topic Re:Incompatible numpy and scipy versions in the Intel channel in Intel® Distribution for Python*

Incompatible numpy and scipy versions in the Intel channel

Orion_P_ — Tue, 02 May 2023 21:36:05 GMT

$ conda list 'numpy|scipy' 
# packages in environment at /opt/anaconda/envs/scipy39-intel: 
# 
# Name                    Version                   Build  Channel 
numpy                     1.23.5           py39h52df89b_7    intel 
numpy-base                1.23.5           py39ha03f565_7    intel 
numpydoc                  1.5.0            py39h06a4308_0    main 
scipy                     1.7.3            py39h4ca98da_8    intel
$ python -c 'import scipy' 
/opt/anaconda/envs/scipy39-intel/lib/python3.9/site-packages/scipy/__init__.py:146: UserWarning: A
NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy (detected version 1.23.5 
 warnings.warn(f"A NumPy version >={np_minversion} and <{np_maxversion}"

There doesn't appear to be a newer scipy than 1.7.3 in the intel channel.

Re: Incompatible numpy and scipy versions in the Intel channel

Rahila_T_Intel — Thu, 04 May 2023 06:02:45 GMT

Hi,

Thank you for posting in Intel communities.

Thanks for reporting this issue. We were able to reproduce it and we have informed the concerned team about it.

Thanks

Re:Incompatible numpy and scipy versions in the Intel channel

Aditya18 — Sat, 06 May 2023 17:18:32 GMT

Hi,

Thanks for sharing. These Scipy and Numpy are optimized with mkl. And in general, mkl optimized libraries have to make sure that other components especially for GPU part(like dppn, dpctl, and numba-dpex) of Intel Distribution for Python are compatible with each other.

And this warning and compatibility is there in official Scipy docs.

This is just a warning. If they are in the Intel channel, they are tested and part of Intel Distribution of Python. We'll check more and confirm the Latest Scipy version release plan.

Please feel free to share if there is any issue apart from the warning.