Language Selection

English French German Italian Portuguese Spanish

Graphics/Benchmarks

Kernel, LF and Graphics: New Features, Hyperledger Global Forum, RadeonSI Gallium3D

Filed under
Graphics/Benchmarks
Linux
  • Linux 4.19 Certainly Is Going To Be A Big Kernel

    At the end of July I outlined some of the changes queued for Linux 4.19 while since then several more notable additions have become aligned for this next kernel cycle following the one week delay of Linux 4.18.

    Linux 4.18 should be out this weekend, which itself is bringing many changes, and in turn will mark the opening of the two week long merge window for 4.19.

  • Linux Live Kernel Patching Is On The Way For 64-bit ARM

    It's been a while since having any advancements to report on the live kernel patching front for being able to deploy primarily Linux kernel security fixes without having to reboot the system. There is some new progress on that front to report on now and that is the 64-bit ARM (ARM64/AArch64) support getting squared away.

  • Linux EDAC Support For AMD's Great Horned Owl

    The latest Linux kernel patch is for supporting ECC error detection via the Error Detection And Correction (EDAC) code with AMD's Great Horned Owl.

    Great Horned Owl is the codename for Family 17h Model 11h hardware, which is what launched back in February as the Ryzen Embedded V1000 series chips.

  • Linux Kernel Expectations For AMD Threadripper 2

    If you have already pre-ordered your AMD Threadripper 2990WX processor or just planning to be an early customer of that high-end desktop processor or the Threadripper 2950X, you may be wondering about Linux requirements from these new high-end AMD CPU offerings. Here's the gist of the Linux support state of AMD Zen+ CPUs for those wanting to get ready for Threadripper 2.

    There's still a few days to go until the global embargo expires for sharing Threadripper 2990WX/2950X performance benchmarks and reviews, at which time you'll be able to see my full Linux analysis of these new processors with plenty of tests... But for today we're just talking in general terms about Linux support requirements.

  • Keynote Sneak Peek for Hyperledger Global Forum – See Who’s Speaking

    Attend Hyperledger Global Forum to see real uses of distributed ledger technologies for business and to learn how these innovative technologies run live in production networks across the globe today. Hyperledger Global Forum will cut through the hype and focus on adoption. Attendees will see first-hand how the largest organizations in the world go beyond experimentation to lead blockchain production applications with measurable impact. Make your plans now to attend the premier blockchain event of 2018.

  • More OpenGL Extensions For RadeonSI Are The Latest In A Flurry Of Interesting Activity

    Marek Olšák has continued his busy work on the RadeonSI Gallium3D driver by implementing more extensions and adjusting various limits/constants to match the behavior of their proprietary driver.

    In the past few days Marek has fleshed out the EQAA GL extensions, ASTC compression for Gallium drivers, and yesterday was EXT_gpu_shader4 support for vintage/legacy OpenGL 2 applications.

  • Intel SSD 660p: 512GB Of NVMe Storage For $99 USD

Games and Wine

Filed under
Graphics/Benchmarks
Gaming
  • Extreme downhill freeriding game 'Descenders' updated with the Ring of Fire update

    The latest update to Descenders adds in the fifth environment to play through named Volcano, which gives you a different way to play. The interesting thing about the Volcano environment, is that you have to unlock it in a very specific way with the developer teasing that "it involves fiery rings of fire". Instead of stunts, this environment uses tectonics instead. In other words, you're in for a bumpy ride.

    In addition to the new environment, which is pretty darn tough to actually reach, they've also added in more new stuff. Along with now having a Linux version that works properly.

    In this update there's also another new bonus level for you to find, they've expanded the "Freeride" mode where you tweak the settings to build your own track to add in the special build crew feature, which gives you certain modifiers like smoother curves or reducing speed wobble. There's also new items for you to unlock as well.

  • It Looks Like A Steam 64-Bit Client Could Finally Be Near

    It looks like Valve could be prepping to finally ship a 64-bit Steam client, possibly coinciding with their long talked about Steam UI/UX overhaul.

    Valve today shipped a new beta release with just one listed change, "Added support for shipping different binaries to 64bit vs 32bit operating systems in Steam self-updater. This support is being added in preparation for future updates."

    The ability for the Steam update to ship different binaries depending upon architecture would be a prerequisite for being able to introduce a 64-bit Steam client while retaining the 32-bit client too... They don't explicitly say it's for shipping a 64-bit Steam client soon, but it's 2018 after all, and there wouldn't be much of a reason to otherwise be adding this capability to the Steam updater...

