The V3DV open-source Vulkan driver being developed by consulting firm Igalia for the Raspberry Pi Foundation is nearing official support for Vulkan 1.0.

V3DV is the modern Vulkan driver for current Broadcom SoCs catering to the Raspberry Pi 4 single board computer and presumably future generations as well. Igalia engineers have been working to wrap up support for Vulkan robust buffer access and multi-sampling with the V3DV code in recent weeks. That went well and more Vulkan demos are now running on the Raspberry Pi

Proposed earlier this summer for GNOME's Mutter was the idea of triple buffering the desktop when the GPU is running behind in order to ideally cause that extra load to ramp up the GPU clock frequencies in order to in turn get back on track with rendering the desktop on-time. A third version of that work is now brewing albeit too late to see with the imminent GNOME 3.38.0 release.

Canonical's Daniel van Vugt has been working on this dynamic triple buffering for GNOME's Mutter

While AV1 adoption is still taking off and finally seeing desktop hardware with AV1 decode , given the time it takes to develop a new high-end video codec it shouldn't come as too much surprise that "AV2" is already being explored.

A Phoronix reader tipped us off this morning that the aomedia.org roadmap for the Alliance for Open Media industry consortium that spearheads AV1 has already started with " R&D for AV2 has begun in earnest ."

There is also already a AV

Intel's open-source SVT-AV1 encoder that is now being used as the basis for the AV1 Software Working Group and used by Netflix and other organizations for being one of the most performant CPU-based AV1 encoders is out with a new release.

SVT-AV1 0.8.5 is out this morning with a few improvements on the encoder front. SVT-AV1 0.8.5 adds multi-staged HME support, new quality speed trade-offs tuned for video

Vulkan 1.2.153 doesn't come with any new extensions or major changes aside from an important fundamental change if you are interacting with their Git repository.

Vulkan 1.2.153 is out this morning with just the usual assortment of documentation fixes/clarifications. About the most noteworthy on that front is adding some developer documentation to the appendices for several extensions like VK_EXT_memory_budget but overall it's a fairly mundane release.

But if you rely on any scripts

Gentoo's Portage 3.0 package manager has been stabilized and ready to shine.

Portage 3.0 removes Python 2.7 support as a significant change but more pressing to end-users is faster dependency calculations. Portage 3.0 should be 50~60% faster for dependency calculations.

This big speed-up comes by making use of the lru_cache functionality in Python 3.2 and newer. The developer who worked on the support found that emerge now dropped from about five and a

Linus Torvalds took some time out of his US Labor Day weekend to issue the fourth weekly release candidate of the Linux 5.9 kernel.

Even with the US holiday kicking off at the weekend, Linux 5.9-rc4 ticked up in activity but that is in part because 5.9-rc3 was smaller than normal due to timing of pull requests.

Linus noted in today's 5.9-rc4 announcement, " rc4 has all the pieces that were missing from rc3,

A few days ago I published a deep dive into the CPU and GPU performance with Blender 2.90 as a major update to this open-source 3D modeling software. Following that I kept on testing more and older NVIDIA GPUs with the CUDA and OptiX back-end targets to now have an 18-way comparison from Maxwell to Turing with the new Blender 2.90.

Given the Blender 2.90 performance changes over Blender 2.8x as outlined in the earlier article, here

Queued up in char-misc-next ahead of the Linux 5.10 cycle is a speed-up for reading from /dev/zero ...

The patch adds a non-iov_iter version of reads for the /dev/zero interface. The write performance is unchanged as it already had a non-iov_iter implementation.

The patch by prominent upstream kernel developer Christoph Hellwig explains that in fact there is a "major speedup" and it's attributed to the iov_iter overhead