Language Selection

English French German Italian Portuguese Spanish

Graphics/Benchmarks

Graphics: Vulkan, Intel and NVIDIA

Filed under
Graphics/Benchmarks
Linux
  • Vulkan 1.2.158 Released With Fragment Shading Rate Extension - Phoronix

    Vulkan 1.2.158 was released this morning with two notable extensions introduced.

    First up is VK_KHR_fragment_shading_rate that allows changing the rate at which fragments are shaded. Multiple pixels can be shaded now by a single fragment shader invocation. The new extension allows controlling the fragment shading rate on a per-draw, per-primitive, or per-region basis. Most notably this can be used by Vulkan-powered games for shading higher levels of detail in a scene compared to others. Or rather lower quality shading in some areas of the scene.

  • Linux 5.10 Continues Bringing Up Support For Intel's Rocket Lake - Phoronix

    Building off Linux 5.9 that featured initial support for Gen12 graphics on next year's Rocket Lake desktop platform along with other early enablement for Rocket Lake like RAPL support and other PCI ID additions, that work has continued for the Linux 5.10 cycle.

    The libata pull adds Rocket Lake PCH-H RAID PCI IDs as one of the additions.

    There is also the platform-drivers-x86 work for Linux 5.10 where Rocket Lake support is added to the intel_pmc_core driver.

  • GCC's Ada Frontend Seeing More Work On NVIDIA CUDA Support - Phoronix

    Should you want to use the Ada programming language for GPU programming, the GCC compiler has been working on CUDA support within its front-end for this safety and security minded language.

    In the past born out of academia there's been CUDA Ada bindings. There has also been Ada/SPARK GPU programming initiatives in the past with various APIs. This latest still ongoing effort is wiring up the GCC Ada front-end with CUDA support.

  • You may want to avoid Linux Kernel 5.9 if you want fully supported NVIDIA drivers | GamingOnLinux

    On the official NVIDIA forum, an employee put out an announcement warning NVIDIA GPU owners that the Linux Kernel 5.9 and later is currently unsupported. It's worth noting they posted that in the CUDA forum, so other workloads like gaming may work as normal.

    In the post they mention Kernel 5.9+ is currently "incompatible" with any of their drivers, and they're suggesting to wait until "mid-November" for a fresh NVIDIA driver update which is expected to bring support for it. They're "working diligently" to get ready to support it.

Graphics: NVIDIA, Radeon and Zink OpenGL-on-Vulkan

Filed under
Graphics/Benchmarks
  • NVIDIA Doesn't Expect To Have Linux 5.9 Driver Support For Another Month

    While NVIDIA is usually quite timely in supporting new versions of the Linux kernel and aim to have out a driver by the end of the release candidates for new series, in the case of the recently minted Linux 5.9 kernel it's taking a lot longer.

    NVIDIA sent out an advisory to their customers on Friday that they don't expect to have Linux 5.9 kernel support in their proprietary graphics driver until mid-November, or about one month after the kernel was released as stable.

  • NGG Appears To Be In Good Shape For RDNA 2 / RX 6000 Series

    NGG (Next-Gen Geometry) ended up being fairly buggy/problematic for Navi but it looks like for the upcoming Radeon RX 6000 (RDNA 2 / Navi 2) launch that it's in better shape.

    The NGG support for Navi 1x involved a lot of driver work from at times enabling it only to disable some functionality to other tweaks in avoiding hangs or other problems around it.

  • Zink In Mesa 20.3 Now Hits OpenGL 3.3, Can Run Blender With This OpenGL-on-Vulkan

    For weeks there have been patches getting the Zink OpenGL-on-Vulkan implementation to OpenGL 4.6 while mainline Mesa has been at OpenGL 3.0 support. Thankfully the out-of-tree patch delta is being reduced and this week in Mesa 20.3-devel the code has been upstreamed getting the support level to OpenGL 3.3.

How Intel's Clear Linux Is Competing Against Late-2020 Linux Distributions

Filed under
Graphics/Benchmarks

As it's been a while since running a fresh Linux distribution comparison with Intel's Clear Linux platform and given all the autumn distribution updates inbound, here is a fresh look at the rolling-release Clear Linux up against a snapshot of Ubuntu 20.10, Fedora Workstation 33 Beta, openSUSE Tumbleweed 20200929, Arch-based Endeavour OS, and Debian Testing from the start of October.

