Eigen (C++ library): Difference between revisions

Revision as of 20:07, 16 April 2024

Eigen is a template header-only C++ linear algebra library. It is one of the fastest and most popular.

Website

Usage

Compilation

Reference
For optimal performance, I recommend using the following flags when compiling.

GCC

-march=native and -mtune=native if running only locally or -march=skylake if distributing to relatively modern (since ~2015) cpus.
- Otherwise, at a minimum
- -mfma Enable fused multiply add
- -mavx2 Enable avx2 vector instructions
-DEIGEN_NO_DEBUG Set preprocessor define for eigen optimizations
-fopenmp OpenMP parallel execution
-O3 to enable optimizations

Data to Eigen

You can use Eigen::Map to create an eigen view for your existing data. This works with aligned or unaligned data, row-order or column-order, and different strides.
See Eigen: Interfacing with raw buffers for an example.

Math

SVD

Eigen comes with a few SVD implementations in its SVD Module.
If you only need low-rank approximations then you may be interested in randomized SVD.
This can be 10-20x faster when calculating low-rank approximations on large matrices.
Github Implementation
Finding structure with randomness paper
Facebook Blog post

@@ Line 8: / Line 8: @@
 For optimal performance, I recommend using the following flags when compiling.<br>
 ====GCC====
-*<code>-march=native</code> and <code>-mtune=native</code> if running only locally or <code>-march=skylake</code> if distributing to (Core ix-5000+) cpus.
+*<code>-march=native</code> and <code>-mtune=native</code> if running only locally or <code>-march=skylake</code> if distributing to relatively modern (since ~2015) cpus.
 **Otherwise, at a minimum
 **<code>-mfma</code> Enable fused multiply add