I’ve recently been building a new computer system and one of the things I have been trying to work out is the overclocking of the system.� In my search for benchmarks and stress tests I came across a popular benchmark in the overclocking community that I thought was pretty neat.� Ever wonder what the 21,865,285th digit of PI was?
Super PI is a program that will calculate PI to various numbers of digits (up to 32M).� The standard benchmark is to run it for 1M digits and compare your time to other computers.� My current CPU (Intel E6600�@ ~2.9GHz) did it in a little over 17 seconds, and my 64-bit workstation at work (Intel Xeon 5150 @ 2.66GHz)�does it in 19.781 seconds.� I’ll add a comment to this once with my final�1M time once I finish overclocking my system and�verify its stability.
There is also a wikipedia entry on the program here.
With the final system setup, I have my Intel E6600 overclocked from 2400MHz to 2835 MHz and my memory overclocked from 800MHz to 1070MHz (5-5-5-15) and I can run the 1M calculation in 17.781s.
On my system, an Athlon XP 2400+, I’m seeing 1m22s. I went through the Services panel and shut down all the services that weren’t (obviously) critical, along with System Tray programs. This knocked of… 1 second! Which makes me happy that they’re affecting the load. Unless the benchmark sets its process priority to high somehow. Time to upgrade the CPU and see the difference.