  • DXVK Introducing Per-Game Configuration Files

    While DXVK is capable of running a great deal of Direct3D 11 games via Vulkan within Wine, a number of games have required various workarounds for either getting the game to properly work in the first place or to run efficiently. Those per-game settings are now being punted off into a per-game configuration system.

  • VK9 - Direct3D 9 Over Vulkan - Completes 27th Milestone

    It's not nearly as far along as DXVK that is allowing D3D11-over-Vulkan and already running great numbers of games in a performant manner under Wine, but the VK9 project for implementing Direct3D 9 over Vulkan has now hit its twenty-seventh milestone.

Graphics: Libinput, Mesa and AMD

Filed under
Graphics/Benchmarks
  • How the 60-evdev.hwdb works

    libinput made a design decision early on to use physical reference points wherever possible. So your virtual buttons are X mm high/across, the pointer movement is calculated in mm, etc. Unfortunately this exposed us to a large range of devices that don't bother to provide that information or just give us the wrong information to begin with. Patching the kernel for every device is not feasible so in 2015 the 60-evdev.hwdb was born and it has seen steady updates since. Plenty a libinput bug was fixed by just correcting the device's axis ranges or resolution. To take the magic out of the 60-evdev.hwdb, here's a blog post for your perusal, appreciation or, failing that, shaking a fist at. Note that the below is caller-agnostic, it doesn't matter what userspace stack you use to process your input events.

  • Mesa 18.2-RC2 Released With 17 Fixes So Far

    One week after branching Mesa 18.2 and issuing the first release candidate, the second weekly RC is now available for testing.

    Mesa 18.2-RC2 has 17 patches queued, including several V3D (formerly "VC5") Broadcom driver fixes, build system updates, fixing the DRISW compilation for Android Nougat, and other small fixes.

  • Mesa 18.2.0-rc2

    The second release candidate for the Mesa 18.2.0 is now available.

  • AMDGPU LRU Bulk Move Functionality Increases Performance In OpenCL And Vulkan
  • AMD Releases 18.Q3 Linux Drivers for Radeon Pro, Including Ubuntu 18.04 LTS Support

    This update also brings Vulkan 1.1 support, and initial support for Ubuntu 18.04 LTS – which is great news for those who recently upgraded to the latest Ubuntu 18.04 LTS package.

    The package is mainly intended for Vega Frontier, Radeon Pro, Radeon Pro WX, and FirePro S/W graphics cards, and the entire driver stack is derived from the 18.20 driver branch which includes both PRO and All-Open driver options.

Graphics: Radeon, AMDGPU, NVIDIA

Filed under
Graphics/Benchmarks
  • AMD Radeon Pro 18.Q3 Linux Driver Released

    AMD has released their Radeon Pro 18.Q3 enterprise driver update, including for supported Linux distributions, as their QA-tested professional-grade driver update for the quarter.

    Radeon Pro Software 18.Q3 is available for RHEL/CentOS 7.5, RHEL/CentOS 6.9, Ubuntu 18.04, Ubuntu 16.04, and SUSE Linux Enterprise Desktop/Server 12 SP3. This is basically the QA-vetted enterprise driver stack of "AMDGPU-PRO" intended for workstation graphics hardware while for the most part those on consumer Radeon GPUs are generally best off riding the latest Linux+Mesa releases for the leading open-source driver coverage.

  • AMDGPU LRU Bulk Move Patches Yielding OpenCL & Vulkan Performance Boosts

    AMD's Huang Rui and Christian Konig are working on LRU bulk move functionality that can yield performance improvements for Vulkan and OpenCL workloads.

    This AMDGPU kernel change is about migrating PD/PT buffer objects onto the LRU (least recently used) list in bulk in a single operation. For end-users, what matters is the performance difference as a result.

  • NVIDIA Vulkan Beta Update Brings Hang Fixes, Device Diagnostic Checkpoints

    NVIDIA has just published their latest Vulkan beta driver release for Windows and Linux.

    The newest NVIDIA Vulkan beta for Linux out today is version 396.51.02. Most notable with this driver update is that it adds support for the VK_NV_device_diagnostic_checkpoints extension. This extension was recently made public by Vulkan 1.1.82 and is for inserting device diagnostic checkpoints.

Toshiba RC100 NVMe SSD Ubuntu Linux Benchmarks

Filed under
Graphics/Benchmarks

Back in June Toshiba introduced the RC100 NVMe solid-state drive as a new low-end offering. The RC100 is now a bit blindsided by Intel's just-launched 660p SSD that delivers incredible storage capacities per dollar, and I'll have some Intel 660p Linux benchmarks in a few days, but for those curious about the RC100 here are some Ubuntu Linux benchmark results for this low-cost NVMe SSD.

