3 ms·
Catching Up with Intel's Python and AI Work at PyCon [audio]
- neves 9y agoIntel Distribution for Python is really cool. Free speed up of code, without the need to do anything: https://software.intel.com/en-us/distribution-for-python/features https://software.intel.com/en-us/distribution-for-python/fea...
- tapia 9y agoWow! How did I never hear of this?! There should be a Package in the AUR (Arch User Repository) for this.
- gigatexal 9y agoI think there's a license involved.
- tapia 9y agoThe AUR only defines rules for installing certain packages, the program itself would still need to be downloaded from Intel. That is how it's done with the MKL package [1] and other proprietary software. [1] https://aur.archlinux.org/packages/intel-mkl https://aur.archlinux.org/packages/intel-mkl
- thearn4 9y agoIs this an alternative to the cpython interpreter, or is it just a bundle with numpy-backed MKL? Not that it wouldn't be handy, but anaconda python does provide this too.
- marmaduke 9y agoIt's the same stack, but compiled with Intels compilers against Intel math libraries (mkl and teh others). If you use standard algorithms it's a good choice.
- tarlinian 9y agoIt's still CPython but many of the scientific software stack pieces (NumPy/SciPy/SKLearn) have been re-written to call Intel optimized libraries (MKL/DAAL mainly).
- thearn4 9y agoSounds interesting, I'll have to give it a shot then!
- neves 9y agoThe same CPython, just with a compiler that makes better user of Intel chips instructions. It is almost the same of Anaconda distribution that has all the math libs included.
- sonium 9y ago> Free speed up of code Well... your paying for the hardware.