jasperzhong / cs-notes Goto Github PK

传说中的零知识证明.

https://www.youtube.com/watch?v=FuKEpOhiVPg&vl=zh-Hant

https://tour.golang.org/list

https://pdos.csail.mit.edu/6.824/papers/tour-faq.txt

https://www.udacity.com/course/computer-networking--ud436

基本保持每日一题. 主要用cxx.

以前面试实习前都会过一遍这本书.

差分隐私

本科时候啃过一部分. 现在的目的是学习跟组合优化有关的部分. 对高级数据结构等内容不感兴趣.

先看第四部分Advanced Design and Analysis Techniques. selected topics也有一些ch值得读.

solutions: https://sites.math.rutgers.edu/~ajl213/CLRS/CLRS.html

讲道理, CLRS的写作非常棒，简洁易懂. 感觉写paper可以学习一个.

https://pdos.csail.mit.edu/6.824/schedule.html

https://github.com/vycezhong/MIT6.824

https://www.youtube.com/channel/UC_7WrbZTCODu1o_kfUMq88g/videos

鸽两年了

https://leetcode-cn.com/problem-list/2cktkvj/

基本每日一题.

https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/index.html

https://pytorch.org/tutorials/beginner/dist_overview.html

必须的

https://en.wikipedia.org/wiki/Roofline_model

https://zhuanlan.zhihu.com/p/34204282

今天了解到这个概念，发现极为有用. 必须彻底掌握.

https://github.com/NVIDIA/Megatron-LM

刻不容缓.

https://www.youtube.com/user/lefticus1/featured

真是个不错的频道

打算按照播放量顺序观看

https://github.com/NVIDIA/cuda-samples

简直是宝藏.

Satisfiability modulo theories

https://horovod.readthedocs.io/en/stable/elastic_include.html

https://en.wikipedia.org/wiki/Cache_coherence

write-through vs write-back

打算系统看一遍

看这本书要注意经常出现的The curx of the problem，帮助理解这是在解决什么问题.

https://github.com/facebookresearch/fairscale

FairScale is a PyTorch extension library for high performance and large scale training.

看上去是几个module的并集.

https://learning.edx.org/course/course-v1:LinuxFoundationX+LFS151.x+2T2020/home

https://mpitutorial.com/tutorials/

写MPI程序让我想起了写CUDA. 有点像.

https://developers.google.com/optimization

简直就是宝藏！！！

OR-Tools is open source software for combinatorial optimization, which seeks to find the best solution to a problem out of a very large set of possible solutions.

https://www.youtube.com/playlist?list=PLeKd45zvjcDFUEv_ohr_HdUFe97RItdiB

notes: https://www.cl.cam.ac.uk/teaching/2021/ConcDisSys/dist-sys-notes.pdf

讲课的是DDIA作者Martin Kleppmann. 太赞了.

话说英国口音真好听（逃

主要是熟悉CLRS这种比较形式化的表达. 感觉这方面自己比较欠缺. 尤其注意下里面的各种定理和证明. 顺便回顾下图算法.

6.29更: 另外一个原因是，DL framework一般把模型表达成计算图. 想在上面做些优化，经常会遇到图论的知识.

https://arrow.apache.org/docs/python/plasma.html