The Toshiba RC100 is a Ball Grid Array (BGA) SSD with the controller and flash memory on a single package. With that the RC100 NVMe SSDs are single-sided and fit on an M.2 2242 PCB rather than the longer and more common M.2 2280 NVMe SSDs. But with being M.2 2242 is limited to PCI Express x2 lanes rather than x4 bandwidth. The RC100 doesn't have any DRAM cache but does support NVMe Host Memory Buffer (HMB) and the drive uses 64 layer BiCS3 3D TLC NAND flash memory.

Read more

X.Org Server 1.20.1

Filed under
Graphics/Benchmarks
  • xorg-server 1.20.1

    This bugfix release fixes several issues in RANDR, Xwayland, glamor, the modesetting driver, and elsewhere. Everyone is encouraged to upgrade. Thanks to all who contributed to this release!

  • X.Org Server 1.20.1 Released With Many Bug Fixes

    As the first point release to X.Org Server 1.20 that debuted in May, xorg-server 1.20.1 is available today with dozens of fixes.

    X.Org Server 1.20.1 is quite a big point release to address the initial fall-out from X.Org Server 1.20 after that release was in development for more than one year and shipped with a ton of changes and new features.

  • X.Org Server 1.20.1 Point Release Addresses Many Bugs including XWayland and Mesa

    The first point release for X.Org Server 1.20 has been released today (xorg-server 1.20.1) and it is chock full of fixes.

    X.Org Server 1.20 was in development for more than a year, but its initial release in May was not exactly smooth, as there were numerous bug fallouts and things that needed adjusting. Fortunately, X.Org Server 1.20.1 is the first point release to address a myriad of issues that were effecting the initial release – which is great news, as this point release is just in time to hopefully make it into the various Linux distros that utilize X.Org Server (Ubuntu 18.10, Fedora 29, etc).

Graphics: Filament and Sway 1.0 Alpha 5

Filed under
Graphics/Benchmarks
  • Google Open-Sources "Filament" PBR Engine Using Vulkan/OpenGL

    Filament is a physically-based rendering engine that has now been open-sourced by Google for Android, Linux, macOS, and Windows systems.

    This physically-based rendering engine is designed to be as small and efficient as possible so that it can scale down and run with ease on Android-based systems. Filament is written in C++ and requires the use of the LLVM/Clang compiler, supports OpenGL 4.1+ / OpenGL ES 3.0+ / Vulkan 1.0 for rendering back-ends, supports a wide range of rendering capabilities, and all-around looks like quite an exciting PBR engine.

  • Sway 1.0 Alpha 5 Brings Multi-GPU Support, Virtual Keyboard Protocol

    The i3-inspired Sway Wayland compositor had already introduced many features ahead of Sway 1.0 while with today's fifth alpha release are yet more new features to advertise.

    Sway 1.0 Alpha 5 was released today and while it's just a few weeks past the alpha 4 milestone, there are more than 250 changes and a number of new features.

Graphics: X.Org Server, RadeonSI, OpenChrome, VKGL, ADriConf/Advanced DRI Configurator

Filed under
Graphics/Benchmarks
  • X.Org Server 1.20 Branch Created, Latest EGLStreams Patches Added

    X.Org Server 1.20 was released back in May while now the "server-1.20-branch" was created at last to allow for X.Org Server 1.21 development to happen on master while letting the point releases to be worked out on the branched code.

    The server-1.20-branch is where the latest code is being staged ahead of the eventual X.Org Server 1.20.1 point release. The branch was just created and has staged some patches so far.

  • RadeonSI Gets Patches For AMD_framebuffer_multisample_advanced (EQAA)

    Last month AMD's Marek Olšák sent out a new extension for the OpenGL registry, AMD_framebuffer_multisample_advanced, and with the latest Mesa patches he has published this week the RadeonSI Gallium3D driver wires in support for this GL extension.

    AMD_framebuffer_multisample_advanced ends up implementing AMD's EQAA: Enhanced Quality Anti-Aliasing. Merged a few months back was EQAA for RadeonSI that is part of Mesa 18.2 but that initial implementation which is activated via an environment variable is just used in place of OpenGL applications otherwise trying to use multi-sample anti-aliasing (MSAA). With AMD_framebuffer_multisample_advanced is now a means for OpenGL games/applications wanting to explicitly target this advanced anti-aliasing mode.

