Language Selection

English French German Italian Portuguese Spanish

Kernel: AMD, Concurrency Cheat Sheet, Habana Labs, and Mesa

Filed under
Linux

  • AMD Secure Memory Encryption Has a Flaw, Now Disabled by Default in Linux Kernel

    According to a report from Phoronix, the Linux 5.15 kernel is receiving a new fix that involves disabling AMD's Secure Memory Encryption, or SME. This feature is normally enabled by default, but due to unexpected boot failures on some AMD machines, SME will now be disabled by default. Devs will update the Linux 5.15 kernel first, but the change will also move to prior kernels.

    AMD Secure Memory Encryption is a feature exposed to AMD's EPYC and Ryzen Pro processors that allows the CPUs to encrypt the memory at a hardware level. AMD says the feature offers no significant impact on system performance and works with any OS and application because it's hardware-accelerated and doesn't rely upon software.

  • [Older] Linux Kernel Concurrency Cheat Sheet

    Navigating Linux kernel APIs can be very time-consuming, so Linux ksplice guru Vegard Nossum put together a very handy time-saving cheat sheet to help traversing Linux kernel concurrency primitives.

    As a child, I spent countless hours playing games on my Nintendo Game Boy. As I grew older and started programming in QBasic on MS-DOS, I also got curious about how Game Boy games worked.

  • Habana Labs For Linux 5.16 To Bring Peer-To-Peer Support With DMA-BUF - Phoronix

    The drama around DMA-BUF code for the Habana Labs AI driver appears to be wrapping up with the Linux 5.16 cycle that is coming up.

    The Habana Labs driver changes were mailed out today for queuing in char/misc ahead of next month's Linux 5.16 merge window. This driver for supporting the AI inference and training accelerators at the Intel-owned company has some exciting updates for this next kernel release.

  • mesa recompiled with gallium drivers

    The guys have been reporting slow Nvidia performance, for example a report from forum member Keef here:

    https://forum.puppylinux.com/viewtopic.php?p=39261#p39261

    I had compiled all of the packages in OpenEmbedded without libllvm, as it is a hug library, about 60MB. However, it meant had to leave out some drivers in the 'mesa' package.

  • Mesa 22.0 Lands Some Patches Toward OpenCL Image Support - Phoronix

    Mesa 22.0-devel is one step closer to having OpenCL image support that is sought after by many individuals for allowing more OpenCL-enabled desktop software to work nicely with this open-source OpenCL component in Mesa.

    For nearly a year there has been the work-in-progress merge request implementing OpenCL 1.2 image support. That work was started by Red Hat's Karol Herbst.

More in Tux Machines

Videos/Shows: GNOME Web, Wayland, Pipewire, Destination Linux

EasyOS version 3.1.13 released

Version 3.1.13 does not have SeaMonkey, instead has BlueGriffon HTML editor and Balsa email client. Let me know how they go. If any problems, they can be changed. I am planning to create Claws and Sylpheed email client PETs for comparison, as Balsa spam filtering seems a bit weak. I received an email that samba is still broken in 3.1.11. I didn't get time to check that one. So that is something we need to look at. I tested booting up in RAM (& optionally disable drives). Works fine, there is one error message during bootup about a missing file, but that doesn't matter. Getting closer to the next major release, that will be 3.2. Read more

today's leftovers

  • Learn more about distributed databases with ShardingSphere

    Apache ShardingSphere is an open source distributed database, plus an ecosystem users and developers need for their database to provide a customized and cloud-native experience. In the three years since it joined the Apache Foundation, the ShardingSphere core team has worked hard with the community to create an open source, robust, and distributed database and a supporting ecosystem. ShardingSphere doesn't quite fit into the usual industry mold of a simple distributed database middleware solution. ShardingSphere recreates the distributed pluggable system, enabling actual user implementation scenarios to thrive and contributing valuable solutions to the community and the database industry.

