AMD A10-6800K and A4-4000 APU Review

By on January 4, 2013, 10:55 PM

Late last year we checked out AMD’s desktop version of Trinity, comprising Piledriver CPU cores along with an on-die Radeon HD 7000 Series (not Graphics Core Next) graphics processor. The A10-5800K debuted at just $130, pitting it at the time against the Core i3-3220. The end result was a typical Intel vs. AMD battle. While AMD had a clear cut advantage in GPU performance, the CPU side of things was more closely contested.

Overall, Trinity picked up where Llano left off, providing an affordable package with enough processing speed for most users while supplying sufficient graphics muscle for most of today's PC games on modest settings.

Now, little over 6 months down the road, AMD is updating their product lineup with a minor refresh. Codenamed Richland, these new APUs offer no substantial changes to either the CPU or GPU. Other than some clock speed improvements, better power management, and a few new software features, everything is the same.

Leading the pack today is the A10-6800K, which as you might have guessed, is stepping in to replace the A10-5800K. Both processors are identical with a few minor exceptions. The A10-6800K is clocked slightly higher and has an updated Radeon HD 8670D GPU. We're also taking a look at the bargain basement A4-4000 at the other end of the spectrum, which we suspect is going to be a slightly slower version of last year’s A4-5300.

Read the complete review.




User Comments: 36

Got something to say? Post a comment
1 person liked this | St1ckM4n St1ckM4n said:

Lol AMD. What r u doing. Stahp.

Guest said:

What memory speed did you use?

ET3D, TechSpot Paladin, said:

Nice advertising on the Anandtech news item. Smooth move.

Richland looks more boring than desktop Haswell, if that's even possible.

gamoniac said:

At 65W TDP, A10-6700 (with the same HD8670D graphic, 3.7Ghz/4.3Ghz turbo, and same price), is a much more interesting APU than A10-6800K. It would have been interesting to see how its power consumption stacks up against comparable Intel products.

gamoniac said:

Nice advertising on the Anandtech news item. Smooth move. ...

OK, I see that. At least Julio Franco was being... frank that he is from TechSpot. That's both respectful and respectable IMO.

mosu said:

Better read this: [link] Where can I donate the money for TechSpot to buy an Intel 3220 chip?

Guest said:

Yes, there is French Hardware review site that noted how the A10-6700 was very energy efficient while outperforming in graphics and games.

This this also the umpteenth review that has a distinct lack of OpenCL / BasemarkCL multiprocessing benchmarks. A number of popular apps like Adobe, Aviary and many other utilizing OpenCL acceleration.

Doesn't the reviewer want his audience to know the full breadth of the capabilities or perhaps it shows an indication of not wanting to embarrass Intel?

ET3D, TechSpot Paladin, said:

At 65W TDP, A10-6700 (with the same HD8670D graphic, 3.7Ghz/4.3Ghz turbo, and same price), is a much more interesting APU than A10-6800K. It would have been interesting to see how its power consumption stacks up against comparable Intel products.

I'm pretty sure that the A10-6800K can be easily underclocked to be as efficient as the A10-6700. I think it's a problem that reviews don't underclock and undervolt CPU's. AMD CPU's often work well while undervolted, and it would be interesting to see what kind of power consumption is possible to get with some power tweaks. I frankly find underclocking/undervolting more interesting than overclocking and would love if it was given more attention, especially in the case of processors that are likely to end up in HTPC's.

Staff
Steve Steve said:

What memory speed did you use?

All Systems were tested with DDR3-1866 memory if they supported it. The A4-4000 only supports 1333MHz memory and that's all we could get it to boot with.

At 65W TDP, A10-6700 (with the same HD8670D graphic, 3.7Ghz/4.3Ghz turbo, and same price), is a much more interesting APU than A10-6800K. It would have been interesting to see how its power consumption stacks up against comparable Intel products.

Expect bugger all difference really. I mean just look at the A10-5700 and A10-5800K in our power consumption graph :S

Yes, there is French Hardware review site that noted how the A10-6700 was very energy efficient while outperforming in graphics and games.

It might outperform in games but it is not very energy efficient.

This this also the umpteenth review that has a distinct lack of OpenCL / BasemarkCL multiprocessing benchmarks. A number of popular apps like Adobe, Aviary and many other utilizing OpenCL acceleration.

Doesn't the reviewer want his audience to know the full breadth of the capabilities or perhaps it shows an indication of not wanting to embarrass Intel?

