I agree, my concern is that the authors and readers of the above aren't even aware of Neanderthal to include it in the Breeze comparisons. A reworked table with an added Neanderthal column might raise awareness, adoption, visibility, and appreciation. It could be hosted somewhere else or anywhere.

@aaelony I agree that would be a good thing. If someone contributes it, I'll include it it neanderthal documentation.

@blueberry here https://gitlab.com/alanmarazzi/numpy-vs-neanderthal you can find the Numpy benchmarks

There is a Dockerfile ready-to-go which is the easiest way to run everything

While I was there I implemented PCA as well in Numpy, might be interesting to compare some "real" use of Numpy with Neanderthal

Let me know if everything's ok

@justalanm awesome! Thank you. I'll try this as soon as I can (but not sooner than a few days, since I'm in the middle of something). Since I normally avoid docker. If I install anaconda on arch linux, is that enough to run this? additionally, can you post your results for the "normal" (not 1000x strawman) examples?

@blueberry if you want there is the specs.txt file for reproducing the conda environment

Anyway, you can always basically copy what the Dockerfile does 😁

I'll post my results soon, would you like me to post Neanderthal results as well?

http://gigasquidsoftware.com/blog/2018/07/01/clojure-mxnet-joins-the-apache-mxnet-project/

whoo hoo!

Thanks to all that made it possible 🙂

nice one

@justalanm sure!