TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

LU Factorization and Linear Systems for Programers

131 点作者 disaster01超过 7 年前

7 条评论

kxyvr超过 7 年前
By the way, if anyone is interested in good open source opportunities, computational linear algebra is nowhere near a solved problem and there&#x27;s good opportunity for impactful contribution. The computational challenges of the algebra versus factorizations is one angle. Dense versus sparse is another. Shared memory parallelization vs distributed memory vs GPUs is another. Even on the GPU, there are different strategies depending on whether or not the entire matrix fits on a single GPU or if we have to use multiple GPUs. Incomplete or multilevel direct methods used as effective preconditioners for iterative methods are also important. Hell, even efficient direct techniques embedded in indirect solvers is important.<p>Part of the way to get started would be to look at something like a general numerical linear algebra book like Numerical Linear Algebra from Trefethen and Bau. There are better computational algorithms than what they present, but they do a good job at introducing important factorizations and why we care about them. Then, have a look at Tim Davis&#x27; book Direct Methods for Sparse Linear Systems. The codes in that book are online. Then, try to reimplement these algorithms in other languages, parallelize them, or make them better. These are good algorithms, but there are better and Tim&#x27;s more recent codes are actively used by both MATLAB and Octave. Then, look for missing routines in open source libraries. For example, I just did a quick look and MAGMA currently lists missing routines between them and LAPACK.<p>Anyway, it&#x27;s not a field for everyone, but it&#x27;s one that good architecture and parallelization knowledge can have a positive impact. Nearly all engineering codes depend on good solvers, so the impact is wide.
评论 #15352653 未加载
评论 #15355447 未加载
评论 #15352921 未加载
dagss超过 7 年前
For programmers, the really interesting part of dense linear algebra is how to achieve high performance, as blocking techniques have to be used to amortize loads from memory to cache.<p>Google for Goto&#x27;s &quot;Anatomy of high-performance matrix multiplication&quot;, one of my favorite programming texts.<p>Also the papers underlying the development of the &quot;Elemental&quot; library for distributed dense linear algebra is worth a look.
评论 #15351075 未加载
评论 #15352609 未加载
评论 #15352507 未加载
dragandj超过 7 年前
The software used in the tutorial: <a href="https:&#x2F;&#x2F;github.com&#x2F;uncomplicate&#x2F;neanderthal" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;uncomplicate&#x2F;neanderthal</a><p><a href="http:&#x2F;&#x2F;neanderthal.uncomplicate.org" rel="nofollow">http:&#x2F;&#x2F;neanderthal.uncomplicate.org</a>
hprotagonist超过 7 年前
Golub&#x27;s &quot;Matrix Computations&quot; remains a must-read reference text here: <a href="http:&#x2F;&#x2F;web.mit.edu&#x2F;ehliu&#x2F;Public&#x2F;sclark&#x2F;Golub%20G.H.,%20Van%20Loan%20C.F.-%20Matrix%20Computations.pdf" rel="nofollow">http:&#x2F;&#x2F;web.mit.edu&#x2F;ehliu&#x2F;Public&#x2F;sclark&#x2F;Golub%20G.H.,%20Van%2...</a>
评论 #15351895 未加载
评论 #15350548 未加载
flor1s超过 7 年前
If you want to learn about this stuff but need a more gentler introduction, check out Robert van de Geijn&#x27;s MOOC at <a href="http:&#x2F;&#x2F;www.ulaff.net" rel="nofollow">http:&#x2F;&#x2F;www.ulaff.net</a> (if there is currently no session, just download the lecture notes and use those, they have lecture videos embedded).
compumike超过 7 年前
For a gentler introduction to LU factorization and how useful the decomposition is for efficiently re-solving the same system multiple times, I&#x27;d offer up the &quot;Systems of Equations&quot; section of the &quot;Ultimate Electronics&quot; book I&#x27;ve been working on: <a href="https:&#x2F;&#x2F;www.circuitlab.com&#x2F;textbook&#x2F;systems-of-equations&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.circuitlab.com&#x2F;textbook&#x2F;systems-of-equations&#x2F;</a><p>I tried to go back and forth between the 5x+2y=3 style equations that most people are familiar with and the matrix forms.
leeoniya超过 7 年前
interactive demo of LU-like and QR-like decomposition of affine matrices:<p><a href="http:&#x2F;&#x2F;frederic-wang.fr&#x2F;decomposition-of-2d-transform-matrices.html" rel="nofollow">http:&#x2F;&#x2F;frederic-wang.fr&#x2F;decomposition-of-2d-transform-matric...</a><p>i mostly copy-pasted the js code from this page as a contribution to this lib back in the day: <a href="https:&#x2F;&#x2F;github.com&#x2F;epistemex&#x2F;transformation-matrix-js" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;epistemex&#x2F;transformation-matrix-js</a>