On Monday, 4 January 2016 at 01:47:04 UTC, Jack Stouffer wrote:
> On Monday, 4 January 2016 at 01:09:30 UTC, Ilya wrote:
>> To be clear: there is NO data in Article example. Only CPU registers are used. It is not fair. -- Ilya
>
> Ok, I see were I made the mistake, I apologize. I believed that since I was only testing the np.mean line of code, that the lazy generation of the D in the D version would have no effects on the comparison. But I forgot the fact that the lazy generation would mean that no memory needed to be accessed in the D code and therefore it's not apples to apples with the Numpy code.
>
> Thank you for clarifying this, I will update the article.

Please add links to
1. Mir: http://code.dlang.org/packages/mir
2. Documentation: http://dlang.org/phobos-prerelease/std_experimental_ndslice.html
3. DlangScience example

Thanks! -- Ilya

On Monday, 4 January 2016 at 02:01:05 UTC, Ilya wrote:
> Please add links to
> 1. Mir: http://code.dlang.org/packages/mir
> 2. Documentation: http://dlang.org/phobos-prerelease/std_experimental_ndslice.html
> 3. DlangScience example
>
> Thanks! -- Ilya

Updated. I also changed the example slightly from the DlangScience example by increasing the array size because some people on HN were complaining that the size of the data wasn't large enough for the benchmark to be meaningful.