Read more

Also: Intel Core i7 1165G7 "Tiger Lake" Linux Performance With The Dell XPS 13 9310

Compute Runtime and Graphics Work in Linux

Filed under
Development
Graphics/Benchmarks

Mesa 20.2.1 and Mesa 20.1.10 Released

Filed under
Graphics/Benchmarks

  • mesa 20.2.1
    Hi list,
    
    
    
    
    I realize that this is a week late, I simply put everything in the calendar one
    week off. Doh. Anyway, mesa 20.2.1 is now available, this release looks much
    bigger than it actually is, because of all of the .pick_status commits. there's
    a bit of everything in here, all and all a nice little .1
    
    
    
    
    Dylan
    
  • Mesa 20.2.1 Released With Initial Batch Of Fixes

    Mesa 20.2 officially released at the end of September as the Q3'2020 open-source driver stack update providing open-source OpenGL/OpenCL/Vulkan support for much of the graphics hardware on the market. For those that prefer waiting for the first point release before upgrading, that milestone was reached today. 

  • Mesa 20.1.10 Is Released With A Handful Of Bug-Fixes

    Mesa 20.1.10 is a small bug-fix release for GNU/Linux distributions that have not yet upgraded to Mesa 20.2.0. There's not much to see, there's seven for the Intel graphics drivers and two on the AMD side.

Raspberry Pi 4 Vulkan Driver "V3DV" Merged Into Mesa 20.3

Filed under
Graphics/Benchmarks
Hardware

Good news for Raspberry Pi 4 users... The V3DV Vulkan driver developed over the past year for newer Broadcom VideoCore hardware with an emphasis on the Raspberry Pi 4 support is now mainlined in Mesa 20.3!

Consulting firm Igalia working under contract with the Raspberry Pi Foundation has been developing V3DV as a Vulkan driver for the Raspberry Pi 4 and presumably future generations of the Raspberry Pi SBC. The driver recently reached a level roughly to Vulkan 1.0 conformance and thus began focusing on upstreaming the driver into Mesa, which is now landed in Git today.

Read more

Does subgroup/wave size matter?

Filed under
Development
Graphics/Benchmarks

This week, I had a conversation with one of my coworkers about our subgroup/wave size heuristic and, in particular, whether or not control-flow divergence should be considered as part of the choice. This lead me down a fun path of looking into the statistics of control-flow divergence and the end result is somewhat surprising: Once you get above about an 8-wide subgroup, the subgroup size doesn't matter.

Before I get into the details, let's talk nomenclature. As you're likely aware, GPUs often execute code in groups of 1 or more invocations. In D3D terminology, these are called waves. In Vulkan and OpenGL terminology, these are called sugroups. The two terms are interchangeable and, for the rest of this post, I'll use the Vulkan/OpenGL conventions.

Read more

Kernel: EXT4, DSA, NTFS and RTX 3080 Support

Filed under
Graphics/Benchmarks
Linux

  • EXT4 "Fast Commits" Coming For Big Performance Boost In Ordered Mode

    After being in development for more than one year, it looks like with Linux 5.10 there will be EXT4 fast commit support.

    [...]

    When running in EXT4 ordered mode with the fast commit patches, Ted reported a ~103% write performance improvement. Thus seeing this new code make it into the EXT4 "dev" branch ahead of mainline integration is quite exciting if you rely on EXT4 ordered journal functionality.

  • The Linux Kernel Preparing To Take Advantage Of The Intel DSA / ENQCMD In Sapphire Rapids

    Expected with next year's Intel Sapphire Rapids Xeon CPUs is the Intel DSA as the Data Streaming Accelerator for high performance data movement and transformation operations. Since the end of 2019 there have been Linux patches surfacing for bringing up the DSA support and now as we roll into 2021 the Linux kernel looks to begin making use of the new capabilities.

    With the Intel DSA block there are new instructions with ENQCMD/ENQCMDS for submitting work descriptors along with the PASID instruction for the Process Address Space ID. We've seen various Linux driver work around the Data Streaming Accelerator as well as supporting these new instructions in the open-source compiler.

  • Paragon Sends Out Latest NTFS Read-Write Linux Driver Patches

    Back in August was the big surprise of file-system driver vendor Paragon Software wanting to mainline their NTFS driver into the Linux kernel that is much more advanced than the existing NTFS Linux driver. While not merged yet, on Friday the latest version was sent out for review.

  •  

  • This Is Your RTX 3080 On Linux

    Phoronix received an RTX 3080 to test on Linux and with a quick install of the NVIDIA 455.23.05 beta Linux graphics driver and CUDA 11.1 they were off to the races.  The results are consistent with the performance on Window, with noticeably improved performance across the board.  There are a handful of benchmarks in which it doesn’t take top spot, specifically when looking at performance per watt, even if it is the fastest overall.

    In machine learning benchmarks like PlaidML Ampere really struts it’s stuff, showing significant improvements compared to previous cards.  In the end, the geometric mean of performance shows the RTX 3080 is 27% faster than the GeForce TITAN RTX, 63% more than the RTX 2080 SUPER and 73% better than the original GeForce RTX 2080.  This makes the RTX 3080 a big winner in performance per dollar.

