While Windows fanatics and Linux fanatics love to interact in some fiery keyboard wars frequently, the makers of the OS themselves perceive the strengths and weaknesses of one another. There is definitely that Microsoft too will get it and in truth, it even encourages customers to check out Linux, albeit the use of its Windows Subsystem of Linux (WSL).

Not simply that, the corporate additionally frequently pushes upgrades that should beef up Linux efficiency or build up its capability. For instance, again in November 2022, Microsoft driven nested virtualization or nested Hyper-V fortify. And in September closing 12 months, Linux boss Linus Torvalds permitted Hyper-V enhancements associated with AMD SEV-SNP visitors and Intel TDX visitors that have been despatched in by way of Microsoft’s Principal Architect at Azure Machine Learning, Wei Liu.

Torvalds has as soon as once more permitted new Hyper-V upgrades from Liu for Linux 6.12 and one in every of them particularly is of extra pastime from the efficiency standpoint which was once introduced by way of Microsoft’s Saurabh Singh Sengar:

Pull Hyper-V updates from Wei Liu:

• Optimize boot time by way of concurrent execution of hv_synic_init() (Saurabh Sengar)
• Use helpers to learn keep watch over registers in hv_snp_boot_ap() (Yosry Ahmed)
• Add reminiscence allocation take a look at in hv_fcopy_start (Zhu Jun)

The patch description explains how the parallelization of the hv_synic_init() serve as can beef up the boot time on processors with many cores and NUMA (non-uniform reminiscence get right of entry to) nodes:

Optimize boot time by way of concurrent execution of hv_synic_init()

Currently on an overly huge device with 1780 CPUs, hv_acpi_init() takes round 3 seconds to finish. This is on account of sequential synic initialization for every CPU carried out by way of hv_synic_init().

Schedule those duties parallelly in order that every CPU executes hv_synic_init() in parallel to take complete benefit of a couple of CPUs.

This answer saves round 2 seconds of boot time on a 1780 CPU device, which is round 66% growth within the current common sense.

Sengar additionally did a up to date presentation about his findings titled “Accelerating Linux Kernel Boot-Up for Large Multi-Core Systems” which delves into the main points of this patch and different optimizations which might be being made. You can in finding it right here at the Linux Plumbers Conference (LPC) web site.

Source: LKML (link1, link2)