As I mentioned in my introduction, one of my biggest concerns when preparing to run the benchmarks for this article was ensuring configuration consistency across all three machines. Given that the hardware in each machine was so similar, this actually ended up being really simple: I used the same hard drive and OS install in each machine! The integrated NICs on the MSI Master LS2 and the Shuttle FB75 ended up being identical (Broadcom NetXtreme) and I compiled in support for the Intel Pro 1000 CT on the Intel D875PBZ.

I'm a big believer in monolithic kernels, so if you're brave enough to dig into my kernel .config you'll notice that support for everything that I require is compiled directly into the kernel and I don't rely on kernel modules. It may be less flexible, but that's just the way I prefer it.

Let's dig into the specific software versions and configuration used in my testing:

• Operating System: Gentoo Linux 1.4.3.13
• Kernel version: 2.6.2 (.config)
• GCC version: 3.3.2 / glibc version: 2.3.2
• Java 2 JDK version: Blackdown 1.4.1
• Apache version: 1.3.29 (./configure --prefix=/usr/local/apache --activate-module=src/modules/php4/libphp4.a)
• MySQL version: 4.0.17 (./configure --with-low-memory --with-mysqld-user=mysql)
• PHP version: 4.3.4 (./configure --enable-track-vars --with-mysql --with-apache=../apache_1.3.29 --with-gd)

Well, that's a reasonable start. I know a lot of our regulars are probably thinking: "Gentoo? Jim, aren't you a slackware lover?" Generally, that is the case. For this article however, I wanted something a little more up-to-date and I was looking for an excuse to give Gentoo a try anyway. I do really like it thus far.

It's important to note that for all compiles the following CFLAGS were used: "-O3 -march=pentium4". Apache, MySQL and PHP were compiled by hand while Blackdown JDK was compiled from portage.

Now that you know a little more about the software installed on my test machines, let's break down the benchmarks I plan on running and what I hope to accomplish with each.

• Kernel Compiling: This is a favorite. We'll abuse make -j and see how hyper-threading impacts compile performance.
• BladeEnc: An mp3 encoder parallelized by using MPI.
• Apache Bench: Bundled with the apache source distribution. We'll take a look at Static HTML web-serving performance.
• MySQL Super Smack: Curious how hyper-threading impacts database performance?
• VolanoMark: Java-based chat room benchmark which should hopefully give us a sense of overall Java server performance.

On to the pretty Excel graphs!