I'm deleting posts related to that change. The point of that benchmark is not to try to optimize the benchmark but to compare performance on the exact same program between several CPUs. If you want to discuss optimizing an old version of a stupid raytracer benchmark to improve your times, great, that's completely useless but feel free to use your time however you want.
CPU: i7-4700MQ @ 2.4GHz Turboboost: 3.5GHz, from Intel XTU
OS: Windows 8.1
Time: 9 minutes, 5 seconds
Screenshot:
http://imgur.com/71UgJlG
Also
CPU: AMD FX-8120 @ 3.1GHz, 4.2 GHz OC
OS: Windows 8.1
Time: 15 minutes, 0 seconds
Screenshot:
http://imgur.com/lVQZPbD
Pushing my computer a tiny bit more every time...
CPU: Core i5 3570K (3.4 GHz) OC'd to 4.4 GHz
OS: Windows 7 x64
Time: 9 minutes, 29 seconds
Screenshot: http://i.imgur.com/hxiyybK.png
Edit: Stock fan is proving to be a beast
CPU: Core i5 3570K (3.4 GHz) OC'd to 4.6 GHz
OS: Windows 7 x64
Time: 9 minutes, 2 seconds
Screenshot: http://i.imgur.com/saaF26J.png
CPU: i5-3320M 3.10GHz
OS: linux 3.12.7-2-ARCH
Time: 13:53
Screenshot:
http://imgur.com/ud9btIh
Edit: changed cpu clock as I've failed at cpu monitoring
CPU: Intel Core i7 4770K OC; 4300MHz (43x100)
OS: Windows 7 Ultimate (x64)
Time: 6 minutes, 59 seconds (419 seconds)
Screenshot:
http://i.imgur.com/7f8K4xD.png
I got an AMD tester on board.
CPU: AMD Athlon X4 750K @ 3.4 GHz
OS: Windows 7 x64
Time: 18 minutes, 26 seconds
Screenshot: http://i.imgur.com/9AVPl9v.png
If I'm not mistaken, most motherboards by default will simply match the CPU cache ratio to the CPU core ratio. The main reason to adjust the value is to set it
lower so that you can achieve a higher core ratio. Nonetheless, I forcibly set my CPU cache ratio to 44 to match my core. The results weren't particularly exciting (well within the margin of error).
CPU: Intel Core i7-4770K @4.4GHz
OS: Windows 8.1 Professional x64
Time: 6 minutes, 51 seconds (411 seconds)
Screenshot:
http://i.imgur.com/wRi8G6J.png
Someone could adjust their cache ratio to various values to see what difference it makes in this benchmark, although it probably wouldn't be terribly useful. Setting the cache ratio above the core ratio would likely provide little to no performance benefit, and setting it lower would only be helpful to the small subset of people that actually care about lowering their cache ratio to maximize an overclock.