Sent in last week were some AMDGPU "fixes" for Linux 5.12. While there are some fixes as part of the series, there are some new (minor) features enabled.

In addition to the previously covered DRM graphics driver features for Linux 5.12 like the Radeon RX 6000 series OverDrive , there are some more patches that were sent in last Thursday as fixes for Linux 5.12.

From that pull are some more AMDGPU changes worth mentioning for this next version of

Dav1d is already the most performant and leading AV1 software decoder we have seen while out today is v0.8.2 that should speed-up the video decode process even more on modern x86/x86_64 and ARM hardware.

While dav1d 0.8.2 is "just a point release", it does pack some interesting performance optimizations for today's hardware. On the ARM front there are ARM32 optimizations to speed up loop restoration and

Greg Kroah-Hartman sent in the staging/IIO updates for the Linux 5.12 kernel and this time around are lighter than normal but still with a few interesting items worth mentioning.

Greg noted in the pull request , " Nothing really huge in here, the number of staging tree patches has gone down for a bit, maybe there's only so much churn to happen in here at the moment. " At least a few items still worth mentioning for staging/IIO in Linux 5.12.

Following the recent major rewrite to KDE's KWin compositor code there are more exciting improvements likely to come for KWin in improving its Wayland compositor support.

KDE developer Xaver Hugl has work-in-progress code striving for Variable Refresh Rate (VRR, a.k.a. Adaptive-Sync / FreeSync) support for the KWin Wayland compositor. Additionally, the tentative patches also aim to get the compositor-induced latency down to around ~1-2ms when using direct scanout with Vsync but without VRR.

This

The block subsystem and related storage changes were merged today for the in-development Linux 5.12 kernel.

The IO_uring changes for Linux 5.12 continue to be quite prominent for this very compelling feature of the Linux kernel. In particular, continued work on making IO_uring even faster . With request recycling and task_work optimizations, IO_uring with Linux 5.12 is now in the range of 10% to 20% faster for workloads that are mostly inline.

Earlier this month at the virtual FOSDEM 2021 conference was an interesting presentation on how European developers are preparing for AMD-powered supercomputers and beginning to figure out the best approaches for converting existing NVIDIA CUDA GPU code to run on Radeon GPUs as well as whether writing new GPU-focused code with OpenMP device offload is worthwhile.

Georgios Markomanolis who is the lead HPC scientist at CSC was the one presenting at this month's FOSDEM virtual conference on preparing for supercomputers with AMD GPUs. Their focus at the moment is

Oracle continues maintaining Solaris 11.4 with monthly stable release updates but there still is no public sign of anything past 11.4 for this operating system that was once exciting during the Sun Microsystems days. But with this week's 11.4 SRU30 release, at least there are many package updates.

Oracle this past week released Solaris 11.4 Stable Release Update 30 in it being nearly three years already since Solaris 11.4 was introduced. There are a number of

There is a lot of new hardware enablement with the ARM platforms and DeviceTree additions for the Linux 5.12 kernel merge window.

The various kernel pull requests from Arnd Bergmann were merged on Saturday night following the nearly week long delay due to the PNW winter weather thwarting the start of the merge window . Some of the ARM/DT highlights for Linux 5.12 include:

- Lightening up the kernel by some 21.4k lines is the dropping of many old/obsolete ARM platforms . Nuked this

Mesa's on-disk shader cache, which is used for speeding up game load times by avoiding the redundant recompiling of shaders on successive loads and also helping performance for software that compiles shaders on-the-fly, is seeing a big improvement with Mesa 21.1 . Mesa 21.1-devel merged this weekend the new single file cache implementation.

Mesa's new single file shader cache has been in the works by Valve contractor Timothy Arceri the past few months. The single file cache doesn