The CFD HOWTO Page

Here you can find information on the different things we do in the lab and how we do them.

The Netboot HOWTO - by Stefan Black
Netbooting is a great way to administer both workstations and Beowulf clusters. The basic idea is that a netbooted computer gets its kernel and network information from some other pre-existing system. In the case of a Beowulf cluster, this system is often known as the "head box". This operation is extremely efficient for Beowulf clusters since it allows for totally "diskless" system operation. Furthermore, a single head box can easily provide boot information for 16 simultaneously booting compute nodes, simulating conventional hard drive performance and providing a large savings in terms of hardware.

The CVS HOWTO - by Dr. Robert McLay
CVS, or Concurrent Versioning System, allows multiple developers to work on a single project without blowing away each other's changes. It is similar to the older RCS system in many ways. Click the link to download a postscript version of Dr. McLay's CVS presentation.

The make HOWTO - by Dr. Robert McLay
The make program is one of the most ubiquitous in any linux computing environment. It can easily be used to compile nearly any type of source code, from the lowliest "Hello World" program on up to the linux kernel itself.

The hash table HOWTO - by Dr. Robert McLay
What can provide an O(1) data lookup independent of the amount of data? The answer is a hash table with a good hash function. Hash tables have been invaluable to the lab for the development of parallel code which runs on PC clusters.

The clustering HOWTO - by John Peterson
Ever wonder what can possibly go wrong when building your own Beowulf cluster? This HOWTO provides several answers to that seemingly innocent question.

---