I'm not sure who makes Photoshop CS6 then :S I'm also pretty sure I couldn't embarrass Intel.

Staff
Julio Franco Julio Franco, TechSpot Editor, said:

@gamoniac & guest, we didn't test the A10-6700 deliberately but because AMD didn't sample us with one ahead of release, instead they opted to send us the AMD A10-6800K and A4-4000 which we have tested and showcased. If you believe the other is the more interesting processor that's perfectly fine, but we have reviewed actual products (including the series flagship) and thus are giving our opinion on the products we have, not the ones we don't.

As for benchmarks, again, it's easy to pick which test we didn't run (out of the 15 or so tests we did), yet you mention OpenCL and Adobe, when we have tested Photoshop CS6 among other multithread benchmarks.

GhostRyder GhostRyder said:

Well I was wondering when I would get to see a review on this processor. Its an interesting APU for sure, its getting up in performance and they are now becoming much more viable for actual gaming performance. But they are still lacking in straight CPU performance and I was expecting/hoping to see much higher gains all around like what was being advertised.

Guest said:

I see no point in even releasing this.

EEatGDL said:

[Sarcasm off during my whole comment] Can somebody explain me what is the actual target market of the A10-6800K? I mean, because of its specs it seems like a mid-high processor; so is it in the same price tag of the IB i5 there? I'm from Mexico and those Richland CPUs still don't have a price to be compared to Intel's same priced here.

I've read the past APU reviews when they were released, but I still don't have it clear what is the purpose of the APU. Are they competing with their own Bulldozers? Why develop both brands in parallel? Are they more graphics-focused than Bulldozer or in a more accessible price tag?

This is no critic, neither opinion. I'm just full of doubts about this AMD alternative; and they haven't been clarified for a while now. I originally thought they were targeting CPUs <= Core i3, and seeing comparisons with i5 and even i7 [graphics] just got me lost.

GhostRyder GhostRyder said:

The point of an APU is pretty much to have a CPU and GPU on the same die that performs on par with having a separate GPU while keeping it clean and simple.

When Comparing the APU like the 6800k to the I5, the Graphics capabilities are higher on the APU by AMD, but the power of the CPU is behind the I5. In short, they are mid range Processors that are deigned for light-medium gamers who want value and a bit of performance. They are not exactly meant to be top of the line, they are a mid-range solution with a price tag cheaper than buying a separate processor and GPU gaming rig, but with similar results.

Guest said:

Better wait for Kaveri APUs that will feature a completely new CPU and GPU architecture. The Richland APUs were merely a minor update to the Trinity platform, offering nothing more than small performance improvements.

Guest said:

Kaveri is supposed to bring CPU performance improvements of ~30%. The GPU is supposed to be a LOT more powerful, too. It's going to be a 28nm processor vs. Richland 32nm.

1 person liked this | JC713 JC713 said:

They lost to a Pentium. Good job...

Guest said:

PCMark 8 is already being released and AMD showed an Acer V5 laptop using Windows 8.1, so how hard would it be to update with something that shows CPU plus GPU compute performance?

We know from TechReport's review of the 1.5Ghz Kabini outperforming faster clocked Ivy Bridge CPUs in OpenCL acceleration, so I guess TechSpot is under the Intel-biased benchmarks mostly influence.

GhostRyder GhostRyder said:

I don't think they are picking benchmarks and stuff that are directly intel biased or anything. Though a few more might be nice of the Opencl persuasion would be nice.

1 person liked this | ET3D, TechSpot Paladin, said:

EEatGDL, these AMD APU's are aimed at PC's where due to price or space constraints adding a discrete graphics card is problematic, but GPU performance still matters. They're supposed to provide a good sweet spot between price, CPU performance and GPU performance. Whether they succeed depends on what's important to you and of course if your local price of AMD CPU's is higher, they'd naturally be less appealing.

1 person liked this | mosu said:

The main point being this, boys and...boys: When it comes to pricing the AMD A10-6700 and AMD A10-6800K both run $142. That is a very reasonable price for what you get. The AMD FM2 platform is also reasonably priced as you can pick up a quality brand name AMD A85X chipset powered ATX motherboard like the Gigabyte GA-F2A85X-D3H for $76.49 shipped. The last key ingredient is the 8GB (2x4GB) DDR3 2133MHz 1.5V memory kit and that runs as low as $68.99 shipped. So, to build a system like this you are looking at $287.48 shipped without messing with any rebates. The Intel Core i7-4770K processor runs $349.99 shipped before motherboard and memory costs. AMD continues to excel at price versus performance! Here you have a processor that can run games faster than Intel and it costs far less.

