↩ Accueil

Vue lecture

Google Axion CPU Performance With The New Google Cloud N4A Instances

Back in 2024 Google rolled out their Axion in-house ARM processors with the Google Cloud C4A instance type. Today they are expanding their Axion offerings in Google Cloud with the N4A instances now out of preview. The Google Cloud N4A instances are designed for scale-out web servers and microservices, containerized applications, back-end application services, databases, data analytics, and cost-effective development/staging/testing environments.
  •  

Systemd Founder Lennart Poettering Announces Amutable Company

Systemd founder and lead developer Lennart Poettering announced the creation of a new company called Amutable. The Amutable company being led by Chris Kühl (CEO), Christian Brauner (CTO) and Lennart Poettering (Chief Engineer) will be focused on delivering determinism and verifiable integrity to Linux systems...
  •  

New Intel Linux Code For DG2 Graphics Can Improve Performance As Much As "A Whopping 260%"

A set of 18 patches were merged overnight to Mesa 26.1 for working around graphics corruption on Meteor Lake and DG2/Alchemist class graphics hardware. Not only are some graphics corruption issues worked around but for that hardware there is as much as "a whopping 260%" performance improvement observed for some graphics workloads...
  •  

ThinkPads On Linux Appear Nearly Ready For Improved Trackpoint Doubletap Handling

Being worked on for a while by Lenovo engineer Vishnu Sankar is nicely handling support for double-tap functionality with TrackPoints on ThinkPads under Linux. The sixth iteration of this enablement work was posted today and is just documentation updates, so it's looking like this new TrackPoint doubletap code could soon be crossing the threshold for the mainline Linux kernel...
  •  

Updated Linux Patches For Managing Out-Of-Memory Behavior Via BPF

Being worked on since last year by Google engineer Roman Gushchin was the latest attempt for the Linux kernel to support managing the out-of-memory "OOM" behavior using BPF programs. It's been a while since there has been anything new to report on that front but published overnight is the latest iteration of those patches...
  •  

AMD Squeezing Out More More ROCm/HIP Performance With New Device-Side PGO

Compiler profile guided optimization (PGO) techniques have paid off well for increasing CPU performance via application/workload-specific profiles fed back to the compiler to make more informed decisions. AMD compiler engineers have been working on crafting device-side PGO for their AMDGPU LLVM back-end for allowing ROCm/HIP workloads to achieve greater GPU performance. An initial merge request is now open for upstream LLVM...
  •  

Intel Panther Lake / Arc B390 Linux Benchmarks Still Coming

Ahead of tomorrow's official availability of new Intel Core Ultra Series 3 "Panther Lake" laptops, the review embargo lifted on Panther Lake and its much anticipated Arc B390 graphics. There have been several Windows 11 reviews of Panther Lake out today, but what about Linux?..
  •  

Revisiting The Linux 6.19 Performance With Some New Wins & Losses

Back at the start of the Linux 6.19 kernel cycle I ran benchmarks showing some scheduler performance regressions with the new kernel. Fortunately, two weeks out from the Linux 6.19 stable release, merged this weekend was disabling the scheduler's NEXT_BUDDY feature due to performance regressions. Here are some fresh benchmarks looking at the latest Linux 6.19 Git state with/without NEXT_BUDDY and comparing it to Linux 6.18 stable for reference.
  •  

AMDGPU Patches Updated For HDMI Gaming Features On Linux With Radeon Graphics

A patch series posted last week for the open-source AMDGPU kernel driver implements HDMI Variable Rate Refresh "VRR" and other gaming features for HDMI displays. With the HDMI Forum blocking HDMI 2.1 open-source support, these HDMI gaming features for the AMDGPU driver were developed via trial-and-error and the limited public knowledge available. A second iteration of these patches are now available for testing...
  •  
❌