скачать видео mp4
-
Информация по загрузке:
