如何在Windows 7上的Octave中准确启用64位索引(64位) [英] How exactly enable 64 bit indexing in Octave on Windows 7(64 bit)

问题描述

我知道这是一个老话题.但是，在用尽了Google的搜索之后，我仍然毫无头绪.我不喜欢 http://deeplearning.stanford.edu/tutorial/上的深度学习教程.在SoftMax回归部分的该教程上，我需要读取4704000字节(即，图像= fread(fp，28 * 28 * 30000，'uchar').脚本正在运行，但速度太慢.八度使用最多500mb.我发现我们应该允许64位索引( https: //www.gnu.org/software/octave/doc/interpreter/Compiling-Octave-with-64_002dbit-Indexing.html ). 我不确定如何在Windows Octave版本3.8.2中启用64索引. 欢迎任何建议或答案.谢谢！

I know this is an old topic. But still after exhausting search on Google I have no clue. I was fallowing along with Deep Learning tutorial on http://deeplearning.stanford.edu/tutorial/. On that tutorial in the SoftMax regression section I need to read 4704000 bytes( i.e images = fread(fp, 28*28*30000, 'uchar'). The script is working but too slow. Octave Using at most 500mb. I found that we should allow 64 bit indexing(https://www.gnu.org/software/octave/doc/interpreter/Compiling-Octave-with-64_002dbit-Indexing.html). I am not sure how exactly I can enable 64 indexing in Windows Octave version 3.8.2. Any suggestion or answers are welcomed. Thanks!

推荐答案

64位索引编制不是您启用的功能.您必须从这样的源代码构建Octave.在Windows中这并非微不足道. "Octave必须使用选项--enable-64进行配置"一词，是指您在实际构建Octave之前运行的配置脚本，而不是可以从Octave运行的某些配置文件或选项.

64 bit indexing is not something you enable. You must build Octave from source like that. That is not trivial in windows. The word configure in "Octave has to be configured with the option --enable-64", refers to the configure script that you run before actually building Octave, not to some configuration file or option you can run from Octave.

此外，还需要从64位源中重建诸如BLAS和LAPACK之类的Octave依赖项，而据我所知，这种依赖关系就显得微不足道了.即使在Linux中，大多数发行版也只能提供32位版本.请参阅最近在邮件列表上的讨论.

In addition, Octave dependencies such as BLAS and LAPACK will also need to be rebuilt from source with 64 bit which I am told is even less trivial. Even in Linux, most distributions will provide 32 bit builds only. See a recent discussion on the mailing list.

这篇关于如何在Windows 7上的Octave中准确启用64位索引(64位)的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！