Posted on 11/09/2017 by Mark Filion
Starting today, and for the next few days, five Collaborans, including Gustavo Padovan, Helen Koike, Gabriel Krisman Bertazi, Robert Foss and Andrew Lee will be taking part and speaking at this year's Open Source Summit North America, "a technical conference where 2,000+ developers, operators and community leadership professionals convene to collaborate, share information and learn about the latest in open technologies". If you see one of them walking the halls, please come say hello!
All five Collaborans attending the conference will also be spreaking during the week! Below are the details for each of their presentations.
Did you recently replace your Hard Disk by a Solid-State Drive in your servers? Are you satisfied with the speed? What if I told you that it could go even faster? This talk will look at NVM Express, an interface specification for accessing SSDs over modern buses via PCI Express (instead of the old SATA) that enables the system to capitalize on the low latency and high level of internal parallelism that SSDs have. Helen will explain the technical details around the NVMe specification, its main concepts and the mechanisms that allow it to achieve a high level of parallelism in storage devices. She will also discuss the latest performance improvements to the NVMe protocol, particularly in virtualized environments, and where its support stands within the Linux Kernel.
Finally, it is possible to run Android on top of mainline Graphics! The recent addition of DRM Atomic Modesetting and Explicit Synchronization to the kernel paved the way, albeit some changes to the Android userspace were necessary. The Android graphics stack is built on a abstraction layer, thus drm_hwcomposer - a component to connect this abstraction layer to the mainline DRM API - was created. Moreover, changes to MESA and the abstraction layer itself were also needed for a full conversion to mainline. This talk will cover recent developments in the area which enabled Qualcomm, i.MX and Intel based platforms to run Android using the mainline graphics stack.
What influences a program's performance? Some reasons are quite obvious, like the algorithm implemented and the number of execution cycles, but what about the order in which libraries were linked? Or the shell environment size? Or even the sequence and which compiler optimizations were applied? In fact, modern computer systems include such a multitude of features and options, whose interaction with each other can affect the workload's performance, that it is surprisingly hard to write code that fully benefits from the potential of the CPU. In this talk, we will discuss how small changes in the code and in the execution environment can impact the execution time and how you can use Linux performance assessment tools, like perf and valgrind, to detect and mitigate such pitfalls.
Isn't it time we had all new and exciting devices running mailing kernel? Robust and efficient Graphics & Media stacks are a must for most of these devices, yet mainline always fell behind in these areas. However, with the inclusion of Explicit Synchronization of buffer sharing in DRM, this is all starting to change. Explicit Synchronization is a way to let the userspace control the synchronization of buffers between drivers, allowing for smarter, and thus, more efficient decisions. And now V4L2 is on its way to support it as well! With Explicit Synchronization, Android is now capable of running on top of mainline Graphics, and soon we will achieve similar results on Media side. There are also some extensions that we've been working on DRM to improve the performance in complex usecases, like Android Apps running inside ChromeOS. In the end, these exciting developments will hopefully help broaden mainline usage in the industry!
It is impressive how much time and resources a team can save by using the OBS to manage their packages creation and distribution. OBS is a generic system to build and distribute packages from sources in an automatic, consistent and reproducible way. Andrew Lee will cover the benefits of using OBS, explain some of it features and workflow for all your packaging and releasing needs, like automatically build package from scratch on multiple target distros and architectures, easy access through QA to the developer's repo to generate new images with the changes for testing before integration into the production repo, vcs-like workflow as branch code, send merge requests and review submissions and flexible to connect additional resources to empower the backend worker(builders) even with different architectures. At the end tips on how to setup and optimize OBS will be provided.
Check out the full OSSummit 2017 schedule here. We look forward to seeing you there!
Many thanks to Google for recording all the XDC2017 talks. To make them easier to watch, here are direct links to each talk recorded at…
Last week, I attended DebConf 17 in Montréal, returning to DebConf for the first time in 10 years (last time was DebConf 7 in Edinburgh).…
More progress is being made in the area of i.MX6, etnaviv and Android. Since the last post a lot work has gone into upstreaming and stabilizing…
Ever since Vulkan was announced a few years ago, the idea of creating a Vulkan benchmarking tool in the spirit of glmark2 had been floating…
Debugging graphics performance in a simple and high-level manner is possible for all Gallium based Mesa drivers using GALLIUM_HUD, a feature…
At Collabora one of the many things we do is build Debian derivatives/overlays for customers on a variety of architectures including 32…