AMD A8-3850 APU Review: Llano Hits the Desktop

[MAC]

IGP Gaming Benchmarks

IGP Gaming Benchmarks

In order to get an idea of how the new Radeon HD 6550D graphics solution in the A8-3850 APU compares to the competition, we decided to test it against the Intel GMA HD IGP (both 733Mhz & 900Mhz variants) found in the Clarkdale chips, the Intel HD Graphics 2000/3000 IGPs in the various Sandy Bridge chips (6 EUs @ 850-1100Mhz/ 12 EUs @ 850-1100Mhz/ 12 EUs @ 850-1350Mhz), and AMD's fastest AM3 platform IGP, the Radeon HD 4290 (40 shaders/700Mhz core clock/UMA+SidePort DDR3-1333) found in the 890GX chipset. We also threw in two low-priced discrete GPUs from AMD and NVIDIA, the Radeon HD 5550 GDDR5 (320 shaders/550Mhz core clock) and the GeForce GT 430 DDR3 (96 shaders/700Mhz core clock).

As we demonstrated earlier, both overall memory bandwidth and the amount of memory allocated to the IGP greatly affects gaming performance, so for these tests with allocated a 1GB UMA frame buffer size, and also tried three different memory frequency/timing settings.

Futuremark 3DMark06

3DMark06 v1.1.0
Graphic Settings: Default
Resolution: 1280X1024

Test: Specific CPU Score and Full Run 3Dmarks
Comparison: Generated Score

The Futuremark 3DMark series has been a part of the backbone in computer and hardware reviews since its conception. The trend continues today as 3DMark06 provides consumers with a solid synthetic benchmark geared for performance and comparison in the 3D gaming realm. This remains one of the most sought after statistics, as well as an excellent tool for accurate CPU comparison, and it will undoubtedly be used for years to come.

In this first test, we're able to see the true power of the A8's integrated graphics processor. It simply demolishes every other IGP and comes within spitting distance of certain lower end discrete cards.

Futuremark 3DMark Vantage

3DMark Vantage v1.0.1
Graphic Settings: Entry Preset
Resolution: 1024X768

Test: Specific CPU Score and Full Run 3Dmarks
Comparison: Generated Score

3DMark Vantage is the follow-up to the highly successful 3DMark06. It uses DirectX 10 exclusively so if you are running Windows XP, you can forget about this benchmark. Along with being a very capable graphics card testing application, it also has very heavily multi-threaded CPU tests, such Physics Simulation and Artificial Intelligence (AI), which makes it a good all-around gaming benchmark.

This IGP's architecture obviously likes Vantage as it is able to eke out a win at every memory setting. We can also see that Vantage seems to benefit quite a bit from the increased memory bandwidth of 1866Mhz.

Far Cry 2

Far Cry 2 1.02
Resolution: 1280x1024
Anti Aliasing: 0
Quality Settings: Medium
Global Settings: DX9 Enabled

Test 1: Ranch Long Demo
Comparison: FPS (Frames per Second)

Far Cry 2 is the hot new new first-person shooter from Ubisoft's Montreal studio, and the first game to utilize the new visually stunning Dunia Engine, which will undoubtedly be used by numerous future games. Using the included Benchmarking Tool, we ran the Long Ranch demo in DX9 mode at 1280x1024 with all settings set to low.

Left 4 Dead

Left 4 Dead (Latest Update)
Resolution: 1680x1050
Filtering: 0X AA / 0X AF
Graphic Settings: Medium
Shader Detail: Medium
Test 1: HWC Custom Timedemo
Comparison: FPS (Frames per Second)

Left 4 Dead is the latest disorienting, fast-paced zombie apocalypse mega-hit from Valve. L4D uses the latest version of the Source engine with enhancements such as multi-core processor support and physics-based animation. We tested at 1680x1050 with in-game details set to low. For benching, we used a pre-recorded 20 minute timedemo taken on the No Mercy campaign during The Apartments mission.

World in Conflict

World in Conflict v1.010
Resolution: 1680x1050
Anti-Aliasing: 0X
Anisotropic Filtering: 0X
Graphic Settings: Medium (DX10)
Test 1: Built-in Benchmark
Comparison: FPS (Frames per Second)

One of the detailed and most visually stunning real-time tactical games in recent history, World in Conflict remains a staple in our gaming lineup . For this test we used the in-game benchmarking tool.