  • OpenChrome DRM Driver To Go Through A GEM/TTM Code Rewrite

    With the OpenChrome DRM/KMS driver for vintage VIA x86 graphics likely to be mainlined in its current code state, the sole developer left working on this driver is going to next rewrite the TTM/GEM memory management code that he also hopes will help in his new ATI RAGE 128 driver initiative.

    The OpenChrome DRM driver that's been struggling for years doesn't look like it will be mainlined in 2018. While it's been sent off for review a few matters have been blocking it from going mainline: the principal challenges are OpenChrome DRM not supporting the modern atomic mode-setting APIs but rather than legacy KMS interfaces and there being a lot of unfinished code left in the driver. When it comes to the unfinished code, the hardware acceleration isn't complete and the upstream DRM driver developers would want that removed before this driver would be hypothetically mainlined.

  • VKGL: An Effort For OpenGL Core Profile Support Over Vulkan

    A few days ago we wrote about GLOVE being open-sourced as OpenGL ES over Vulkan and we were then pointed out to another project: VKGL.

    GLOVE currently is working for OpenGL ES 2.0 over Vulkan with plans for supporting not only newer versions of OpenGL ES but also to achieve at least some desktop OpenGL support. Think Silicon's GLOVE project is quite promising and is cross-platform. Now there is also VKGL that seems to be the project of just one independent developer for getting desktop OpenGL (using a core profile context) over Vulkan.

  • Advanced DRI Configuration Picking Up New Features

    It's been a while since having anything to report on ADriConf but fortunately this graphical utility for configuring some open-source Linux graphics driver features is progressing.

    ADriConf, The Advanced DRI Configurator, is a GUI utility for the configuring open-source graphics drivers that aims to be modern and more featureful compared to the long-standing DriConf. Recent work to ADriConf includes new shortcuts in the menu, build issue fixes/improvements, new test cases added, and proper handling as well with the closed-source graphics drivers.

Graphics: AMDVLK and Mesa

Filed under
Graphics/Benchmarks
  • AMDVLK Radeon Vulkan Driver Updated With 8-Bit Storage Support

    Another weekly code drop has occurred for the "AMDVLK" open-source AMD Radeon Vulkan Linux driver with its XGL/PAL/LLPC.

    On the Vulkan front, the XGL code drop on Friday now exposes the VK_KHR_8bit_storage extension for allowing 8-bit types in uniform/storage buffers and push constant blocks. The updated XGL code also has some fixes and other minor work but the 8-bit storage support is the main addition.

  • EGL Device Support Coming Together For Mesa

    Emil Velikov's latest Mesa work is on implementing support for EGL Device extensions for enumerating and using EGLDevices.

    The extensions Emil has implemented in patch form include EGL_EXT_device_base, EGL_MESA_device_software, EGL_EXT_device_drm, and EGL_platform_device. The EGLDevice extensions allows for bringing up EGL without any underlying/native windowing system API.

  • Adreno A6xx Gallium3D Support Coming Together

    For the past number of months there's been Adreno A600 series support coming together within the MSM DRM kernel driver in large part thanks to Qualcomm / Code Aurora contributing code themselves. Quietly coming together as well is the A6xx Gallium3D support for allowing OpenGL acceleration.

    When it comes to the latest-generation Adreno A6xx hardware, most of the open-source talk has been on the MSM Direct Rendering Manager driver front but it's great to see the Gallium3D/OpenGL driver support being pieced together too. A Phoronix reader pointed out that this work is being staged via this wip/a6xx branch.

A Look At The Clear Linux Performance For July 2018

Filed under
Graphics/Benchmarks

Given our fascination with Intel's Clear Linux performance in the plethora of performance benchmarks we frequently run at Phoronix and this open-source operating system being maintained in a rolling-release style, here's a look at how the performance for this x86_64 Linux distribution evolved over the past month.

For those curious about the performance of Intel's Clear Linux for July 2018, here are benchmarks of its state as of 1 July 2018 to where it was at in ending out the month on 31 July. Tests were done on nine different Intel systems from desktop systems to Xeon workstations/servers.

Clear Linux began July with build 23370 that was using the Linux 4.17 kernel, GCC 8.1.1, and Python 3.6.5 as some of the important versions to point out for this testing. At the end of July they were up to build 24090 with the Linux 4.17.11 kernel, GCC 8.2.0, and Python 3.7, among many other package upgrades and changes throughout the month.

