Q & A : Benchmarks : Performance vs. Power (aka Power Efficiency) |
What is it?
A chart, available in all benchmarks that compares the performance of the device with its power (TDP). It allows us to view how the device’s performance varies with the amount of power it dissipates (and thus consumes) and how competing devices perform for the same amount of power (aka power envelope).
Why do we measure it?
Power Efficiency is very relevant today, in many cases even more relevant than Cost Efficiency as the drive for lower energy usage and lower CO2 emissions come to the fore.
It is also used to estimate long-term Cost Efficiency that is in addition to the intial cost. The long-term cost includes things such as energy consumption, maintenance and other indirect costs (e.g. cost of power supply, cost of cooling system, etc.) that may well be higher over the life of the product than the initial cost of the product!
What do the results mean?
- The result is a ratio of performance (MIPS, MFLOPS, MB/s, etc.) over power (TDP W, etc.)
- Where higher indexes mean better performance (MIPS, MFLOPS, MB/s, etc.) the higher the ratio of performance over power the better the Power Efficiency.
- For timing indexes, e.g. latencies (ns, us, etc.) where lower indexes mean better performance the lower the ratio of timing over power the better the Power Efficiency.
Typical Results from Desktop Processors on the Market
Testing various current desktop processors or just checking out the reference results makes the differences in architectures and implementations very clear.
Rank | Desktop Processor | Performance / Power | Power Efficiency | Commentary | |
#6 | AMD Phenom X3 720 | 28164 MOPS @ 95W TDP | 296.5 MOPS/W | AMD’s design does not offer a very good performance per watt anymore. | |
#5 | AMD Phenom X4 970 | 47463 MOPS @ 125W TDP | 379.7 MOPS/W | Same issue with the quad core version due to a rather high TDP. | |
#4 | AMD Phenom X6 1055 | 56713 MOPS @ 125W TDP | 453.7 MOPS/W | Good performance per watt given the high end six cores design. | |
#3 | Intel Core i3 550 | 38807 MOPS @ 73W TDP | 531.6 MOPS/W | A correct placement for the entry level Core series due to efficient power design. | |
#2 | Intel Core i5 760 | 54388 MOPS @ 95W TDP | 572.5 MOPS/W | Very good performance per watt given the average TDP of this chip, that makes you forget the bad reputation of the P4. | |
#1 | Intel Core i7 975 | 96463 MOPS @ 130W TDP | 742 MOPS/W | Usually high end processors do not have an efficient power design, but this model is the best example to demonstrate the contrary. |
You can check your own processor’s Power Efficiency or view how other processors measure up using the Performance vs Power tab in Sandra’s benchmarks.
Typical Results from Mobile Processors on the Market
Testing various current mobile processors or just checking out the reference results makes the differences in architectures and implementations very clear.
Rank | Mobile Processor | Performance / Power | Power Efficiency | Commentary | |
#1 | Intel Core i7 620M | 36881 MOPS @ 35W TDP | 1053.7 MOPS/W | Intel has the crown also on mobile processors due to the its latest 32nm technology. | |
#2 | Intel Core i3 350M | 27841 MOPS @ 35W TDP | 795.4 MOPS/W | Beating the AMD competitor by a whisker, the entry level Core series places itself into 2nd place. | |
#3 | AMD Phenom II N950 | 27807 MOPS @ 35W TDP | 794.4 MOPS/W | With this product AMD succeeds makes it to the podium with very good performance. | |
#4 | AMD Turion II Neo K625 | 9961 MOPS @ 15W TDP | 664.1 MOPS/W | As performance is not being considered here, this low wattage processor may be a good choice. | |
#5 | Intel Atom N550 | 4545 MOPS @ 8.5W TDP | 534.7 MOPS/W | The dual core design of the original Atom has the lowest performance but an acceptable power draw. | |
#6 | VIA Nano L3100 | 5951 MOPS @ 25W TDP | 238 MOPS/W | VIA is last here given the single core design and above average power consumption. |
You can check your own processor’s Power Efficiency or view how other processors measure up using the Performance vs Power tab in Sandra’s benchmarks.
Most Popular Processors, Chipsets & Memories
Most popular Processors as benchmarked by users (past 30 days): | Most popular Chipsets (and thus Mainboards) as benchmarked by users (past 30 days): | ||||||||||||||||
|
|
||||||||||||||||
For a complete list of statistics, check out the Most Popular Hardware page. For a list of more products, see SiSoftware Shopping. |
Typical Results from Desktop Chipsets & Memories on the Market
Testing various current desktop memory sub-systems or just checking out the reference results makes the differences in architectures and implementations very clear.
Rank | Desktop Chipset / Memory / Processor | Performance / Power | Power Efficiency | Commentary | |
#5 | AMD 890FX / 2x 2GB DDR3 PC3-8500 / AMD Phenom X4 970 |
12800 MB/s @ 19.6W+4.08W TDP (chipset+memory only) | 524.6 MB/s/W | The ratio does not improve when using lower spec memory because the performance is also affected. | |
#4 | AMD 890FX / 2x 2GB DDR3 PC3-10600 / AMD Phenom X6 1055 |
13100 MB/s @ 19.6W+4.8W TDP (chipset+memory only) | 536.9 MB/s/W | A relatively high TDP prevents the high-end AMD chipset to be on a higher place. | |
#2 | Intel H55 / 2x 2GB DDR3 PC3-8500 / Intel Core i3 550 |
13000 MB/s @ 5.2W+4.08W TDP (chipset+memory only) | 1400.9 MB/s/W | An entry level choice with a very good performance per watt. | |
#1 | Intel P55 / 2x 1GB DDR3 PC3-10600 / Intel Core i5 760 |
15700 MB/s @ 4.7W+4.8W TDP (chipset+memory only) | 1652.6 MB/s/W | A good power design easily places this combination on the first place. | |
#3 | Intel X58 / 3x 2GB DDR3 PC3-12800 / Intel Core i7 975 |
26300 MB/s @ 28.6W+7.38W TDP (chipset+memory only) | 730.9 MB/s/W | Because the chipset contains the IOH, despite its top performance, it is on third place here. |
You can check your own chipset/memory’s Power Efficiency or view how other chipsets/memory measure up using the Performance vs Power tab in Sandra’s benchmarks.
Please let us know what you thought of this article by voting using the icons/links below. Thank you for reading.