AMD sent in a batch of fixes for the AMDGPU kernel driver today for Linux 4.3. One notable change with this AMDGPU DRM driver update is that it marks the Iceland/Topaz graphics processor support as experimental so it's no longer enabled by default until the support has been better vetted.
Since this summer we've known that Canonical developers have been looking at Vulkan in regards to supporting this forthcoming graphics API by Unity 8 and Mir. Since then we've seen work done in Mir to support renderers other than OpenGL with this Ubuntu display server. As another sign of working towards Vulkan, more of Mir's OpenGL code continues to be re-factored.
Given the recent releases of FreeBSD 10.2 and NetBSD 7.0, plus the H2'2015 Linux distribution updates rolling around, I've just started work on a new BSD vs. Linux operating system performance comparison.
First up are the BSD distributions for testing... The test system being used for this comparison is an Intel Xeon E5-2687W v3 Haswell-E plus AMD FirePro system. Given the new release of NetBSD 7.0, I decided to try that out first.
Enlightenment DR 0.20 Alpha has been released as the first step towards E20 with one year having passed since E19.
Enlightenment E20 in its current state has full Wayland support with much better, more featureful support than what's found in E19. That's why Wayland support was removed from E19 rather than for any nefarious reasons.
Earlier this week I posted a number of openSUSE Leap benchmarks of their different kernels: debug, default, desktop, and vanilla. Here's some follow-up tests with more results from comparing the openSUSE 42.1 Leap Beta kernel builds.
The tests are very similar to the article earlier this week, just with many more data-points now after seeing the performance differences from the initial test suite.
A few weeks ago at IDF, Displaylink released drivers for USB monitors on Linux. This has been something SemiAccurate has been asking them about since, well it has been years now.
The idea is simple, transfer video data over USB rather than a dedicated video port. This requires a bit of compression, CPU load, and of course their proprietary hardware on the monitor side. That isn’t a big deal, the chips are fairly inexpensive and since you are buying a USB monitor or dock, it comes with the device out of the box. On the plus side it means your monitor will work everywhere, or at least it will now.
While there's just a handful of names that Phoronix readers are familiar with when it comes to AMD's open-source Linux driver developers and those from AMD who communicate with the community in our forums, it turns out there are many more developers at AMD becoming involved as part of their new AMDGPU driver stack.
Those unfamiliar with the openSUSE kernel flavors can see this Wiki page. Tested for this article was the default, desktop, debug, and vanilla kernel options. During testing, installed by default was the -desktop kernel when installing the openSUSE 42.1 Leap Beta and thus that was what ended up being used in last week's Linux distribution comparison, although the Wiki page indicates that the -default kernel should be the default for desktops and servers. The 4.1.6-10 kernel was used when testing the -debug, -default, -desktop, and -vanilla kernel images obtained from the official openSUSE Leap repository.
This six-way Linux distribution comparison is looking at the out-of-the-box performance of this set of popular Linux distributions while using the default package sets and running all tests on the same system. For this comparison an Intel Xeon E5-2687W v3 Haswell system with 16GB of DDR4 memory, 80GB Intel SSD, and AMD FirePro V7500 graphics were used for benchmarking.