Athlon II Or Phenom II: Does Your CPU Need L3 Cache?

Test Setup And Efficiency Considerations

Efficiency Considerations

Typically, we would measure idle and peak power consumption and look at the system efficiency by tracking power required to complete a specific workload (typically PCMark Vantage). From this, we could calculate the resulting performance per watt. However, we had to modify several variables which wouldn’t normally appear in real life. We reduced the Phenom’s clock speed and had to disable Cool’n’Quiet so the Phenom II X4 965 would run at a constant 2.6 GHz instead of its 3.4 GHz default. Since the slowest Phenom II X4 starts at 3.0 GHz, hardly anyone would run it at decreased speeds. In addition, we decreased the Phenom II’s memory speed to DDR3-1066 to match AMD’s specification for the Athlon II X4.

Now, there is a measurable power consumption benefit if you opt for a processor without L3 cache. Cache alone accounts for roughly one-third of the processor’s silicon real estate. This becomes clear when you look at the processor’s power ratings. AMD rates the Phenom II at 95W to 140W while the Athlon II X4 is a 95W part. The peak power of our test system equipped with the Phenom II X4 965 at 3.4 GHz hit 226W while the 2.6 GHz Athlon II X4 maxes at 170W.

Idle power tells a similar story. We measured 84W for the Athlon II X4 620 and 85W for the same system running a Phenom II X4 965. In these cases, Cool’n’Quiet was enabled, so processors throttled the voltage and clock speed to as little as 800 MHz. Since most sections of the processors that aren’t in use are largely switched off, idle power doesn’t differ much between our contenders.