Socket 478

First up is the P4 platform, these tests were performed by Lazyman.

So how does this Samurai compare to other coolers? A Samurai deserves a formidable opponent; a cowboy would be appropriate I may say. The Cowboy this time is the MCX-478-V from Swiftech.

CPU: P4 2.8 “C” and overclocked to 3.4ghz using 1.7v vcore


(H) = fan speed high, (L) =fan speed low, the 0 value is used when the system crashed during testing when overclocked.

What can I say? The cowboy is wounded by the Samurai sword with 3 C higher. May be we did not give the cowboy the Colt 45 larger caliber six-shooter (a faster Tornado fan) which he handles well. Nevertheless, they both have the 80mm fan with blades tuning in similar speed and the Samurai is ahead.


Socket 754

Piotke installed the Samurai on his A64 setup and compared it to the current reigning champ, the Thermalright SLK-948U. The Thermalright was equipped with a low noise Papst fan, the 8412 n2gm to be exact. It spins at 2700rpm, which places it between the Samurai’s fan High and Low setting.

CPU: A64 3000+ overclocked to 2.4ghz using 1.55v vcore


(H) = fan speed high, (L) =fan speed low

In the graph above you can also see the dBA rating of the fans, the Samurai offers a nice balance when running at its most silent setting, quite a good showing against Thermalright’s top contender.


Socket A

Last but not least, Piotke installed the HSF on one of the more popular budget platforms; the opponent in this test is the new Alpha S-PAL8055F which we will be reviewing in detail very soon. The fans used with the Alpha:

High = Delta AFB0812SH-F00 (80x80x25.4mm) - 4000RPM / 40db(A)

Low = Sanyo Denki 109P0812M701 (80x80x15mm) - 2000RPM / 21db(A)



At “High” fan settings the difference is negligible, but when going for silence the Alpha leads the way, but wait... since we are not using the same fans it is essential we also take a look at the data below:



This explains a lot, the Samurai runs a lot quieter then the Alpha and even at “High” fan settings it’s more silent!

Onto the conclusion ->