Share on Twitter
November 4, 2012
In these challenging times for silicon chip maker AMD, and as part of its initiative to reduce its operating expenses by
cutting a lot of its excess staff, Advanced Micro Devices has closed its Dresden, Germany-based Linux research center (OSRC).
The lab focused mainly on the enterprise development of Linux kernels optimized for AMD microprocessors. The OSRC was founded
in April 2006 and was acting as a go-between the Linux operating system development community and the global AMD processor design
AMD's OSRC played a particularly critical role in ensuring the support of the next-generation AMD products and solutions
OSRC was specialized mainly in OS virtualization, cloud integration, memory management,
multi-core scheduling and optimal performance measuring tasks to make better use of future multi-core architectures.
Generally speaking, the laboratory was overseeing security patches to support AMD’s server processors’ features in Linux
operating systems such as Red Hat and CentOS.
The OSRC had about 25 employees who helped integrate important changes for new AMD platforms into Linux distributions
such as Red Hat Enterprise Linux (RHEL) and also Suse Linux Enterprise. Some of the OSRC kernel developers also worked on
open source virtualisation solutions such as the Xen hypervisor.
But some of the Linux developers from AMD Research are also located in Austin, Texas. From now on, the U.S. team will have
to incorporate support for all functionality of AMD code-named Steamroller and Excavator high-performance as well as Jaguar
low-power cores into the Linux operating system. This should be implemented in the next three to four months said AMD.
Overall, AMD plans to lay off about 15 percent of its workforce this quarter to reduce operating costs, and in an effort
to save $1.3 billion in 2012. If that happens, the company could reach the break-even point next year.
In other Linux news
Ask just about any software developer or system integrator and most will tell you that everyday is different in the field
of information technology and application development. There's always something new to learn and that's an important ingredient
that keeps them motivated and excited.
But there's one area where they've been lacking a lot of interest lately and that's with the GNOME project. If you're new
to Linux, GNOME is a desktop environment and graphical user interface that runs on top of a computer operating system. It's
composed entirely of free and open source software, just like the Linux operating system is.
GNOME is an international project that includes creating software development frameworks, selecting application software
for the desktop, and working on the programs that manage application launching, file handling, along with window and task
But lately, GNOME has been losing a lot of key players and some in the Linux community are taking this as a major concern.
Not that Linux is dying-- on the contrary. Linux is very alive and well, growing in popularity at more than ten percent a year,
especially in data centres and in mission-critical enterprise applications.
And that's not about to change, on the contrary.
But the GNOME project has hit many road blocks in the last year, and could be ready for a major falling out with developers
after the release of GNOME 3.
At the core of the issue is the radical rewrite of the whole desktop. And that's where the majority of the Linux community
draws the line.
So it's not just a few unhappy users that are ditching the project. Linux application builders and software developers are
also leaving as well.
For instance, GNOME project developer Benjamin Otte notes that core Linux developers are leaving GNOME development in droves,
and that the project is extremely understaffed, has no specific goals and is losing market share at an alarming rate.
Of course, not everybody agrees with Otte, but the fact remains that he is a well respected member of the Linux community,
and as such, many people are following his comments with interest.
But it gets worse-- in the last week, two major Linux distributions have jumped ship entirely, preferring to create their
own desktops. Canonical's Ubuntu team has the Unity desktop and popular newcomer Mint Linux has created not one, but two new
desktop projects. So Otte does have a point here.
If that wasn't demoralizing enough for GNOME developers, Linus Torvalds himself called GNOME 3 an unholy mess, going on
to add that he's never met anyone who likes it. And after all, Torvalds IS the creator of Linux.
At times, Torvalds is known to have a few outbursts here and there, but taken together, all those rumors of GNOME being in trouble
are starting to paint a picture that is looking more desperate with each passing day.
And to think that GNOME 3 was originally rejected by the project developers themselves says a lot about the desktop. The
complete re-imagining of the application that was to be GNOME 3 was initially set aside because it went against the underlying
philosophy of the GNOME development community-- incremental improvements.
And make no mistake, it was incremental improvements and a little of 'rocking of the boat' that brought many users to GNOME
in the first place. But now the situation has gone out of hand.
But eventually, GNOME 3 did launch and the project made a fundamental mistake that has now cost it not just mindshare,
but marketshare as well. And to throw the proverbial fuel on the fire scenario, the GNOME 3 developers decided to abandon
the current users it did have to chase other users it didn't.
"GNOME itself is like the protagonist of a romantic comedy, chasing someone it will never catch even as it misses what
was always there-- its old but faithful user base."
And what makes that decision all the more confusing is that, as Otte points out, GNOME is chasing users that are moving
to devices GNOME doesn't even currently work on-- tablets and smartphones.
So the question is, exactly what happened that made GNOME developers seemingly abandon all sense of sanity and in the end,
managed to design a desktop interface that no one wants?
And of course, it's very easy to compare GNOME to KDE and its transition from KDE 3 to 4, which was similarly disruptive
to work flows and generated a similar amount of negative press. But while KDE 4 may have been a bumpy ride initially, it was
always very clear where KDE 4 was headed-- it just took a bit longer to get there but it did in the end, and that's what really
counts in the field of software development.
The more likely candidates for inspiring GNOME 3's demise is Apple's iOS and Google's Android operating systems. The touch
screen came like a blinding white light that obliterated developers' interest in anything as mundane as the desktop and laptops
people use to do actual work every day of the week.