X3: Terran Conflict

X3: Terran Conflict 1.2.0.0
Resolution: 1680x1050
Anti-Aliasing: 0X
Anisotropic Filtering: 0X
Graphic Settings: Medium
Test 1: Built-in Benchmark
Comparison: FPS (Frames per Second)

Real world gaming scenarios show basically the same results as the two 3DMark tests did: the A8 and its IGP are able to come close to or beat certain mainstream DX11 GPUs. The past generation of AMD IGPs and even Intel's own HD 3000 don't have a snowball's chance in hell of keeping up. The HD 6550D is actually the first non-discrete graphics processor that we'd consider playing games with.

In our opinion, AMD has simply shattered almost every preconception about what an integrated graphics chip can do. To make matters even better for the Fusion architecture, Intel is still mired in the DX10.1 generation without any hope of DX11 reinforcements until Ivy Bridge arrives sometime in 2012.

 

[MAC]

Power Consumption / Temperature Testing

Power Consumption

For this section, every energy saving feature was enabled in the respective BIOSes and the Windows Vista power plan was changed from High Performance to Balanced.

For our idle test, we let the system idle for 15 minutes and measured the peak wattage through our UPM EM100 power meter.

For our CPU load test, we ran Prime 95 In-place large FFTs on all available threads for 15 minutes, measuring the peak wattage via the UPM EM100 power meter.

For our overall system load test, we ran Prime 95 In-place large FFTs on all available threads for 15 minutes, while simultaneously loading the GPU with OCCT v3.1.0 GPU:OCCT stress test at 1680x1050@60Hz in full screen mode.

As you can see, our A8-3850 sample achieved power consumption numbers that were middle of the pack. In fact, that were eerily similar to its close cousin the Phenom II X4 840. We are slightly surprised that AMD were not able to achieve better results, especially since this is a chip that was fabricated using the new 32nm manufacturing process. Yes, this is a processor with a whopping 1.45 billion transistors, by far the most on this list, but the large graphics portion should be power-gated when a discrete GPU is installed, so that doesn’t explain the lack of improvement. Overall then, not the greatest CPU performance per watt from this APU.

Frankly though, this processor has a huge chunk of its die dedicated to graphics, and it was designed to be used in a simple system without a discrete GPU, so let’s take a look at the power consumption when using the IGP.

First and foremost, the Sandy Bridge chips in this graph have an almost unfair advantage in the form of the Intel DH67BL motherboard. This Intel-manufactured motherboard has unmatched idle power consumption, easily 10W less than comparable motherboards from the big three motherboard manufacturers. When you take that into consideration, the A8-3850’s idle numbers are very impressive. When you load up the CPU, power consumption rises greatly, and it’s not really competitive with its main competition, the Core i3-2100 series. Thankfully, the power consumption doesn’t rise much when you also tax the integrated GPU, so overall you end up with a system that should peak at about 150W max. That is pretty darn good efficiency for the level of gaming performance that you are getting compared to the discrete GPUs.

Temperature Testing

For the temperature testing, since we were not given a default cooler from AMD, we used a Thermalright Ultra-120 Extreme with a Thermalright TR-FDB-1600 fan. The ambient temperature was 24°C/75°F. The application used to monitor temperatures was HWiNFO v3.82-1300. Keep in mind that the thermal sensors in most modern processors are not really accurate at measuring idle temperatures, hence the very small delta between the room temp and the idle results.

Idle CPU + Idle IGP: The system was left to idle for 15 minutes.
Idle GPU + Load IGP: OCCT v3.1.0 GPU stress test was run at 1680x1050 for 15 minutes.
Load CPU + Idle IGP: Prime 95 In-place large FFTs was run for 15 minutes.
Load CPU + Load IGP: Prime 95 In-place large FFTs and OCCT v3.1.0 GPU stress test were run for 15 minutes.

AMD are known for making cool-running processors, and the A-series APUs appear to be no different.
Our A8-3850 sample was downright chilly when idle, and barely warmed up when we were fully loading the integrated GPU. Even when both CPU and GPU portions where loaded, the temperatures never even reached 45°C.

 

[MAC]

Overclocking Results

Overclocking Results

Since we've barely had a week with this chip, the amount of stability overclocking we've been able to do has been quite limited. However these are some very preliminary results which were achieved over the course of a few hours before launch.