Read more

Syndicate content

More in Tux Machines

RISC-V and NVIDIA

  • Open-Source RISC-V-Based SoC Platform Enlists Deep Learning Accelerator
    SiFive introduces what it’s calling the first open-source RISC-V-based SoC platform for edge inference applications based on NVIDIA's Deep Learning Accelerator (NVDLA) technology. A demo shown at the Hot Chips conference consists of NVDLA running on an FPGA connected via ChipLink to SiFive's HiFive Unleashed board powered by the Freedom U540, the first Linux-capable RISC-V processor. The complete SiFive implementation is suited for intelligence at the edge, where high-performance with improved power and area profiles are crucial. SiFive's silicon design capabilities and innovative business model enables a simplified path to building custom silicon on the RISC-V architecture with NVDLA.
  • SiFive Announces First Open-Source RISC-V-Based SoC Platform With NVIDIA Deep Learning Accelerator Technology
    SiFive, the leading provider of commercial RISC-V processor IP, today announced the first open-source RISC-V-based SoC platform for edge inference applications based on NVIDIA's Deep Learning Accelerator (NVDLA) technology. The demo will be shown this week at the Hot Chips conference and consists of NVDLA running on an FPGA connected via ChipLink to SiFive's HiFive Unleashed board powered by the Freedom U540, the world's first Linux-capable RISC-V processor. The complete SiFive implementation is well suited for intelligence at the edge, where high-performance with improved power and area profiles are crucial. SiFive's silicon design capabilities and innovative business model enables a simplified path to building custom silicon on the RISC-V architecture with NVDLA.
  • SiFive Announces Open-Source RISC-V-Based SoC Platform with Nvidia Deep Learning Accelerator Technology
    SiFive, a leading provider of commercial RISC-V processor IP, today announced the first open-source RISC-V-based SoC platform for edge inference applications based on NVIDIA’s Deep Learning Accelerator (NVDLA) technology. The demo will be shown this week at the Hot Chips conference and consists of NVDLA running on an FPGA connected via ChipLink to SiFive’s HiFive Unleashed board powered by the Freedom U540, the world’s first Linux-capable RISC-V processor. The complete SiFive implementation is well suited for intelligence at the edge, where high-performance with improved power and area profiles are crucial. SiFive’s silicon design capabilities and innovative business model enables a simplified path to building custom silicon on the RISC-V architecture with NVDLA.
  • NVIDIA Unveils The GeForce RTX 20 Series, Linux Benchmarks Should Be Coming
    NVIDIA CEO Jensen Huang has just announced the GeForce RTX 2080 series from his keynote ahead of Gamescom 2018 this week in Cologne, Germany.
  • NVIDIA have officially announced the GeForce RTX 2000 series of GPUs, launching September
    The GPU race continues on once again, as NVIDIA have now officially announced the GeForce RTX 2000 series of GPUs and they're launching in September. This new series will be based on their Turing architecture and their RTX platform. These new RT Cores will "enable real-time ray tracing of objects and environments with physically accurate shadows, reflections, refractions and global illumination." which sounds rather fun.

today's leftovers

GNOME Shell, Mutter, and Ubuntu's GNOME Theme

Benchmarks on GNU/Linux

  • Linux vs. Windows Benchmark: Threadripper 2990WX vs. Core i9-7980XE Tested
    The last chess benchmark we’re going to look at is Crafty and again we’re measuring performance in nodes per second. Interestingly, the Core i9-7980XE wins out here and saw the biggest performance uplift when moving to Linux, a 5% performance increase was seen opposed to just 3% for the 2990WX and this made the Intel CPU 12% faster overall.
  • Which is faster, rsync or rdiff-backup?
    As our data grows (and some filesystems balloon to over 800GBs, with many small files) we have started seeing our night time backups continue through the morning, causing serious disk i/o problems as our users wake up and regular usage rises. For years we have implemented a conservative backup policy - each server runs the backup twice: once via rdiff-backup to the onsite server with 10 days of increments kept. A second is an rsync to our offsite backup servers for disaster recovery. Simple, I thought. I will change the rdiff-backup to the onsite server to use the ultra fast and simple rsync. Then, I'll use borgbackup to create an incremental backup from the onsite backup server to our off site backup servers. Piece of cake. And with each server only running one backup instead of two, they should complete in record time. Except, some how the rsync backup to the onsite backup server was taking almost as long as the original rdiff-backup to the onsite server and rsync backup to the offsite server combined. What? I thought nothing was faster than the awesome simplicity of rsync, especially compared to the ancient python-based rdiff-backup, which hasn't had an upstream release since 2009.