Linux 3.10 kernel officially released; brings better support for SSD caching

Linus Torvalds officially released the latest Linux kernel, version 3.10, late yesterday evening. According to Torvalds, this version of the kernel is said to have one of the biggest changes in years based on commit count and actual number of lines. In his announcement on the Kernel Mailing List, Torvalds noted that he was thinking of marking this as yet another release candidate, but decided against it.

This release brings some great new features and changes that help make Linux more competitive to the other two major operating systems. One of the more prominent features is Bcache, a storage functionality that lets systems equipped with both SSDs and HDDs make use of the fast SSD to cache data that's used frequently, eliminating read-write delays caused by slower devices such as hard disk drives. Bcache is file-system agnostic as it operates on the block device level, which means it doesn't care what file system a hard drive uses since it only reads blocks of data.

The Linux kernel gets one of the biggest changes in years

The new kernel also adds an assortment of power management improvements for AMD processors. Better support for AMD's frequency sensitive power-save bias has been added, which allows the system to make more power-efficient frequency changes according to feedback from the hardware. Systems with AMD's processors should benefit from this.

Support for ARM’s big.LITTLE architecture has also been improved. This architecture pairs low-power processor cores with high-power ones to better meet processing and power-consumption demands, and Linux 3.10 adds a cpufreq driver to handle frequency switching. The Samsung Galaxy S4’s Exynos 5 Octa chip follows this architecture.

Linux 3.10 also brings better support for Intel’s new Haswell line of processors, besides a few optimisations for better performance.

This Linux release adds support for timer-free multitasking, which lowers the timer interrupt’s trigger frequency to one interrupt per second from 1,000 interrupts per second, reducing the interrupt processing load on the system. The increase in performance is not really noticeable during everyday usage, but performance-intensive applications such as scientific calculations will run faster.