My attention was just drawn to the existence of this product:
Does Intel actually fork the CPython code, re-implement Python in some way, or otherwise specialize Python itself for its own CPUs? Or is this merely their own re-distribution of some kind of installer for Python plus a bunch of scientific-computing/AI libraries? I couldn’t figure out a clear answer for this in all the marketing-speak; I figure this sort of project would benefit from (if not require) collaboration with the dev team, so I hoped someone around here would know.
Whoa!? Is it a breakthrough on the hardware side?
Or is it simply part of Intel’s programming learning curriculum?
There seems to be some attractive words and a video below🥰
Let’s take a look (I don’t have an Intel CPU…)
I think what they are doing is adding the “data parallel” API to python AI and data analytics libraries. Would be surprised if core python had been changed beyond being compiled with Intel x86 compiler.
The “Develop fast, Python* code” irritates me. It sounds really weird. Like “Develop fast Python* code” but with a confusing speedbump preventing it from being (read) fast.
Thanks for posting this link. I wonder - should Python* be pronounced as Python-star, or Python-asterisk
Intel has open-sourced its main data-parallel implementations (Intel Python · GitHub).
Seems this is mainly further optimized Cython code for numpy, and similar libs, interfacing with their “oneAPI” libraries. Seems a pretty interesting project -
