Page 2:Fine Tuning Power Management
Page 3:Test System
Page 4:Update: A “Broken” Cool'n'Quiet Implementation
Page 5:Assessing Voltage
Page 6:Tweaking And Undervolting
Page 7:Voltage Ramps, Continued
Page 8: Measured Power Consumption And Methodology
Page 9:Test Results: Idle
Page 10:Test Results: Idle, Continued
Page 11:Test Results: Load
Page 12:Voltage Ramp And Power Consumption
Page 13:Voltage Ramp And Power Consumption, Continued
Page 14:Phenom X4 955 And Conclusion
Editor's Note: This story first went live in November of 2009. After its publication, we went back and forth with AMD's power experts, discussing some of the conclusions drawn. With feedback from AMD, we've retested to clarify where optimizations are most likely to be made. We've also addressed a problem with Cool'n'Quiet observed in the previous piece, which turned out to be caused by an implementation of the technology, rather than the technology itself. Special thanks to AMD for its input.
Choosing the right processor is never an easy task. Like other choices in life, there are so many to choose from, and many factors to be considered.
Today, performance and heat are not the only factors we all have to consider. Power consumption is becoming increasingly important. However, most reviews and articles about processors still focus on performance for the most part. Often times, review benchmarking is performed without power management features turned on (an especially flagrant faux pas in the era of Intel's Turbo Boost, which requires those capabilities be enabled).
Is this a bad thing? It doesn't have to be in the context of testing specific theories. Performance is very likely the penultimate factor we all consider when buying a processor, or any other component for that matter. But as concerns about power consumption becomes increasingly important, we should try to always look at performance with power management features enabled.
Many view power management features as detriments to processor performance. Indeed, most reviews and articles that include performance evaluation with power management features turned on usually shows slightly lower numbers than when those features are turned off. However, many do not realize that, just like hardware and software, we can fine tune power management features either for more performance or lower power consumption. It’s not uncommon to tweak the BIOS, drivers, and operating systems for maximum performance, so why not the same with power management features?
Getting to Know Power Management
This article is an answer to that question. In order to fine tune power management, first we have to understand how it works. Power management throttles down the processor (or any other component for that matter), slowing it down to save energy when the workload is low. When a user alters the workload (by opening an application, for instance), power management throttles the processor back up to a higher clock rate until the workload changes again (basically, when the application or task is complete).
Although clock throttling is the most obvious effect of power management, it’s not the largest influence in lowering power consumption. Reducing voltage is what has the most significant impact. By throttling down the clock, processor manufactures can settle in on a much lower voltage for their processor. In addition to lowering power consumption, throttling clock and voltage will also reduce operating temperatures. Less voltage equals less power consumed and thus less dissipated as heat.
Today's processors offer fine-grained power management, able to run at a number of different clock rates and at many voltage levels. These different clock and voltage settings are called p-states (you know them as SpeedStep on Intel CPUs and Cool'n'Quiet on AMD chips). A modern processor can have at least two p-states--one for performance and another for power-saving mode. However, it’s not uncommon to see processors with more than two p-states. In addition to p-states, power management settings also include transition time and a workload level required to switch between p-states.
- Fine Tuning Power Management
- Test System
- Update: A “Broken” Cool'n'Quiet Implementation
- Assessing Voltage
- Tweaking And Undervolting
- Voltage Ramps, Continued
- Measured Power Consumption And Methodology
- Test Results: Idle
- Test Results: Idle, Continued
- Test Results: Load
- Voltage Ramp And Power Consumption
- Voltage Ramp And Power Consumption, Continued
- Phenom X4 955 And Conclusion