One of the major advantages Ivy Bridge has over its predecessors is finer manufacturing technology using 22 nm process and 3D transistor structure. By improving the manufacturing process Intel promised not only to increase performance, but also lower the power consumption. This was implied in respect to computational as well as graphics cores. In order to estimate the scale of this improvement, we ran a few of our traditional power consumption tests.
The graphs below (unless specified otherwise) show the full power draw of the computer (without the monitor) measured after the power supply. It is the total of the power consumption of all the system components. The PSU's efficiency is not taken into account. The CPUs are loaded by running the 64-bit version of LinX 0.6.4 utility. The graphics cards were loaded using FurMark 1.10.1 utility. Moreover, we enabled all power-saving technologies to correctly measure computer's power draw in idle mode: C1E and Enhanced Intel SpeedStep for Intel processors, and AMD Cool’n’Quiet for AMD processors.
In idle mode Socket Fm1 systems consume about 4-5 W less than Intel competitors. However, the new Ivy Bridge processors with integrated graphics cores suddenly started to consume much more power than their predecessors. We have already witnessed a similar situation when there is a discrete graphics accelerator installed into an LGA 1155 platform.
Once we load the graphics cores of the processors to their maximum, the situation changes dramatically. AMD processors become more power-hungry, and the power consumption of the Intel HD Graphics cores becomes directly proportional to their performance. In other words, Ivy Bridge consume more power than their predecessors during graphics 3D load.
However, when both cores - processor and graphics – are fully utilized, Ivy Bridge does become more energy-efficient. While the difference between Ivy Bridge and Sandy Bridge was not too serious, when we tested the same processors in systems with discrete graphics accelerators, the ones manufactured with new production process demonstrated their energy-efficiency more confidently. As for AMD solutions, they can’t compete against Intel in performance-to-power consumption ratio. As we know quad-core Llano processors don’t get even remotely close to Core i5 in computational performance, but their power consumption is significantly higher.
Another example of complex operational load is HD video playback. The numbers here are very close to what we have just seen in idle mode, the biggest difference being 7-8 W more. Hardware video decoders built into the graphics core allow the computational cores to remain in power-saving modes, that is why all tested systems remain pretty energy-efficient in this case. Although I have to point out that processors with a lot execution units (HD Graphics 4000) or shader processors (Radeon HD 6550D) in their cores consume much more power.