Remember, this is a brand new platform for us and like it or not, it has quite a few intricacies. Tweaking the APU bus speed on this platform changes the core clock, the GPU clock, the memory frequency, the NCLK, and can also affect the PCI-E, SATA USB interfaces. So there is a lot to tweak and learn before we can give you any real pointers.

In the mean time though, these are the highest results that we achieved, using voltages that we would not necessarily recommend for 24/7 use:

Highest CPU Overclock - Click on images to enlarge - Highest Bus Overclock​

As you can see, nothing too dramatic so far, but this Lynx platform is definitely showing some promise. We achieved these results while using the IGP, and at these bus speeds the GPU clock should theoretically be running at between 918MHz and 936Mhz, a hefty increase over the stock 600Mhz. Apparently, it is much easier to hit high bus speeds when using a discrete GPU, so we will be checking that out as well. By the way, we have seen some insane memory frequencies with Llano (think above DDR3-2300), and it is not even particularly difficult if you have good memory modules.




EDIT:

As promised here are our stable overclocking results.

Highest Stable CPU Overclock

Click on image to enlarge

Highest Stable APU Bus Overclock

Click on image to enlarge

Highest Stable Memory Overclock

Click on image to enlarge​

 

[MAC]

Conclusion

Conclusion

And here comes to a close a review that seemed like it would never come. AMD has been touting Llano and Fusion in general for what seems like an eternity, and we are very happy to have a final product in our hands. Unlike the low wattage Brazos platform which never really piqued our enthusiast interest, the Lynx platform is definitely worthy of our attention.

The chip that we took a look at today, the $135 A8-3850, really stands out as a phenomenal bargain. It is not going to set any records in CPU benchmarks, but it is the new golden standard when it comes to IGPs. Take a look at these two tables to see what we mean:

The natural competitor for the A8-3850 is the dual-core/four-thread Core i3-2100 series, which headlines as Intel's mainstream-level Sandy Bridge offerings. As luck would have it, we just happened to have a Core i3-2120 on hand. This little Intel chip retails for $150, which is a little more expensive than AMD's top-end APU, but it nevertheless allows us to see what the strengths and weaknesses of each processor family is.

There's no reason to beat around the bush here, the "Stars" K10.5 cores simply cannot compete with the newer Sandy Bridge microarchitecture when it comes to lightly threaded applications. The Llano chip does occasionally score the odd victory thanks to its four native cores, but even those victories are often by a very slight amount. To make matters worse, in all of our gaming benchmarks it was a slaughter in favour of the Sandy Bridge processor, but only when using a discrete GPU.

When we benchmarked the integrated GPU's, Llano's raison d'être became crystal clear. The Radeon HD 6550D graphics solution that AMD have embedded in the A8-series leaves in the dust anything that Intel have produced so far in the IGP realm. The A8-3850 had on average 80% higher frame rates than the Core i3-2120. If we increased the Llano's memory frequency from DDR3-1333 to DDR3-1866, we could have easily realized a further 15% framerate improvement. This synergy between the GPU and CPU that has been personified in AMD's APUs is becomingly increasingly important as software manufacturers begin harnessing the parallel processing horsepower of modern graphics architectures. While there was once only a handful of GPU-accelerated programs available, now everyday programs like Microsoft Word, Internet Explorer and Flash can benefit from GPU acceleration.

Still not good enough for you? With the new Dual Graphics feature, AMD is allowing users to augment the integrated graphics processor with an entry-level discrete card for increased performance. Now currently this feature is hit or miss due to driver inconsistencies but when it worked, there was a noticeable framerate increase across the board. We wouldn't bank on Dual Graphics being an integral part of the Lynx platform but it could be an excellent way to upgrade in the future without spending too much money.

If you're building a new budget-friendly system from scratch that will be used for medium-level gaming, HD movie watching or a long list of other seemingly benign tasks, this new Lynx platform is a no-brainer. The A8-3850 is a terrific all-in-one chip that provides unmatched value when you consider that the integrated GPU is truly equal to a $50 graphics card. Intel's lower-end Sandy Bridge models are absolutely faster in CPU-centric benchmarks, but when you take an overall look it is hard not to come to the conclusion that Llano's graphics capabilities more than makes up for its processing shortcomings. In other words, whereas most users likely will not notice that the A8-3850 is slightly slower than another processor in 2D applications, they will notice that it has much better framerates, decoding abilities and GPU computing performance. To us, that equates to a better overall computing experience.

AMD A8-3850 APU Review: Llano Hits the Desktop Comment Thread​