Intel® Distribution for Python*
Support and discussions for achieving faster Python* applications and core computational packages.
391 Discussions

How to use Numpy in Intel® Distribution for Python*?

HeMing
Beginner
1,208 Views
  1. My question

I haven't found the technical documentation of Numpy on intelpython3_full. I don't know which classes, methods, and functions of intelpython3_full can be called in python or used in numpy.

I need a function and method documentation of Numpy, pandas, and scipy on intelpython3_full, similar to daal4py's technical documentation.

 (Https://intelpython.github.io/daal4py/contents.html)

  1. Question details

On this page:

https://software.intel.com/content/www/us/en/develop/blogs/faster-random-number-generation-in-intel-...

There is a case:

HeMing_0-1631242238596.png

Cannot find "random_intel" at www.numpy.org.

Refer to: from numpy import random, random_intel usage, I use: from numpy import sin, sin_intel, the result is an error:

HeMing_3-1631243076534.png

Where can I find more function documentation ending in _intel?

  1. Further details

I delete the virtual environment of intelpython3_full, and I can run "random_intel" successfully

HeMing_2-1631242640501.png

Explain that "random_intel" may be a function or class in Numpy.

Are there more functions ending in _intel? Please recommend the documentation link, thank you very much! !

  1. System information:

win10 Professional 20H2,

Anaconda3-2019.10-Windows-x86_64

  1. Work done

According to the intel documentation, I installed intelpython with conda create -n idp intelpython3_full python=3.7.

And I found the technical documentation of daal4py (https://intelpython.github.io/daal4py/contents.html)

By studying this document, I can use daal4py smoothly.

0 Kudos
1 Solution
Huiyan_C_Intel
Employee
338 Views

Hi Heming,

https://www.intel.com/content/www/us/en/developer/articles/technical/faster-random-number-generation..., this article is outdated, random_intel is deprecated, and we use mkl_random instead. The numpy in Intel Distribution for Python is built to use mkl_random by default.


So, to use numpy in Intel Distribution for Python, you just use the same API that standard numpy supports and you can get performance boost, there are no additional numpy packages or numpy APIs.


If you are interested in optimizations in oneMKL, you can refer here: https://www.intel.com/content/www/us/en/develop/documentation/onemkl-linux-developer-guide/top.html


I will close this ticket, feel free to re-open if you have further questions. Thanks!


View solution in original post

5 Replies
RaviranjanK_Intel
Moderator
1,173 Views

Hi,


Thanks for posting in Intel Forums.


Also thanks for your feedback. We are looking into it. We will get back to you soon.


Regards


Raviranjan


HeMing
Beginner
1,145 Views
RaviranjanK_Intel
Moderator
994 Views

Hi,


we had taken your feedback and reported to the internal team. Please let us know if there is any other issue we could help with?


Thanks


AthiraM_Intel
Moderator
548 Views

Hi,


Thank you for your feedback. We have provided your feedback to the relevant team. At this moment there is no visibility when it will be implemented and available for use. Please let me know if we can go ahead and close this case?


Thanks


Huiyan_C_Intel
Employee
339 Views

Hi Heming,

https://www.intel.com/content/www/us/en/developer/articles/technical/faster-random-number-generation..., this article is outdated, random_intel is deprecated, and we use mkl_random instead. The numpy in Intel Distribution for Python is built to use mkl_random by default.


So, to use numpy in Intel Distribution for Python, you just use the same API that standard numpy supports and you can get performance boost, there are no additional numpy packages or numpy APIs.


If you are interested in optimizations in oneMKL, you can refer here: https://www.intel.com/content/www/us/en/develop/documentation/onemkl-linux-developer-guide/top.html


I will close this ticket, feel free to re-open if you have further questions. Thanks!


Reply