  • AWS DocumentDB not MongoDB-compatible, says MongoDB Inc
  • Winners in the Month of LibreOffice, November 2021!

    At the start of November, we revved up a new Month of LibreOffice, celebrating community contributions all across the project. We do these every six months – so how many people got sticker packs this time? Check it out…

Programming Leftovers

  • Cracking the Spotify Code

    Spotify offers a little picture that, when scanned, opens almost anything searchable with Spotify. Several lines are centered on the Spotify logo with eight different heights, storing information in octal. Many visual encoding schemes encode some URI (Uniform Resource Identifier) that provides a unique identifier for that specific song, album, or artist when decoded. Since many URIs on Spotify are pretty long (one example being spotify :show:3NRV0mhZa8xeRT0EyLPaIp which clocks in at 218 bits), some mechanism is needed to compress the URIs down to something more manageable. Enter the media reference, a short sequence encoding a specific URI, generally under 40 bits. The reference is just a lookup in a database that Spotify maintains, so it requires a network connection to resolve. The actual encoding scheme from media reference to the values in the bars is quite complex involving CRC, convolution, and puncturing. The CRC allows the program to check for correct decoding, and the convolution enables the program to have a small number of read errors while still having an accurate result. Puncturing is just removing bits to reduce the numbers encoded, relying on convolution to fill in the holes.

  • Day 7: Neural Nets in Raku (Part 1) – Raku Advent Calendar

    Thinky the Elf was sitting in his office, it had been a closet but he’d been given it as his office after the great baked beans incident. It wasn’t his fault. He was right that feeding the reindeer beans would give them a jet boost but Santa had not been all that happy about it. And his tendency to stare of into space while suddenly having a thought wasn’t great while working on the shop floor meant it was safer to put him out of the way to do some thinking. Recently he’d been thinking about how to sort children into naughty or nice. This was Santa’s big job all year and Thinky thought that there must be a way to simplify it, he’d spent some time watching videos on YouTube and there was one that gave a brilliant description of Neural Networks (jump to 20 minutes for that bit but it’s an interesting video). As Thinky watched this he couldn’t help thinking about Raku and how the connections between nodes felt like Supplies.

  • Rblpapi 0.3.12: Fixes and Updates

    The Rblp team is happy to announce a new version 0.3.12 of Rblpapi which just arrived at CRAN. Rblpapi provides a direct interface between R and the Bloomberg Terminal via the C++ API provided by Bloomberg (but note that a valid Bloomberg license and installation is required). This is the twelveth release since the package first appeared on CRAN in 2016. Changes are detailed below and include both extensions to functionality, actual bug fixes and changes to the package setup. Special thanks goes to Michael Kerber, Yihui Xie and Kai Lin for contributing pull requests!

  • LLVM Now Has "Official" Support For Targeting NEC's Vector Engine (VE) - Phoronix

    The LLVM compiler infrastructure supports not only a growing number of CPU architectures but continues to lead when it comes to its support for different accelerators. Back in 2019 NEC was working to upstream their SX-Aurora VE "Vector Engine" Accelerator and now as of this week that target is considered officially supported upstream. NEC originally launched the SX-Aurora Vector Engine (VE) back in 2018 as a PCI Express accelerator card and supporting up to eight vector processors per server. The NEC SX-Aurora has its own architecture for the "VE" and is backed by HBM2 memory. The current VE processor is rated for 1.53 TB/s of memory bandwidth and a double precision peak performance of 3.07 TFLOPS or 4.91 single precision TFLOPS.

  • 5 Excellent Free Books to Learn CSS - LinuxLinks

    Web pages are built with HTML, which specifies the content of a page. CSS (Cascading Style Sheets) is a separate language which specifies a page’s appearance. CSS code is made of static rules. Each rule takes one or more selectors and gives specific values to a number of visual properties. Those properties are then applied to the page elements indicated by the selectors. Here’s our recommended books to learn CSS.