What's new
  • Please do not post any links until you have 3 posts as they will automatically be rejected to prevent SPAM. Many words are also blocked due to being used in SPAM Messages. Thanks!

AMD Richland Review; A10-6800K & A10-6700 Benchmarked

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,264
Location
Montreal
IGP Gaming Benchmarks (Super Street Fighter IV / Torchlight)

IGP Gaming Benchmarks (Super Street Fighter IV / Torchlight)


As with previous in-game tests, we have selected a number of games for our IGP-only testing suite. As you may expect, these benchmarks are run without a discrete card installed. All applications are tested at moderate detail settings (remember, these aren’t fully fledged discrete cards so we can’t expect miracles) in both DX11 and DX9 environments at a resolution of 1080P.



RESULTS: Even compared to Intel's Haswell processors, Richland is in another league altogether. It uses one of the first IGPs that can actually be considered gaming-grade at 1080P provided detail levels are kept to reasonable levels. This makes us seriously excited to see what Kaveri can bring to the table.
 
Last edited:

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,264
Location
Montreal
System Power Consumption

System Power Consumption


Our power consumption numbers are broken down into two categories: one which simply stresses all of the CPU cores with WPrime and another which puts a high amount of load on both the CPU cores and the IGP. The latter will only be included if a given processor includes a dedicated internal graphics sub-processor.

For the CPU power consumption test, we use the standard testing system (with an NVIDIA GTX 670 installed) and wait until the system and discrete GPU are at idle speeds in order to log the idle power consumption. After this, WPrime 1024M is looped for 15 minutes while the power consumption is logged with a calibrated power meter to determine the peak watts.

IGP power consumption testing follows very much the same route as above but with some changes. First and foremost, the GTX 670 is removed and the video output is run through the processor’s graphics engine. In order to fully load the graphics cores and the primary processing stages within the CPU, we run the Unit Benchmark (in DX9 mode) from Civilization V for exactly 15 minutes.

Please note that after extensive testing, we have found that simply plugging in a power meter to a wall outlet or UPS will NOT give you accurate power consumption numbers due to slight changes in the input voltage. Thus we use a Tripp-Lite 1800W line conditioner between the 120V outlet and the power meter.




AMD claims that Richland APUs have the same TDP value as their Trinity predecessors and from what we can tell, both APU generations lineup pretty evenly in the power consumption department. Granted, the A10-6800K consumed a few watts more than the A10-5800K but with its clock speed increase, we were expecting much more than that.

While AMD seems to have wrangled the 32nm manufacturing process into some increased perf per watt numbers, it just can't compete with Intel's 22nm Tri-Gate technology. There's a desperate need for 28nm and luckily that will come when Kaveri is introduced.


Richland may not be the most efficient thing going right now but at least AMD has improved their positioning somewhat. It is also interesting to see the gap between the A10-6800K and A10-6700. It stands at about 21W which is quite a bit but not the 35W TDP numbers would have you believe.​
 
Last edited:

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,264
Location
Montreal
Overclocking Results

Overclocking Results


Before we get too far into this section, let's bring you back in time to nine months ago when our overclocking outing with Trinity hit a stability wall at 4.77GHz. At the time, this was as far as the architecture would go on air cooling, even after several hours of tweaking. Our A10-6800K sample had slightly different plans.

AMD's solution for overclocking their APU's is blissfully simple: buy a good cooler, increase the multiplier, add some voltage and you're off to the races. Sure, there are several finer-grain controls like Bus Speed modifications that will keep enthusiasts busy for days but the basic process makes overclocking a straightforward, rewarding exercise for newcomers.

For this section, we set aside about two hours, one of which was dedicated to stability testing in order to ensure the final clock speeds you see below go beyond a simple CPU-Z screenshot. Voltage was also capped at an offset of +0.13V, a setting which ensured heat didn't go above the 80C mark on our Noctua NH-U12S cooler.


While just over 5GHz is a reasonably good achievement, our sample was actually able to boot into Windows at 5.4GHz but we had to dial things back in order to achieve complete stability in games. With a bit more tweaking, this particular A10-6800K should have no issue hitting at least 5.2GHz on air and potentially even higher on water cooling.

Our results were achieved with only one sample so actual median overclocking headroom on these chips may be higher or lower than this. However, the nearly straight stock / OC clock speed correlation between our A10-5800K and A10-6800K seems to point towards an ability by AMD to dynamically scale their frequencies in Richland without sacrificing its overclocking ability. That's certainly good news for anyone who wants more performance out of their APUs.

