Testing & Performance :

For comparison purpose, I am using an "odd" or the forgotten heatsink which has been serving me well for over a year and a half.



JMC does not have any A64 coolers, the Phoenix 70 is for P4 S478 which performs admirably in my Prescott system for a long time exceeding a few well known coolers such as Thermalright SLK948 and Swiftech MCX-478V. It performs closely to the famed XP-90, and only replaced by the X-500 I tested recently.

Some P4 S478 heatsink will fit onto the A64 retention bracket mimicking S478 mounting, so on went the Phoenix 70 without a problem.


Testbed:

A64 Test Setup - Ambient temp during test ~26°C
CPU	AMD A64 Venice 3000+ @ 2.65 & 2.71 GHz 1.55 & 1.62 Vcore
Mainboard	DFI nF4 Ultra D
Memory	2*256Mb PC4400
Other	PCX6200 softmod to 6600GT Two Optical drives 80 GB PATA Hard Drive 3Com modem 2x80 mm exhaust; 2x80 mm intake; 1x80 top blower Kool Solutions Chill Vent CCV-28 AS5 TIM is used for all tests X-Dreamer case

Ambient temp = Room temp, NOT system temp



My fellow reviewer piotke is on his "second tour" of duty here with [M]; I like to follow his recent Cooler review setup in the areas of room temp, Vcore and CPU speed for two reasons; 1) a certain continuity for our readers and 2) welcome back piotke :)

With the continuous 90°F (>30°C) weather in the Mid-West U.S.A., the temp stress test is revised to using Battle Field 2 instead of the boring Prime95. I've got to have some fun, don't you agree? Besides, playing games will also increase case temperature when both CPU and GPU are working hand in hand. Tried as hard as I could to maintain 26° room temp to +/- 1°C. All tests are performed with case panel close. And, yes during the test, room temp did fluctuate in between 26 and 27°C. Although I do have the readings in decimal point, I figure since I am not a scientist and operate in regular household the decimal point reading could well be laughable: D Additionally, if I care so much for my rig to be unstable because of 0.5°C, I'd better go see a doctor.

Test 1 - Vcore 1.55, 2.65Ghz

CoolJag SFO-H with standard 70mm fan

CoolJag SFO-H with modified 80mm fan mount

Phoenix 70



Test 2 - Vcore 1.62, 2.71Ghz

CoolJag SFO-H with standard 70mm fan

CoolJag SFO-H with modified 80mm fan mount

Phoenix 70

The SFO-H never did reach the rated higher fan RPM. 4687 rpm is the highest during test which is quite slower than the rated 5800, which might explain the slightly higher load temp. There is negligible different between the use of 70 and 80mm fan.

Sound level in this test is irrelevant since the tests are done in a cased environment where PSU, case fans, graphic card, HDD and CPU fan should be considered independently. Anyway, I did record the noise level about 18" or 45 CM away (from front of the case), dBA readings were taken when the house A/C is not cycling.



What I think of the cooler is next ->