Call xxxI instead of xxx to support strided layout. (2nd take)

[N5N3]

Replace #54 (my local git history get messed).

[Crown421]

Very nice, great improvement. Could you increment the version in the Project.toml as well?

Unfortunately there is currently an issue with the TagBot. Will try again tomorrow to fix it and then merge.

[Crown421]

Ah, could you perhaps add a few words and maybe an example to the Readme about the new capabilities for strided arrays?

[N5N3]

Well the newly added funs's accuracy looks terrible on mac (Not supprising though.)
Should we only define them on linux/windows? I have no mac at hand so it's hard for me to test which func is broken.

[Crown421]

Well the newly added funs's accuracy looks terrible on mac (Not supprising though.) Should we only define them on linux/windows? I have no mac at hand so it's hard for me to test which func is broken.

Are you referring to the failed tests on CI or something in addition? I have a Mac in my office, will check it out once I go back in. Since it only fails on nightly, I think there is a bit of time, might even get fixed somewhere in Base.

[N5N3]

Are you referring to the failed tests on CI or something in addition?

Yes.

Since it only fails on nightly, I think there is a bit of time, might even get fixed somewhere in Base.

I'm not sure if this is related to the version of MKL_JLL . If not, then perhaps this is bug in Base .

[Crown421]

Will check separately, especially since the same issue is already on main. Tag bot should be working now, so I will merge this one.

[Crown421]

Docs added, ready to merge.

[Crown421]

Improved the test display, turns out its the gamma function that is failing on mac:
https://github.com/JuliaMath/IntelVectorMath.jl/actions/runs/3832441706/jobs/6522750253#step:6:353