LinuxBIOS at Four
LinuxBIOS runs on a wide range of platforms. Fifty supported motherboards are in the source tree, but we have found that many motherboards are so similar that a LinuxBIOS for one motherboard can work on another. Companies build code for one motherboard, run it on another motherboard and do not always get around to telling us.
LinuxBIOS works on 64-bit and 32-bit CPUs. CPUs supported include the Alpha, K8, K7, PowerPC, P4, PIII, PII, Cyrix (VIA), Geode (now AMD) and SC520 (AMD). Chipsets are too numerous to list. Form factors of mainboards range from the smallest PC/104 systems to the largest K8 systems. An IBM PPC 970 port is in progress.
One of the most common phrases we heard from chip vendors in the first few years was “we'll never tell you that.” “That” being CPU information, chipset information, motherboard information or any combination of the three. The designs for these three systems constitute highly guarded secrets. It seems amazing, even now, that vendors are able to let us build a GPLed BIOS that by its nature exposes some of these secrets.
How was it possible for us to get this type of information? Simple, businesses are not charities. If there is no business case for releasing this information to us, they do not do it. If, however, there is a business case, then it happens—sometimes with astonishing speed.
From what we can see, the two factors in our success were competition and the creation of a market. Competition gave us a wide variety of choices as to motherboard, chipset and CPU. Once there was a reasonable market, vendors were concerned about being left out.
The experience at LANL is revealing. LANL's last two large cluster RFPs have specified LinuxBIOS as a mandatory requirement. Spending on these RFPs has come in at over $19 million US. Companies that had decided not to become involved in LinuxBIOS could not respond to these RFPs. Companies that had the foresight to get involved in LinuxBIOS early in the game were equipped to respond. Foresight, in this case, conferred a competitive advantage.
LinuxBIOS has come a long way in four years—as one person put it, from “I'm Possible” to “In Production”. LinuxBIOS is used on everything from the largest Linux clusters yet built to the small—test instruments, MP3 players and portable clusters.
LinuxBIOS makes it possible to build systems without PC hardware baggage. The systems can be optimized for Linux and thus can be more compact and simpler. There is increasingly a business case for such systems.
LinuxBIOS is now in its second version, with four years, at least six CPUs and over 50 motherboards' worth of experience behind it. It now takes only days in some cases to do a port to a new system; originally, it took months. LinuxBIOS' impact on the world of computing is only beginning.
So many people have contributed to LinuxBIOS that it is easy to slight them by listing some and not all. Nevertheless, a few contributors stand out as having made LinuxBIOS possible. First, of course, is Stefan Reinauer and the OpenBIOS effort; Jeff Garzik, who got the STPC BIOS Project set up on SourceForge as FreeBIOS; Ollie Lho, who did so much to get our first workstation platforms going in 2000; Steve James and Linux Labs, who worked with us and expedited the shipment in 2000 of our first LinuxBIOS cluster; Greg Watson, who did the PowerPC port; and Eric Biederman, who has done so much to get our really hard platforms up and stable and who has done so much to create version 2.
This paper is released under LAUR 03-8165. This research was funded by the Mathematical Information and Computer Sciences (MICS) Program of the DOE Office of Science and the Los Alamos Computer Science Institute (ASCI). Los Alamos National Laboratory is operated by the University of California for the National Nuclear Security Administration of the United States Department of Energy under contract W-7404-ENG-36.
Ronald G. Minnich has been working in high-performance computing and clustering for 15 years. He recently realized that one of his first clusters, a 16-node SPARC cluster, has a total power equivalent to one-fourth of one of the 2,048 processors in his newest cluster; his new cluster has 10,000 times the power of his first one. Ron started working with UNIX in 1976, with Linux in 1993 and built his first PC cluster in 1994.
|Non-Linux FOSS: libnotify, OS X Style||Jun 18, 2013|
|Containers—Not Virtual Machines—Are the Future Cloud||Jun 17, 2013|
|Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer||Jun 12, 2013|
|Weechat, Irssi's Little Brother||Jun 11, 2013|
|One Tail Just Isn't Enough||Jun 07, 2013|
|Introduction to MapReduce with Hadoop on Linux||Jun 05, 2013|
- Containers—Not Virtual Machines—Are the Future Cloud
- Non-Linux FOSS: libnotify, OS X Style
- Linux Systems Administrator
- Validate an E-Mail Address with PHP, the Right Way
- Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer
- Senior Perl Developer
- Technical Support Rep
- UX Designer
- RSS Feeds
- Introduction to MapReduce with Hadoop on Linux
Free Webinar: Hadoop
How to Build an Optimal Hadoop Cluster to Store and Maintain Unlimited Amounts of Data Using Microservers
Realizing the promise of Apache® Hadoop® requires the effective deployment of compute, memory, storage and networking to achieve optimal results. With its flexibility and multitude of options, it is easy to over or under provision the server infrastructure, resulting in poor performance and high TCO. Join us for an in depth, technical discussion with industry experts from leading Hadoop and server companies who will provide insights into the key considerations for designing and deploying an optimal Hadoop cluster.
Some of key questions to be discussed are:
- What is the “typical” Hadoop cluster and what should be installed on the different machine types?
- Why should you consider the typical workload patterns when making your hardware decisions?
- Are all microservers created equal for Hadoop deployments?
- How do I plan for expansion if I require more compute, memory, storage or networking?