DAY 1:08/18 (01) Introduction of parallel processing a. Shared Memory Multipleprocessor b. Distributed Memory Multiplecomputer (02) Concept of CUDA a. CUDA supported GPUs b. What is CUDA ? (03) CUDA installation a. CUDA driver & toolkit b. CUDA SDK c. Meet DeviceQuery
(04) CUDA programming model a. CUDA execution flow b. Index system in kernel function c. Streaming Multiprocessor(SM) d. NVCC - NVidia C Compiler e. CUDA Runtime library f. example: CUDA memory management (05) CUDA memory hierarchy a.Register b. Shared memory c. Local memory d. Global memory e. L1/L2 cache f. Constant memory g. Texture memory DAY 2:08/19
(06) Your First CUDA program : Vector Addition a. sequencial b. 1 B 1 T c. 1 B N T d. M B N T e. CUDA event functions (07) CUDA performance tunnig basic : Vector inner productor a. sequencial b. naive c. atomic function d. shared memory e. tree reduction f. kernel execution time distribution analysis (08) Optimizations a. memory optimization b. configuration optimization c. instruction optimization DAY 3:08/20 (09) Profilers a. cuda-memcheck b. NVVP - NViadia Visual Profiler (10) CUDA Stream, multiple GPUs, hybrid models a. streams b. with openmp c. with MPI (11) Cuda Libraries a. cublas b. curand c. cufft 更多課程詳細資訊,請見研碩網站:www.yanshuo.cc
- 新聞稿有效日期,至2015/07/30為止
聯絡人 :黃小姐 聯絡電話:02-2912-3028 電子郵件:frances@yanshuo.cc
上一篇:高普考季!各科考前猜題題庫 免費下載
下一篇:聯想控股「投資風」重磅襲港 獲市場追捧
|