With all of this being said, we're intrigued by how high our sample will clock. Expect additional updates in our forum comment thread over the next 48 hours as we try to squeeze a bit more out of this chip.
 
Last edited:

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,264
Location
Montreal
Conclusion

Conclusion


Richland couldn’t have been released at a better time for AMD. With Intel’s Haswell stealing the headlines and Kaveri still months away, AMD needed something that would draw attention towards their APUs and refresh the eight month old Trinity lineup. The new A-series does exactly that; it may not use a completely new architecture but Richland’s targeted improvements go a long way towards legitimizing AMD’s current positioning.

On the mobile side, Richland has the capability to be a superstar. The lower power TDP thresholds, improved clock cycles, an expanded feature set and low costs should make it a winner for entry level notebooks. Within the desktop market its true goals remain rather nebulous since its improvements over Trinity are quite minor. However, we can’t forget AMD only required a simple clock speed bump to achieve Richland’s 5% to 15% improvement over Trinity without augmenting TDP numbers. By comparison, Intel’s transition to a brand new architecture netted Haswell very similar gains over its predecessor, albeit without frequency changes.

From a performance standpoint, the A10-6800K is an improvement over the A10-5800K, most notably in the graphics processing department. Its higher clock speeds also help out in general processing tasks, the one area where APUs seriously lag behind the competition. Overclocking the 6800K is also a dream come true; just increase the voltage, play around with multipliers and you’ll be running at 4.7GHz or higher in no time. That’s a far cry from the hoops we had to jump through when testing Haswell.

The A10-6700 by comparison is priced in line with the 6800K at $149 yet consumes less power, outputs less heat and performs almost identically. Its lack of an unlocked multiplier may turn some budget-focused enthusiasts off but anyone looking for a perfect mid-range HTPC companion need look no further. Granted, some of the A8 and A6 APUs could conceivably fit the HTPC mold as well but the 6700’s HD 8670D GPU boasts better gaming abilities alongside more available power for AMD’s proven HD decoding features. IGP performance doesn't stop at the levels we saw in this review either since these APUs have proven to scale well with higher memory speeds.


But will these changes tempt current Trinity users to switch over? Not a chance. Anyone with a Llano will likely be in the same boat. Even with AMD’s newfound performance boost, the architecture –like Trinity- still represents a step backwards from Llano in terms of some processing capabilities. This causes no small amount of headaches for Richland when it’s placed alongside similarly priced Sandy Bridge and Ivy Bridge processors, particularly from a performance per watt standpoint. Once the Haswell microarchitecture cascades down into lower price points, things will really start to get interesting.

Without a doubt, AMD’s focus on GPU compute shows great promise with new software developers jumping onboard the OpenCL bandwagon almost daily. However, if there’s any hope for survival their architecture evidently needs some balancing. Currently Richland exhibits the same issues Trinity did: the Radeon cores significantly outpacing everything Intel has at a comparable price point but in general processing tasks, these APUs fall behind all too frequently.

The issue here lies with the Bulldozer / Piledriver architecture which just isn’t properly optimized for running programs using legacy compilers. Intel has realized legacy development tools aren’t going anywhere and Haswell was designed to address this reality. AMD on the other hand moved in the opposite direction and is paying a heavy price. Their APUs are supposedly optimized for tomorrow’s computing environments but Trinity and now Richland are being sold today, in a market containing a healthy mix of software using legacy and newer compilers and instruction sets.

While pricing on the A-series APUs continues to be extremely competitive, their usefulness for many DIY desktop builders is questionable. For budget gamers, the IGP won’t pack enough power, allowing AMD’s FX-4300 to act as a worthy alternative since it provides all of the A10-6800K’s performance and costs some $30 less. That’s money which can be put towards a current generation discrete graphics card rather than on an outdated HD 6000-series part for Dual Graphics compatibility. With that being said, for low power small form factor builds, Richland can offer some great performance in a compact and efficient package.

The A10-6800K is a capable, well-rounded APU which boasts performance increases in several key areas and great overclocking headroom, making it a good choice for small, overclocker-friendly systems that aren't compatible with discrete GPUs. Its sibling, the A10-6700, balances out performance and TDP very well, though its overclocking headroom has been curtailed. As such, Richland may not be the processor architecture that AMD required to take the fight to Intel but it serves perfectly well as a bridge between Trinity and Kaveri.
 
Last edited:
Top