4 ms·

I think Eigen is more popular these days. https://eigen.tuxfamily.org/index.php?title=Main_Page https://eigen.tuxfamily.org/index.php?title=Main_Page

• galangalalgol 4y ago
  I replaced lapackc with eigen and it was largely performance neutral. But nicer interface and header only. Those benchmark charts seem to mostly imply blaze is better at fine grained parallelism. But I largely use threadpools with coarse grained parallelism for stuff like this. Still quite interesting, not sure why I haven't heard of it before. I'd like to see nalgebra in the benchmarks...
  • jey 4y ago
    Agreed. Some of the upsides of eigen compared to lapack relate to the use of expression templates to eliminate temporaries. And the ability to specialize code for small fixed size arrays and matrices. If you have large matrices and don’t make use of the expression templates, I would expect performance to be about the same.
    • galangalalgol 4y ago
      I was doing some medium sized level 3 stuff with a small temporary and a small decomposition. But that over and over again pretty fast.
  • badpun 4y ago
    Last I tried eigen, adding those headers to the project added a full 3 seconds to my compile (it was under a second before), seriously harming ability to quickly play with the code, as you would say in NumPy.