The Google Pixel 4a smartphones launched in mid-2020 and now in mid-2025 it looks like we might finally be close to seeing mainline kernel support for the Pixel 4a devices and other hardware making use of Qualcomm Snapdragon 730/730G/732G SoCs.

Last week I published some initial benchmarks of the Amazon/AWS Graviton4 processors now available within the EC2 cloud using the new "R8g" instances. That initial comparison was a 64 vCPU comparison of Graviton4 against AMD EPYC and Intel Xeon 64 vCPU AWS instances. In today's article is a look at the 96-core Graviton4 bare metal performance using the "r8g.metal-24xl" AWS instance type. The Graviton4 r8g.metal-24xl performance was then compared in today's article against various bare metal AMD EPYC, Ampere Altra Max, and Intel Xeon processors in the lab at Phoronix.

Following recent Vulkan 1.1 support within Mesa for the PanVK driver for open-source Arm Mali Vulkan driver support, Vulkan 1.2 is now being advertised.

This week Intel announced "200S Boost" for Core Ultra "Arrow Lake" K-Series desktop processors as effectively a new overclocking profile rolling out to existing Z890 motherboards via a BIOS update. Enabling the 200S Boost profile is said to help with low-latency workloads like gaming by allowing higher fabric / die-to-die / memory frequencies. While some Windows benchmarks have begun emerging for the Intel 200S Boost mode and some limited gains, I was curious about the performance under Linux so here are some 200S Boost benchmarks with the Core Ultra 9 285K on Ubuntu 25.04.

Linus Torvalds is sharing some of his classic and straight-to-the-point wisdom today over file-systems with case-folding / case-insensitive file and folder support.

While last week Intel released an update Compute Runtime for GPU compute with the OpenCL and Level Zero APIs on Windows and Linux, today they released a new preview version for readying a shiny new feature: Ultra Low Latency Scheduling "ULLS" for Lunar Lake Xe2 graphics.

GCC 15.1 was just released as the newest annual feature release to the GNU Compiler Collection. This first stable GCC 15 release brings a COBOL compiler front-end, many C and C++ language support improvements, support for new CPUs and ISA capabilities, better Rust programming language support, debugging enhancements, and a whole lot more.

A set of Linux kernel patches posted today by longtime Linux kernel developer Ingo Molnar are looking to remove support for "ancient" 32-bit CPUs. In particular, if these patches are accepted, the Linux kernel would be ending support for old i486 CPUs as well as early i586 CPU models.

Nearly two years ago patches for casefolding / case insensitive file and folder support on Bcachefs were posted by a Valve/Linux developer. That support was upstreamed into the Bcachefs kernel driver but it turns out that it never properly worked. Patches now set for merging into the Linux 6.15 will fix that case insensitive file/folder opt-in support so that it is now properly supported.

One of the interesting new features merged to the Linux kernel last year was the DRM Panic infrastructure so that Linux can display an error screen akin to Windows' "Blue Screen of Death" when encountering problems. With follow-on kernel releases it's been extended to add QR code error messages and other improvements. But DRM Panic does require the support/cooperation of the different Direct Rendering Manager drivers and so far Intel graphics haven't been supported.

Intel today released a new version of the Intel Extension for PyTorch in order to apply optimizations to PyTorch for benefiting Intel's hardware. With the Intel Extension for PyTorch v2.7 release, there is support for new large language models (LLMs) as well as various performance optimizations and other enhancements.

Following the COSMIC Alpha 6 release from February, System76 today released COSMIC Alpha 7 as their last planned alpha release for this open-source, Rust-written desktop environment designed around the needs of their Pop!_OS Linux distribution.

The Framework Laptop 13 with AMD Strix Point is now shipping that as detailed in our review earlier this month can provide for a very capable Linux laptop for Linux developers, creators, and enthusiasts. But for those hesitant about the high price and still weeks away before they have shipped all their pre-orders, if you are principally concerned about battery life, and/or after proven build quality backed by on-site warranty and other warranty/support options, the Lenovo ThinkPad X1 Carbon Gen 13 Aura Edition ends up being a solid option for a very reliable and well-engineered laptop for Linux use. Here is a look at the Lenovo ThinkPad X1 Carbon Gen 13 Aura Edition on Linux that is powered by Intel Lunar Lake.

A new software project covered on Phoronix last year was SCALE for natively compiling CUDA applications for AMD GPUs. This "clean room" implementation of CUDA building off the open-source LLVM codebase continues going strong and out this week is SCALE 1.3 with more features and hardware support for compiling CUDA software for AMD GPU execution.

Completely separate from the big performance regression I noted earlier this week for the Linux 6.15 Git kernel and fixed yesterday in the upstream codebase, another significant performance issue was also uncovered and fixed this week in Linux 6.15 Git.

The latest bit of hardware enablement coming out of Asahi Linux and queued for introduction in the mainline kernel the next cycle is PCI Express (PCIe) support for the Apple M2 Pro SoC.

New Linux kernel patches have been posted adding the necessary Device Tree files so that Linux is able to boot on the Qualcomm Snapdragon X1 powered Dell Inspiron 14 Plus laptop.

Mesa has supported on-disk shader cache for years to help speed-up game load times and overall system efficiencies. They had shifted from a multi-file cache layout to a single file cache for greater space savings. Steam also added support for the single-file cache. But now upstream Mesa is shifting back from the single-file cache default to the multi-file cache over performance issues.

At the end of last year the upstreaming efforts began for a Raspberry Pi HEVC decoder driver for getting H.265/HEVC accelerated decode working on the Raspberry Pi 4 and Raspberry Pi 5 single-board computers with the mainline kernel. Nearly a half-year later that effort is still ongoing but yesterday brought the third iteration of this driver.