cuda

[nnop]

architecture

• http://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html#virtual-architecture-feature-list

• Pascal GPU 架构详解 by 卜居

[nnop]

capability

• https://developer.nvidia.com/cuda-gpus

[nnop]

posts

• “CUDA Tutorial”
• CUDA Toolkit Documentation
  • Programming Guide
  • Best Practices Guide
• SC11 Tutorial: High Performance Computing with CUDA
• An Easy Introduction to CUDA C and C++
• An Even Easier Introduction to CUDA
• CUDA 简明指南
• CUDA名词介绍：sp，sm，thread，block，grid，warp等

[nnop]

courses

• Nvidia - Existing University Courses
• Caltech - CS 179: GPU Programming
• NYU - CSCI-UA.0480-003: Parallel Computing
• NYU - CSCI-GA.3033-004: Graphics Processing Units (GPUs): Architecture and Programming
• CMU - 15-418/618: Parallel Computer Architecture and Programming
• UMW - CPSC 425: Parallel Computing
• Stanford - cs193g: Programming Massively Parallel Processors with CUDA

[nnop]

Thrust

• http://cs.boisestate.edu/~alark/thrust_intro/presentation.html

[nnop]

algorithm

• Optimizing Parallel Reduction in CUDA

[nnop]

• block之间完全独立运行
• CUDA Parallel Thread Management

[nnop]

• Numba - 3.2. Writing CUDA Kernels
• Organizing threads
• CUDA系列学习（五）GPU基础算法: Reduce, Scan, Histogram

[nnop]

• Mark Harris @ NVIDIA Developer Blog

[nnop]

Unified Memory

好处

• 简化编程
• 数据结构无需deep copy
• 可以用c++数据结构

NOTICE

• 由于不再用cudaMemcpy(), kernel执行完后一定要同步

caffe

• Incorporating cuda unified memory into caffe #2775
• Switch to Unified Virtual Address memory copies #555

[nnop]

optimization

• Optimizing CUDA
• Understanding and Using Atomic Memory Operations