All of a sudden, developers everywhere were seized by a kind of touch-screen mania that has warped computing visions
not just in the GNOME world itself, but in the Windows world as well.
Windows 8 exhibits the same kind of iOS/Android envy that GNOME displays, it just manifests it in a different design. To be sure,
Apple created its very proprietary iOS operating system exclusively for the iPhone and the iPad. It didn't rewrite the OS X
desktop that runs on its popular Mac desktop computers and laptops, nor did it try to re-imagine the desktop computing paradigm.
Apple did what it does best and created something entirely new that was always designed with touch screens in mind. And the iOS
was uniquely designed for smartphones and tablets-- nothing else. End of the story.
What's really sad about GNOME is that it *had* a sizable user base. It had the support of companies well beyond its Red Hat
stable release, including Red Hat rivals such as SUSE and even phone maker Nokia. Both companies have since all but dropped
their commitments to GNOME.
Yet, it abandoned GNOME to chase what its developers call innovation in a segment where Linux users don't want innovation.
In the last six months alone, GNOME 3 experienced a slow but very steady stream of distros dropping it. Ubuntu has moved on to
Unity. Mint continues to pull in GNOME 3 refugees with Cinnamon.
Linus Torvalds himself and a few others have jumped ship for XFCE, KDE, LXDE or countless other small desktops. When Fedora
18 ships in two weeks, it may well be the last of the big distributions to remain with GNOME 3. But the question is, just exactly
how long will Fedora stick with GNOME 3? Will Fedora 19 drop it like a hot patato?
It will be interesting to see in the next few months just what the outcome of all this will be. But one thing is for sure:
GNOME is causing a lot of concerns in the community right now, and a lot of people and key Linux developers are abandoning
the GNOME project for other desktop solutions.
In other Linux news
Linux developers over at Red Hat say they are porting OpenJDK to ARMv8, the 64-bit ARM architecture (also known as A64).
According to Andrew Haley of Red Hat, the change is taking place because the current OpenJDK ARM situation is rather unsatisfactory,
and Red Hat wants to do better with the A64.
As the people at ARM created a new instruction set for the 64-bit version rather than just extending the 32-bit instruction
set, it simply means that to take advantage of the A64, Red Hat needs new compilers and Java virtual machines.
Haley explained that for 32-bit ARM, there currently are two versions of the HotSpot Java virtual machine, one owned by
Oracle which is proprietary and one which is free that was written for ARM.
Also, the Oracle VM with its JIT engine performs better than the free one with its lightweight JIT. "We really don't want
this situation for A64," says Haley. Red Hat will be writing a port which will be entirely free software and will be submitted
as an OpenJDK project which others can take part in.
The Linux developers have already faced one issue-- no real hardware exists for ARMv8 yet so they have created a simple CPU
simulator to use for development testing.
But they do expect to find some answers when they run on real hardware for the first time. The template interpreter, the
foundation for any HotSpot port, is almost completed and Haley says they hope to finish it by Christmas so that it is ready
for preliminary public access.
They are unable to release anything at present as the developers are under a non-disclosure agreement, but Haley says that
should be lifted soon.
This interpreter release will be a complete implementation but it won't be fast. There are two different compiler technologies
to be worked on, C1 (the client JIT) and C2 (the server JIT).
The client JIT compiler wasn't initially in the developers' initial plans, but it's easier to write and should help the team
learn more about the new CPU architecture, which they will be able to apply when they develop the server JIT.
It should also open the way for tiered compilation where code is first compiled by the client JIT and heavily used methods
are then recompiled again by the server JIT.
Haley notes that "to the best of our knowledge, there hasn't been a full port of HotSpot done as an open project since OpenJDK
was freed five years ago" and that makes the project both exciting and a little scary. "But we're confident that this is going
to work" he concludes, inviting comments and questions via the firstname.lastname@example.org email address as there is no
mailing list set up yet.
In other Linux news
Icinga, which started out as a fork of the open source monitoring software Nagios, is beginning to go its own way with
the upcoming Icinga 2. The developers have released a technology preview of their work to date. This first release of Icinga
2 carries the version number 0.0.1.
The developers will be working on the replacement for Icinga's core framework in parallel with enhancing Icinga 1.x. The
developers are building the new system in C++ and moving to a heavily modularised architecture.
The Red Hat developers' intention is that the complete rewrite will correct some shortcomings of the current version such
as the complicated configuration of the tool and scalability issues in large deployments. The preview is far from complete, as
befits its preview status.
According to the development plan, Icinga 2 should be receiving major features in 2013 such as dynamic reconfiguration,
real time event support, a legacy-compatible RPC interface to make it easier to migrate to the new release, and business process
monitoring. Icinga 2 will also offer some support for Windows platforms.
Overall, the technology preview of Icinga 2 runs on Windows and Unix-like systems such as Linux. Source code is available for
download from Sourceforge, licensed under the GPLv2. The technology preview is only meant for testing purposes; for those
who wish to install network monitoring in their organization now, the most recent stable version is Icinga 1.8.
Source: American Micro Devices.
Share on Twitter
All logos, trade marks or service marks on this website are the property of their respective
companies or owners.
Linux News Today.org is read by over 450,000 people involved in the field of Linux application development,
professional Web hosting services, Linux
security, Linux Web development, etc.
Inquire about our reasonable advertising rates
on our news website. One of our advertising representatives will be in touch with you. Simply email us to learn
about our ad rates and how we can help drive relevant traffic to your website. Advertising space is limited.