Before I start, this is not a review, just what I did to my own system. As far as being accurate, I tried to keep everything constant so I and others could get an idea of real world performance difference, but by no means is the a lab or professional review, so you can leave those posts at the door.

My shopping list:

http://bgm.bgmicro.com/prodinfo.asp?...ri=FAN&stype=2
5 80mm Panaflo L1A's

http://www.pcmods.com/list.asp?CtgID=24
Panaflo Tails because I'm lazy

http://www.kdcomputers.com/
Alpha 8045, fan grills, rubber spacers, screws, Sanyo Denki 36cfm/29db, and a Sunon 36cfm/30db

I have an economic case, the Enlight 7237. Nothing fancy because I don't look at it. Cooling included the Vantec CCK6035D with a 7K screamer on it with a Sunon 36cfm/30db fan as exhaust and the standard weak Magic fan as intake. I had enlarged the air holes in the front bezel. The HSF was loud enough to hear across the apartment.
I used both MBM5 and mbprobe to determine temps, and they're accurate compared to what I read in BIOS. Blah, Blah, Blah. That was the standard you can't compare disclaimer.
With a room temp of 21C, Idle was 26C case and 51C cpu
Load consists on 15 minutes of Toast at high Priority.
If you don't think that's enough, try it before you say anything. http://www.overclockersonline.com/?page=downloads

Load was 28 case, and 63C.
This is a standard range for a 1.4GHz Tbird and the Gigabyte GA-7DXR. The cpu temp sensor rest against the back of the cpu resulting in a lot higher reported temps. Usually about a 10C difference compared to a sensor that measures air temp in the socket, this info I got from people who own multiple systems.

Typing next post..............