Returning...

After a relatively long period of exams and projects I'm about ready to resume working on the project.

I've suspected for a while that a big part of the reason I was getting a lot of variance in my benchmark results was that I'm running them on my i7 laptop with Intel Turbo Boost. I downloaded a small intel gadget that shows your current processor speed (which isn't the one reported through eg Process Explorer) and it showed my clock speed changing at least once a second. Although it should certainly be possible to benchmark on this machine, I'm afraid I'd either have to reboot after every test, run really long tests, run in a temperature controlled environment, etc.

I decided it might be a good idea to try benchmarking on another machine, so I spent most of yesterday installing Fedora 16 on a Core2Duo laptop (I won't be working directly from this machine but controlling it through SSH and SCP which makes Linux preferrable over Windows). I won't go into the details of yesterday, but just imagine the fun I had with a BIOS that seems to be broken for USB boot, resizing partitions, corrupt MBR's, a machine that refuses to boot even from CDROM, finding out "not enough free space" meant I crossed the limit of 4 primary partitions per MBR based drive, programs crashing while they're partitioning my hard drive, and lots more. However, All's Well That Ends Well and I hope to resume doing more useful work tomorrow or the day after.