Tag Archives: GNU/Linux benchmarks

Wine 4.18 Released With Many Bug Fixes, Some Feature Work


WINE --

While three weeks have passed since the previous Wine development release compared to the usual two week cadence, Wine 4.18 is out today and isn’t too busy on the feature front but there are more than three dozen bug fixes.

The delay and Wine 4.18 not being particularly big appear to be due to WineConf taking place last week in Toronto keeping many of the developers busy. New Wine 4.18 feature work includes implementing more VBScript functions, cleanups/improvements to the Apple macOS Quartz code, and fixes for test case failures.

On the bug fixing front there are 38 known bug fixes for Wine 4.18 and include addressing items in games like The Witcher 2, Lego: Stunt Rally, Lego Island 2, World of Warcraft, Halo 2, and Need for Speed: Carbon.

The list of Wine 4.18 changes can be found via WineHQ.org. Wine 5.0 meanwhile should be coming up as the next stable release of Wine in early 2020.


Fedora 31 Release Held Up By Installer + DNF Bugs


FEDORA --

Fedora developers had been trying to ship Fedora 31 for their original release target of next Tuesday, 22 October, but that isn’t going to happen due to remaining blocker bugs.

At today’s meeting they decided F31 is a “No-Go” due to open issues.

Among the blocker bugs at this point are upgrade issues over libgit2.so.28()(64bit) not being found, a DNF exception around armv7hcnl, a dnf-yum upgrade issue from Fedora 30, dnfdragora having complaints about a locked process after updates, and a Fedora 31 KDE Live graphics issue when booting in BIOS mode, among other proposed blockers.

The current Fedora 31 blockers are outlined here.

Fedora developers will again meet next week to see if Fedora 31 is then ready for release on their back-up target of 29 October or we’ll be challenged by more weeks of delays.


Libdrm 2.4.100 Released With Bits For Intel Elkhart Lake, Tiger Lake Graphics


MESA --

AMD open-source developer Marek Olšák on Wednesday released libdrm 2.4.100 as the newest feature update to this Mesa DRM library.

On the AMD front there are a number of RAS tests added, a new amdgpu_cs_query_reset_state2 interface, and other expanded AMDGPU test coverage.

Meanwhile over on the Intel front they have added the necessary PCI IDs for bringing up support for Gen11 graphics bearing Elkhart Lake as well as the initial bits for Tiger Lake Gen12 / Xe graphics. Of course, most of the interesting driver enablement work happens within Intel’s DRM kernel driver and Mesa OpenGL/Vulkan drivers while libdrm is just the tiny shim between the two needing PCI IDs, etc.

Libdrm 2.4.100 was also going to drop GNU Autotools support in favor of Meson, but at the last minute the Autotools support was restored by Marek at least for this release while Meson continues to be offered. The libdrm 2.4.100 release also has a few Meson build system updates and xf86drm clean-ups/fixes.

The list of libdrm 2.4.100 changes can be found via the release announcement.

You may recall there was recently interest in changing the Mesa DRM library versioning scheme but that didn’t go into effect for Wednesday’s release. While there was support for a year.number versioning scheme similar to Mesa itself, ultimately Marek said he would go with a 2.year.number versioning scheme. That was done at the request of Intel developers to allow for a major version bump should their driver interfaces ever radically break and need to change the library versioning number.


Arch Linux Nears Roll-Out Of Zstd Compressed Packages For Faster Pacman Installs


OPERATING SYSTEMS --

The upcoming release of Arch’s Pacman 5.2 is bringing support for compressing packages with Zstd which ultimately will provide faster package installs on Arch Linux.

Similar to other Linux distributions beginning to make use of Facebook’s Zstd (Zstandard) compression algorithm for faster compression/decompression of packages, Arch Linux is doing the same. Their findings mirror that of others in allowing faster compression/decompression performance with a similar compression ratio for binaries to that of XZ.

With Pacman 5.2, the Zstd package compression is being rolled out for Arch Linux while ahead of that is a bulletin today that users need a libarchive within the past year. Since September 2018 has been a libarchive on Arch Linux supporting Zstd, but if for some reason you are running an Arch Linux install with a very outdated libarchive, you will want to update soon to avoid issues moving forward.

More details on this notice via ArchLinux.org.


Highly Threaded Linux Software Running Under CFS Quotas See Big Performance Fix


LINUX KERNEL --

Thanks to a Linux kernel fix that is likely to be back-ported to the various stable series, highly threaded software running under CFS quotas for enforcing CPU limits are about to be much faster. At least in a synthetic test case, the kernel fix yields a 30x improvement in performance.

Spotted by the Kubernetes community but affecting others with highly threaded workloads and making use of a CFS quota to restricted shared CPU resources, it turns out that highly-threaded applications are routinely not getting “their fair share” of the CPU, leading to lower than expected performance and higher latency.

This has been a known bug for more than one year and a kernel bug report on unexpected CFS throttling since late 2017. The issue is believed to be recently fixed up for mainline Linux 5.4 and pending for back-ports after the patch was volleyed around the kernel mailing list for a few months.

There is the fix that is a few dozen lines of code that removes the expiration of CPU-local slices:

It has been observed, that highly-threaded, non-cpu-bound applications running under cpu.cfs_quota_us constraints can hit a high percentage of periods throttled while simultaneously not consuming the allocated amount of quota. This use case is typical of user-interactive non-cpu bound applications, such as those running in kubernetes or mesos when run on multiple cpu cores.

This greatly improves performance of high-thread-count, non-cpu bound applications with low cfs_quota_us allocation on high-core-count machines. In the case of an artificial testcase (10ms/100ms of quota on 80 CPU machine), this commit resulted in almost 30x performance improvement, while still maintaining correct cpu quota restrictions.

Thanks to Phoronix reader Mark for pointing out this recent kernel change.