An nameless reader shared this document from ZDNet:
On the Linux Plumbers Conference, the invite-simplest assembly for the head Linux kernel developers, ByteDance Linux Kernel Engineer Cong Wang, proposed that we use AI and machine learning to tune the Linux kernel for basically the most outcomes for suppose workloads… There are hundreds of parameters. Even for a Linux professional, tuning them for optimum efficiency is a lengthy, hard job. And, obviously, varied workloads require varied tunings for quite loads of items of Linux kernel parameters… What ByteDance is engaged on is a first strive to automate the entire Linux kernel parameter tuning job with minimal engineering efforts.
Particularly, ByteDance is engaged on tuning Linux reminiscence administration. ByteDance has stumbled on that with machine learning algorithms, comparable to Bayesian optimization, computerized tuning may per chance per chance well additionally even beat most Linux kernel engineers. Why? Correctly, the foundation, as Wang wryly establish it, “is to not establish Linux kernel engineers out of trade.” No, the aim is “to liberate human engineers from tuning efficiency for every particular person workload. Whereas making better choices with historic knowledge, which contributors in general combat with. And, final, but by no methodology least, secure better solutions than those we attain up with the usage of our recent trial and error, heuristic systems.
Briefly, ByteDance’s machine optimizes helpful resource usage by making real-time adjustments to issues esteem CPU frequency scaling and reminiscence administration.
Read extra of this yarn at Slashdot.
—
Online:
Slashdot details agency contributed to this document, published by ORDO Files editors.
Contact us: [email protected]
Our Requirements, Terms of Use: Traditional Terms And Conditions.
To keep away with any confusion creating from varied time zones and daylight saving adjustments, all instances displayed on our platforms are in Coordinated Universal Time (UTC).