AMD Kaveri A10-7850K & A8-7600 Review

[SKYMTL]

GPU Compute Performance (Civ V / MediaEspresso)

GPU Compute Performance (Civ V / MediaEspresso)

With many CPUs using integrated graphics processors, parallel co-processing has become a hot topic of conversation. More and more applications now include support for OpenCL, DirectCompute and other features which help speed up performance in order to quickly finish certain tasks.

In this section, we will be benchmarking a number of applications which support (or claim to support) GPU compute in an effort to highlight the performance benefits which come with this technology. All of these tests are conducted on a system WITHOUT a discrete GPU installed.

Civilization V: Gods & Kings Leader Benchmark

Once again we are using Civ V in an odd way but its included Leader Benchmark includes a feature which uses the compute shader features on the GPU to rapidly compress large texture files. This puts a large amount of stress upon the graphics subsystem and should illustrate how well various manufacturers have implemented GPGPU features onto their cores.

MediaEspresso

Much like MediaCoder, MediaEspresso is a program used for video transcoding from one source to another. In this case, we take a high resolution 600MB 1080P video and convert it to a different MPEG-4 format suitable for mobile devices.

RESULTS:

 

[SKYMTL]

GPU Compute Performance (MuseMage / Photoshop CS6)

GPU Compute Performance (MuseMage / Photoshop CS6)

MuseMage

MuseMage is a fully featured image editing program that uses GPU acceleration for its enhanced suite of color effects, filters, adjustment tools and other features. In order to achieve our results, we used MuseMage’s handy Batch script to apply 15 separate and consecutive image modifications to a 50MB JPG file. The results below represent the amount of time it took to complete this task.

Photoshop CS6

In our previous Photoshop CS6 benchmark, we deliberately disabled the GPU acceleration features so we could get an apples to apples CPU comparison. However, in this test, we are enabling that acceleration to see what affect it has upon the benchmark numbers. Please remember: Photoshop’s GPU features only improve performance on SOME (rather than all) editing tools so performance will not scale in a linear fashion due to the CPU still playing a role.

 

[SKYMTL]

IGP Gaming Benchmarks (3DMark06 / 3DMark11)

IGP Gaming Benchmarks (3DMark06 / 3DMark11)

Integrated graphics processors have never been thought of as legitimate gaming devices but that stigma has begun to change. In this section, we test 3DMark06 and 3DMark11 to see how modern IGP systems reach to DX9 and DX11 environments.

RESULTS:

 

[SKYMTL]

IGP Gaming Benchmarks (Deus Ex: HR / Dirt 3 / Skyrim)

IGP Gaming Results (Deus Ex: HR / Dirt 3 / Skyrim)

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:

 

[SKYMTL]

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:

 

[SKYMTL]

IGP Gaming Benchmarks (BF4 / CoD Ghosts / Dirt Showdown)

IGP Gaming Benchmarks (BF4 / CoD Ghosts / Dirt Showdown)

 

[SKYMTL]

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.

Efficiency was one of AMD’s major focuses this time around and it looks like they’ve accomplished their goals….wit the A8 7600 at least. Their lowest-tier Kaveri APU is ridiculously efficient, particularly when it’s used in 45W mode.

The A10 7850K provides a disappointing counterpoint by consuming nearly as much as the 32nm A10 6800K. That’s certainly not an optimal result considering these two APUs’ neck and neck benchmark performance. 28nm was supposed to bring about better power consumption numbers but that didn’t happen with the A10 7850K.

 

[SKYMTL]

Conclusions

A8 7600 Conclusion

When I first opened the box, I couldn’t believe my eyes. “Who in their right mind samples a low-end APU processor to highlight their new lineup?” I thought. After years of bleeding edge Intel and AMD samples being sent in an effort to showcase the best of the unaffordable best, getting an A8-7600 just seemed odd.

I quietly progressed through the benchmarks and now, after more than 20 hours of testing, I get it. This particular APU isn’t about record setting overclocking, extreme gaming or pushing the boundaries of x86 performance. It’s a stark reminder of reality. While part of me misses the race towards the top, that’s not what folks are looking for these days. AMD’s goal this time around was to showcase how ridiculously capable their more affordable APUs are.

The A8-7600 is the real standout of AMD’s new lineup. It epitomizes everything we could want from Kaveri by incorporating the graphics and multimedia performance of a 95W A10-6800K into a part with a TDP of just 65W. More importantly, it can be easily configured down to 45W without taking a massive cut in the areas that count the most to its intended market: GPU accelerated decoding, transcoding and gaming. I’m not saying forget about the K-series parts but the A8 7600 makes one hell of a convincing case for itself in HTPC systems and simple set-top boxes. Intel doesn’t have a thing that can touch it within their current Haswell product stack.

By combining graphics horsepower that’s focused in key areas and “good enough” (for its segment at least) CPU performance into an efficient, affordable package, AMD has created the perfect poster child for their design mantra and a great companion for SFF systems. It may not have the overclocking chops of its more expensive siblings but the A8 7600 is infinitely more capable of displaying the best Kaveri has to offer.

A10 7850K Conclusion

AMD seems to have made it a point to keep the A10 7850K well away from most press outlets; they figured the A8 7600 story personified Kaveri’s strengths in today’s market. Sure, that makes sense but readers of these pages want a better perspective about what’s available and simply ignoring this APU would be doing them a disservice. Plus, the A8 will only be available a month or more after the other parts are launched. So inquiries went out to unofficial sources and our sample was literally airlifted in from overseas this weekend. With the APU in hand, I now understand why the A10-7850K was almost swept under the rug.

With its Steamroller x86 cores and 512 core GCN-based graphics capabilities, the A10 7850K is a potent addition to AMD’s lineup from a gaming and GPU compute perspective. It absolutely screamed through games, routinely outpacing the A10 6800K by a wide margin and its performance in the few available GPU accelerated applications was nothing short of amazing. These are two key elements which go a long way towards legitimizing AMD’s focus on graphics horsepower. It gives me hope for the future. With the sample arriving less than 24 hours ago there wasn’t time for overclocking but I expect some great results there as well.

On the other hand, the majority of benchmark results were not up to par by current standards. The 7850K’s CPU scores point towards a worrying trend: a real lack of inter-generational x86 performance increases. There were some IPC improvements that actually made things close but on average, it didn't deliver one iota more performance than its A10 6800K predecessor or provide anything in the way of efficiency increases. The K-series part didn't even distinguish itself all that well from the 65W, $55 less expensive A8 7600. This isn't just a poor showing; it's devastating for an APU that costs 20% more than the product it’s supposed to replace.

Despite AMD’s best efforts we’re still years away from the day when broad-scale availability of GPU accelerated applications. This means raw CPU horsepower is still a critical element in this equation. Yes, things will change and more accelerated programs will become available but we've been saying that for three years now. Until more developers jump on board, for all its innovative features and big ideas, Kaveri will remain relegated to entry level markets unless its low CPU frequencies are addressed and HSA-supporting software becomes available. And quickly.

Kaveri isn’t lacking potential and competes well on a price / performance level with some Intel offerings but I constantly found myself wondering how great the A10 7850K could have been if AMD had at least managed to match the previous generation’s clock speeds. With OpenCL 2.0 still in its infancy, HSA features inaccessible in the current driver stack, legacy instruction sets still prevalent in today's software and the glacial speed of GPU accelerated application development, there’s very little reason to spend the extra money for an “unlocked” 95W APU. Especially when a significantly less expensive 65W model can offer just as much, if not more.