We started our new review series dedicated to mainboards on AMD A85X chipset for Socket Fm2 processors with a not very traditional article called “FM2 Platform: Gigabyte GA-F2A85X-UP4 Mainboard and AMD A10-5800K Processor Review”. In fact, the part discussing the actual mainboard is quite common. Yes, we started using new Microsoft Windows 8 operating system, partially updated the benchmark list, but nothing really changed in the overall perspective. The uniqueness of this article is that most of it is dedicated to the specifics of the new platform and AMD A10-5800K processor in particular. We were disappointed that in the nominal mode the system consumes about the same amount of power as a testbed with one of the LGA 1155 mainboards and an Intel Core i5-3570K processor offering incomparable performance. An even more shocking discovery was lying ahead. It turned out that an overclocked Socket FM2 platform can consume even more power than that!
We decided to figure out which of the platform components is the one to blame for making it some incredibly energy-inefficient and which one contributes the most to the performance increase. It is a known fact that AMD processors are not the best in terms of power consumption and performance ratio, so it didn’t come as a surprise that computing cores bear the principal responsibility for the high system power consumption. AMD A10-5800K overclocking boosts the computing cores performance, but not dramatically, because its nominal frequencies are set at very high values to begin with. At the same time, the graphics performance in games and other applications doesn’t increase at all, while the total power consumption snowballs rapidly.
The experiments with the graphics core integrated into the processor produced very interesting results. It turned out that it is the increase in its clock frequency and not CPU overclocking that determines the increase in the system power consumption in idle mode. Moreover, graphics core overclocking may have a negative effect on performance in computing tasks, although this negative effect is minimal and may be neglected altogether. However, it was quite disappointing that graphics core overclocking doesn’t serve its true purpose effectively – to ensure the performance boost in games and graphics applications. The acceleration proved to be minimal and could only be noticed in diagrams and tables, but not during the actual system operation.
However, the main character here turned out to be the memory. All our previous reviews showed that faster system memory had very little effect on system performance. This is an unbreakable rule that applies to all platforms no matter what their functionality is and who is the developer. We investigated whether it made sense to select faster memory for LGA 1156 Clarkdale processors, which memory was the best for Socket AM3 Phenom II X6. We compared the modules for LGA 2011, and which modules would be the best choice for LGA 1155 processors. The specific recommendations could vary, but the overall conclusion remained the same: one shouldn’t neglect high-speed memory, but it could be wrong to overestimate its potential. Why has everything changed for Socket FM2 all of a sudden?
At first I could say that not everything has changed, because the new platform wasn’t developed from scratch and therefore inherited some aspects from the previous systems. The performance gain from using high-speed memory is often quite minimal and the specific numbers in most cases depend heavily on the load type created by the application. The gain is barely noticeable in computing-intense applications, more noticeable during data archiving, but most of the potential can be unveiled only in graphics applications. The thing is that the graphics core uses part of the operating memory for video memory and its higher speed provides this phenomenal performance boost. Namely, the speed gets bumped up by at least one third when the memory frequency changes from 1333 MHz to 1867 MHz. It is also extremely important that despite this significant performance increase, the power consumption doesn’t rise.
Despite the overall positive feedback the Socket FM2 platform received, we were quite disappointed with the AMD A10-5800K processor. Substantial power consumption and low computing power are definitely not the features to be proud of. However, we didn’t really have high expectations in this aspect, but merely hoped that the relatively powerful graphics core would outweigh the shortcomings and make it look good. True, integrated AMD Radeon HD 7660D graphics is quite capable of replacing entry-level discrete graphics accelerators, which will help you to put together an inexpensive non-gaming computer system, which could still be used for occasional gaming. However, this system will never become a fully functional gaming platform, because its graphics sub-system is not powerful enough to satisfy the needs of contemporary 3D games, even at the lowest resolution and with the minima image quality settings.
So, now that we know all this, we can proceed with a more traditional mainboard review. Asus mainboard have long been primary rivals to the products from Gigabyte, therefore our choice of the next review subject was actually quite predictable. Gigabyte GA-F2A85X-UP4 mainboard is the top model in Gigabyte’s Socket FM2 line-up that is why for the sake of a fair comparison we selected Asus F2A85-V PRO mainboard, which tops the FM2 series of Asus products.