The Current Intel Coffee Lake Mitigation Performance Impact With Linux 5.9

Filed under
Graphics/Benchmarks

Of the many new features in Linux 5.9 with its debut set for this weekend, one of the performance-related changes is Intel FSGSBASE support finally being mainlined. A half-decade after the Linux patches first appeared for this feature present in Intel CPUs going back to Ivy Bridge, the mainline kernel is now patched for this feature that can help out I/O and other context switching heavy workloads. Given many of the same workloads were negatively impacted by the CPU security mitigations of recent years, here is a look at the current mitigated vs. unmitigated performance difference on the Linux 5.9 kernel with an Intel Core i9 9900K CPU for reference on how the mitigation impact is on recent versions of the Linux kernel.

Read more

Graphics: Mesa, AMD and More

Filed under
Graphics/Benchmarks
Linux
  • OpenCL 1.2 Support Merged For Mesa's Gallium3D Clover While OpenCL 3.0 Is Being Tackled

    With this quarter's Mesa 20.3 the Gallium3D "Clover" state tracker providing OpenCL support finally can handle version 1.2!

  • AMD Renoir On Linux Could Soon See Lower Power Consumption During Video Playback

    It looks like for Linux 5.11 there could be lower power usage during video playback on DCN 2.1 hardware, namely AMD Renoir.

    Sent out on Wednesday were the latest set of AMD DC patches for their "display core" code shared between Windows and Linux. Of the patches, catching our eye was a DCN 2.1 power optimization for video playback.

    That change amounts to enabling ODM Combine and fullscreen MPO on DCN2.1 hardware, which is most notably Renoir. The full-screen MPO support is regarding multi-plane overlays.

  • Mike Blumenkrantz: Cache Harder

    It’s just that kind of week.

    When I left off in my last post, I’d just implemented a two-tiered cache system for managing descriptor sets which was objectively worse in performance than not doing any caching at all.

    Cool.

    Next, I did some analysis of actual descriptor usage, and it turned out that the UBO churn was massive, while sampler descriptors were only changed occasionally. This is due to a mechanism in mesa involving a NIR pass which rewrites uniform data passed to the OpenGL context as UBO loads in the shader, compacting the data into a single buffer and allowing it to be more efficiently passed to the GPU. There’s a utility component u_upload_mgr for gallium-based drivers which allocates a large (~100k) buffer and then maps/writes to it at offsets to avoid needing to create a new buffer for this every time the uniform data changes.

Syndicate content

More in Tux Machines

New Systemd 247 Is Out For Linux Operating System As Major Release

Systemd, a controversial system and service manager for Linux operating systems, has a major version release as Systemd 247. Speaking of new changes, systemd 247 has added a new service called systemd-oomd to monitor and take action on processes when memory or swap goes above the configured limits. Systemd, a controversial system and service manager for Linux operating systems, has a major version release as Systemd 247. Speaking of new changes, systemd 247 has added a new service called systemd-oomd to monitor and take action on processes when memory or swap goes above the configured limits. Read more

Comparing the similarities and differences between inner source and open source

Open source software (OSS) has been around since the 1990s and has thrived, quickly growing to become mainstream. It is now more well understood around the world than it has ever been before. Some refer to it as FOSS to highlight the Freedom part of open source (Free and Open Source Software). And in 2014, at OSCON, the term "inner source" was debuted, and people started talking about how to use the principles of open source, but inside of a company. It raised several questions for those unfamiliar with the term, which I hope to answer with this article. For example, what is similar about the two, what is different, the company roles involved in the two, is inner source taking the energy away from open source, etc. These are all fair questions, and as my organization practices both and is involved in both movements, I want to take some time to share insight with this audience as a developer, as a company, and as an open source enthusiast. Read more

4 questions about AI ethics and how open source can help

As a high school student, I've become very interested in artificial intelligence (AI), which is emerging as one of the most impactful innovations of recent times. This past summer, I was selected for the AI4ALL program, where we learned how to develop AI systems using Python. For my final project, I created an object-detection program and integrated it with a virtual drone simulation. Throughout the project, I was able to use open source frameworks, including TensorFlow, Keras, Scikit-learn, and PyTorch, to aid in developing the object-detection machine learning (ML) algorithm process. Read more

MuditaOS: A Beautiful and Minimal Open Source Mobile Operating System

MuditaOS is a beautiful, minimal open-source operating system for mobile phones. Unlike other mobile phone operating systems, however, the developers behind MuditaOS are not interested in smartphones. Instead, they aim to take us to the era before the smartphone craze that has lead to distrust for big tech companies, but with a far cooler style. Developing this mobile operating system has admittedly been a challenge to the team and they are excited to have come up with a beautifully designed E Ink OS, which they have open-sourced to meet the modern user’s desire for transparency without compromising quality. When the company developed and open-source the operating system, they announced that open-sourcing the OS goes along the lines of their “You’re happy – I’m Happy” philosophy and this makes even more sense when you understand that Mudita is from the Sanskrit word “Mudit” which translates to ‘Happy‘. Read more