Longtime NVIDIA engineer Thierry Reding who has been involved with the open-source Nouveau driver efforts largely from an embedded/mobile Tegra angle last week sent out the newest patch series.

The latest work by Reding is focused on supporting synchronization file descriptors (sync FDs) and synchronization objects (syncobj) with the Nouveau DRM kernel driver.

A new kernel interface is proposed for submitting push buffers that can optionally return a sync FD or sync object to user-space. These synchronization FDs/objects are practical for synchronization operations between the

Earlier this summer building off the latest Windows Subsystem for Linux 2 advancements by Microsoft, NVIDIA released early support for CUDA / GPU compute on WSL2 . This week NVIDIA offered up a new version of their CUDA WSL support.

The updated NVIDIA CUDA implementation with Windows Subsystem for Linux brings better performance particularly for smaller workloads, DirectML API for DirectX 12 GPU acceleration, and support for PTX JIT. The PTX JIT support allows developers to run the PTX representation on WSL and loads from the driver store directly.

The better performance

Going back a number of months have been patches for bringing up the Broadcom BCM2711 display pipeline as found with the Raspberry Pi 4 SBC. That work still hasn't been mainlined but a fifth round of patches has now been sent out for review.

Eighty patches hit the kernel mailing list this morning for working on this BCM2711 display pipeline support for the VC4 DRM driver. The most notable beneficiary to this work is the Raspberry Pi 4. Due to this SoC now having two

Open-source consulting firm KDAB has released Hotspot 1.3 as their GUI utility for visualizing Linux perf reports.

Hotspot is a utility for consuming data generated by Linux's perf subsystem and serves as a replacement to the command-line perf report functionality.

Hotspot 1.3 brings improvements for dealing with large and complex perf.data files, it is much faster than previous releases, and also properly supports perf analysis data that has been compressed using Zstd. On the user-interface side the Qt application has improved

One of the most interesting projects out of Google Summer of Code 2020 has been the ongoing work for allowing individual code files to be compiled in parallel , building off work last year in addressing GCC parallelization bottlenecks . The final report for GSoC 2020 on this work has been issued.

This work has been focused on being able to parallelize the compilation of large source files compared to traditionally with making use of multiple jobs from the build system for compiling multiple files in parallel. For large source files and

In addition to Intel talking a lot today in virtual briefings about the new 11th Gen "Tiger Lake" mobile processors , they were also talking at a higher level extensively on their second-generation "Project Athena" laptop innovation program and their new Intel Evo branding for premium laptops.

Project Athena has been about encouraging their OEM partners to design thin-and-light laptops that perform well and meet a number of different criteria from battery life to bezel thickness and other hardware/software attributes. With Intel Evo they

After a small dip in July , how did the Steam on Linux gaming marketshare end out for August prior to many gamers returning to school and others still being isolated at home? A small uptick but still under the 1% threshold.

Valve just updated their Steam Survey with the August 2020 numbers. From July's 0.86% marketshare, Valve's survey is reporting a 0.03% increase for Linux gaming putting the overall percentage now at 0.89%.

In

While Mesa 20.2 will hopefully be out next week, Mesa 20.1.7 is out today as the newest stable release for this collection of open-source Linux graphics drivers.

With Mesa 20.1.7 there still are a fair number of changes even with being quite late in the cycle for this Q2'2020 driver series. Among the highlights with Mesa 20.1.7 are:

- A new option for clamping division by zero and through the

NVIDIA software engineer Zi Yan who specializes in the Linux kernel memory management subsystem today sent out a set of patches proposing the addition of 1GB THP support for the Linux kernel.

The NVIDIA engineer is proposing 1GB (transparent hugepages support for Linux on x86_64 hardware for being more flexible in reducing translation overhead and increasing the performance of applications that have a large memory footprint. Unlike Linux's huge page-tables (HugeTLB) code for supporting large memory footprints, the 1GB THP approach doesn