A test + new Piledriver microarchitecture Richland-DT: AMD A10 6800K
The new micro-architecture Piledriver + introduces the third generation of APU A-Series, Richland-DT , and corresponds to an improvement over the existing, both the APU A-Series second generation “Trinity” and the CPU FX-Series of second generation “Vishera”; optimizing consumption and performance, achieving higher frequencies. Along with this, it has greatly improved the memory controller driver, increasing performance per CPU cycle as graphics performance, but requiring the use not only of RAM in dual channel configuration [the promise that we will try to prove].
Test was used to evaluate various aspects critical platform are performed in triplicate and the average result is reported. Together will be a platform to Overclock.
|Processor||AMD A10 6800K|
|Motherboard||Asus F2 A85-V PRO|
|Memoirs||KHX1600c9d3t1k2/8G Kingston HyperX (2x4GB) / 8G (2x4GB)|
|Video Card (integrated)||8670D|
|Source of Power||Seasonic M12II-520 Bronze|
|Processor Cooling||Thermaltake Contac 30|
|Storage||Kingston HyperX 120GB|
|Monitor||LT22B350 Samsung LED 22 “1080p|
|Cabinet||Cooler Master HAF 912|
|Operating System||Microsoft Windows 8 Pro|
|Drivers|| AMD Catalyst â„¢ 13.5 Beta
Profile AMD 13.4 CAP 1
|Benchmarks, games and utilities|| Benchmarking HD x264
AIDA64 Extreme Edition v2.400.1900
3D MArk 11
Lost Planet 2
Resident Evil 5
Unigene Heaven 3 and 4
Conditions of Testing
He climbed the processor frequency by 6%.
Processor Arithmetic Performance.
This is critical for applications that make heavy use of floating point calculations and, like most suite of numerical and data analysis.
Super Pi – 32 Million.
Performance nothing out of the ordinary, under Stock performance is comparable to an A8 5600K. When scale overclock performance by 9%.
Note is heat production of this processor in less than 25% of the test we reached both under stock as overclocking (here was really hell), about 55 Â° C, so that often the test failed (make a little magic, forcing cooling to endure the 20-odd minutes of the test).
SiS Sandra Processor Arithmetic
Again, performance is nothing out of the ordinary, matching the performance of this new processor with one of the older generation to overclocking. Arithmetic performance according to this scale test 3%.
wPrime – 1024 Million.
Not too good, not too bad, but for the price is pretty good, owns approximately 40% of i7 4470k stock performance.
We increase the yield from 1.111 GB / s 1.172 Gb / s, which represents an increase of 5%.
Performance bad, thinking that the evolution of the platform is aimed directly at improving this point.
Multimedia Performance Processor.
Very poor performance for expected and overclock scaling on average, less than 2%.
SiS Sandra Multimedia processor.
Very good results, that put him in the top 76% ranking performance as Sis Sandra, and OC achieved a great improvement of almost 10%. But when compared to the A8 A8 5800K or 5600K, its performance is not so good, so as happened in the previous tests, we expected more.
System Tests and Charts.
This point will be measured by some test suites and generally this point will give us an idea of â€‹â€‹the potential of the platform.
Reading speed ASS Disco.
The ideal maximum speed our SSD is 525MB / s, and we achieve the 508-509 MB / s, which represents 97% of the theoretical value, not bad.
AIDA64 reading speed RAM.
Result less than expected, improvements were estimated also to compare test was conducted in conjunction with a lower performance chipset, A75, the result is mainly due to the deficiency of AMD drivers.
AIDA64 RAM writing speed.
Result less than expected, improvements were estimated, ditto reading speed.
11 Direc3D Video Processing
Result lousy, explained why the graphics driver in use is Microsoft Basic Render, since the processor and GPU are not recognized, and the graphics driver is not approved
3D Mark 11
Points lower than expected, it stands out.
Bench HARD 3D info
He was one of the few tests that lacking feature supports OpenGL, the result is terrible, considering that in this test an Asus Ati x1950 delivers 89.1 fps and an old Nvidia Geforce 8600 GTS 42.1 fps.
Monocore behavior is better than multicore, but still bad.
Other graphical tests failed.
Lacking support various test controllers failed in their measurements, using both OpenGL and DirectX 9, 10 or 11, to name Unigene Heaven in version 3 and 4, Lost Planet 2 and Resident Evil 5 STALKER.
Chronicle of a disaster: Drivers / Drivers.
My first approach to this review was pretty naive, just turned on the computer with the processor and motherboard started by default and highlight some shortcomings of this release, the Windows graphics driver worked horribly, but because AMD had received information from I should make updates to the BIOS, I told myself it was just a temporary thing, well, have a resolution of 1024 Ã— 768, not bad, considering that once, when Windows instalabas optabas hopefully have a resolution of 800 Ã— 600, so without thinking 2 times updated the BIOS … But nothing changed.
Well, I said, will be the drivers, so install the AMD Catalyst â„¢ 13.5 Beta, as I had recommended AMD itself, restart and again disappointment, everything was running horrible, but at least had reached maximum resolution 1400 Ã— 900, and still with hopes again followed the recommendations of AMD, install 13.4 CAP 1 profile (something not everyone does, but it is highly recommended), and finally got a resolution of 1920 Ã— 1080 (1080p), but something was wrong, Windows animations went horrible, little fluid graphics and with defects that made me remember the artifacts of video cards, but there was still some faith … but after testing is over, the CPU performance was not of good GPU performance was disappointing, plus the inability to use DirectX or OpenGL, ended up killing a good impression, it’s just a bad product today, and do not say tomorrow, because the great problem of AMD are your drivers, and this review is more than clear. Along with this, to crown the situation, we are informed, just a couple of days, that there are some new drivers recommended. the AMD Catalyst â„¢ 13.6 Beta, but at this point, the die was cast …
Let’s get honest, personally I’m using AMD, and I have an appreciation special good for the brand, because when money is tight, always delivers a product quite competent, but AMD is wrong at the time of launch of the new processors Piledriver +, and A-Series APU third generation Richland-DT, almost in a self-destructive urge again the mistakes themselves the launch of the first generation of Bulldozer FX processors , launches a green product, we will leave bad impressions at the beginning , but their performance can change dramatically by the former and never well weighted theme Controllers / Drivers.
In general we can not speak of a bad product because it is not, but disappointing for what was expected, and I repeat, this performance should be attributed largely to the drivers, but something that is unforgivable is the great heat production , which just surpassed the acceptable times.
- Cryptographic performance is not too bad.
- Arithmetic performance, multimedia and energy efficiency well below expectations, is supposed to be an improvement.
- Drivers, and therefore lack of performance associated with Catalyst and fayas various synthetic tests and real environments.
- Price, the suggested is $ 142, while an alternative as the A8 A8 5600k or 5800k are under 100 dollars.
- Processor Heat Production, Overclock borders on unacceptable, and stock is quite unkind, we recommend a cooling system between good and great to not suffer.
Relationship Performance / Price
For these very poor results, and false expectations generated in this review we will grant you a zippy cardboard, or not at all recommended, but may improve dramatically with some decent drivers.
You can also comment on this review in our forums.Tags: 680K A10 AMD Benchmark, A-Series APU third generation, AMD benchmark, APU A-Series de tercera generación, Benchmark AMD A10 680K, fm2, piledriver, Richland-DT