More good news about Wayland 1.5 is that it's passing all of Intel's automated test-cases. Ullysses Artie Eoff at Intel shared their automated test pass rate was at 100%. In terms of their manual test pass rate, it was at 84%, which is a +15% improvement over the Wayland 1.5 Alpha. Those wishing to find out more about the Wayland 1.5 release candidate results can find the information shared on the Wayland-devel list. Wayland 1.5 with Weston 1.5 should be officially released in the next few days.
Now that kernel development activity is settling down for the Linux 3.15 kernel, here are some benchmarks of the EXT4, XFS, F2FS, and Btrfs file-systems compared to the stable Linux 3.14 kernel performance.
The benchmarks in this article are a Btrfs, F2FS, XFS, and EXT4 file-system comparison between the Linux 3.14 and 3.15 Git (as of 6 May) kernels when carrying out these disk workloads on a 240GB OCZ Vertex 3 solid-state drive.
It's no secret that open source has shaken up the software world, not least for the savings it's brought both organizations and consumers. Now it's starting to look like open source hardware could have a similar, game-changing effect.
Though still nowhere near as ubiquitous as FOSS, open hardware is gaining ground rapidly -- especially with the booming popularity of open source 3D printing -- and some very compelling benefits are becoming clear.
The focus of Intel's DRM Color Manager work is to have a common interface for all color correction / enhancement properties for different hardware, the color manager one be one umbrella DRM property, and DRM drivers can register the color correction properties of its hardware during initialization time.
Most of the performance changes to be found between Mesa 10.1 stable and the current Mesa Git code just past the 10.2 branching was around the HD 7850 graphics card that uses the RadeonSI Gallium3D driver while the other three graphics cards used the R600g driver. With R600g and our assortment of Linux gaming and OpenGL benchmark tests run, we didn't see any better performance in the code beyond where it's at with Mesa 10.1.
In conjunction with its Project Skybridge and K2 announcement, AMD said that today it “demonstrated for the first time its 64-bit ARM-based AMD Opteron A-Series processor, codenamed ‘Seattle,’ running a Linux environment derived from the Fedora Project.” The Fedora-based Linux environment is said to enable development — and migration between — applications based on both x86- and ARM-based processors using common tools.
The NVIDIA 337.19 Beta was released today and it features a bug-fix for HDMI at 4K resolutions in certain configurations, nvidia-settings command-line controls for over/under-clocking support, several cosmetic fixes for the NVIDIA Settings GUI for clock controls, support for the GLX_EXT_stereo_tree extension in certain configurations, and Unified Back Buffer (UBB) and 3D Stereo support with the composite extension for Quadro graphics cards.
The xf86-video-r128 driver supports all of the old ATI Rage 128 graphics cards including the Rage Fury AGP, XPERT 128 AGP, and XPERT 99. The Rage 128 was ATI's best graphics processor back in 1998 and fabbed on a 250nm processor, supporting 32MB and 64MB video memory configurations, and its core was clocked around 100MHz... Its OpenGL compliance stands at version 1.2. While it's hard to believe the Rage 128 is still being used in any production capacity, especially with modern Linux environments, the open-source X.Org driver for it has been revived.
Overall these results aren't too interesting for the Linux 3.15 kernel when it comes to Haswell graphics, but in a few cases there were some slight performance changes as illustrated above. At least Linux 3.15 betters off the Broadwell support, there's now per-process address space support for better security, and a variety of fixes and other improvements that landed for this kernel cycle.
Those wishing to help in testing this latest Intel 3.0 X.Org driver, which will hopefully be released before Wayland takes over the Linux desktop, can be easily built from the xf86-video-intel Git. Ubuntu users can easily test the latest Intel open-source driver code via the Oibaf PPA. It's not too much of an issue that the 3.0 official release hasn't happened yet given that most distributions ship some form of the Git code from the past few months, and that most of the exciting Intel Linux graphics work happens within the kernel DRM and Mesa, but still it's a bit surprising how long this release is taking to materialize.
For those anxious to see some NVIDIA Tegra K1 performance numbers, hopefully this puts the four-plus-one Cortex-A15 performance a bit into perspective... Again, in the coming days will be clean results from Ubuntu 14.04 LTS throughout, power numbers, other GPU/GPGPU benchmarks, and other interesting data from this Jetson TK1 ARM development board. Stay tuned!
There's a big belief that OpenGL 5 will be about optimizing this cross-platform, widely-used graphics API. All of the major hardware companies are working towards reducing OpenGL driver overhead and making other OpenGL improvements as a result of AMD's Mantle API. Mantle is still Windows-only and used by just a handful of games for now with AMD's Catlayst driver on GCN GPUs, but it's ignited a conversation about increasing the performance potential out of OpenGL. DirectX 12.0 is also going to be optimizing the performance potential of Microsoft's 3D graphics API.
David has now posted working patches for his DP MST code on the DRI-devel mailing list. Right now his code has just been tested on a Lenovo Ultrabook boasting Intel "Haswell" graphics and it's working when connected to external hubs. There's still code that's a work in progress but overall it seems to be working fine. Right now this initial "preview code" works for Intel Haswell hardware with certain DP MST hubs.
A few links have been sent in to our news tip box with this page, which reads, "Open Source Mali-200/300/400/450 GPU Kernel Device Drivers." While the page mentions open-source drivers, it's only about the kernel portion of the driver and it's always been that way with ARM -- and most other ARM-based graphics vendors. The kernel portion is open, the user-space components are closed. Without an open user-space, having an open kernel driver is only of limited use, and will not be accepted into the upstream Linux kernel.
The pull request was sent in by AMD's Christian König and tries to add support for their newly-announced Mullins APUs. This pull comes shortly after AMD announced open-source Mullins/Beema APU support, one day after the APUs were announced. Support for the new AMD graphics hardware might come in still for the Linux 3.15 kernel even though it's past the merge window since the GPU is based upon AMD's Kabini and mostly deals with adding new PCI IDs and enums to the kernel driver. There's also the libdrm and Mesa RadeonSI Gallium3D updates needed in user-space for complete support.