Guest said:

After 3 years on my old Asus gaming laptop I'm looking to upgrade to a 6800k system (I've decided I want a "budget" box for light gaming and day-to-day work, and then I'll get myself a new ultrabook for when I'm travelling around)

Is a 6800k + 2133mhz ram + ssd going to be quick enough for these games at 1080p with reasonably high settings?

Fallout 3/New Vegas

Skyrim

Arkham City & Arkham Asylum

Mass Effect 1,2,3

I'm not exactly a power user. The only upcoming game I've got any interest in is Rome 2 Total War. I'll also use the machine for a small bit of video editing (just vids I upload to YouTube) and university work.

DKRON said:

If they combined the AMD's Radeon HD 8670D with Intel's I5 we could actually see some steps towards good framerates (at low res)

GhostRyder GhostRyder said:

@the guest above

Yes and no.

Fallout, it will do a decent job

Skyrim: This game is heavily CPU dependant, it will be decent, but high might be pushing it

Arkham series: Not to bad, you could definitely get away with low-medium

Mass Effect: Yea should be enough for medium at least

Guest said:

I could not find the listing of the speed of the 8GB DDR3 ram. But this is a huge factor in the performance of the A10 chips.

GhostRyder GhostRyder said:

The ram used In the above benchmarks I believe was 1866.

1 person liked this | ET3D, TechSpot Paladin, said:

Mosu, comparing to a Core i7-4770K is silly and completely undermines your argument. You should stick to showing how the AMD APU's beat similarly priced Intel configurations in performance, or beat in price similarly performing configurations.

By the way for anyone who wants a good power comparison Tom's Hardware has one here: [link]

GhostRyder GhostRyder said:

Mosu, comparing to a Core i7-4770K is silly and completely undermines your argument. You should stick to showing how the AMD APU's beat similarly priced Intel configurations in performance, or beat in price similarly performing configurations.

By the way for anyone who wants a good power comparison Tom's Hardware has one here: [link]

I agree, the point of an APU is not to be the most powerful brute force processor and compete with the I7, it is a solution giving great gaming/media performance in a small package. The CPU does its job with the on board graphics and gives you a complete package to play games decently on a budget without having to worry about a separate GPU in the mix. The I7 is meant for high end gaming and a separate GPU, the on board is there for low power when not gaming.

Guest said:

First, let me say thank you for the review. However I would've liked to have seen the A10-6800k paired with some 2133mhz RAM for these benchmarks. APUs are heavily dependant on memory speed, and in getting the 6800k vs the 6700 you're paying for that built in 2133mhz memory controller and higher clock speeds. Also, the 6800k overclocks much better than it's Trinity counterpart - another review I read saw the 6800k get up to 5.0ghz easily and this gave it a sweet performance boost. Richland was a mild refresh to the Trinity lineup, nothing more. AMD is still planning on releasing Kaveri in 4Q13 - expect to see great things out of that.

I also agree with what someone said about using OpenCL benchmarks, as the APUs tend to shine in those. In your Adobe benchmark did you have OpenCL enabled?

Finally, for anyone talking about power consumption - First of all, it's a desktop. No one cares about battery life on a desktop. Secondly, Intel's TDP is up quite a bit on their new Haswell architecture, and doesn't do so well in the wattage arena either.

ET3D, TechSpot Paladin, said:

I agree that OpenCL benchmarks are interesting, though I'd rather stick to practical ones than theoretical ones. The interesting thing in the results I've seen is that Intel crushes the AMD APU's in some OpenCL benchmarks. (In others AMD beats Intel.)

"No one cares about battery life on a desktop." That's your spin, but people do care about power consumption on desktop, and especially in an HTPC.

Staff
Steve Steve said:

I agree that OpenCL benchmarks are interesting, though I'd rather stick to practical ones than theoretical ones. The interesting thing in the results I've seen is that Intel crushes the AMD APU's in some OpenCL benchmarks. (In others AMD beats Intel.)

"No one cares about battery life on a desktop." That's your spin, but people do care about power consumption on desktop, and especially in an HTPC.

Sorry I am a bit late to the party but for all those complaining about the lack of OpenCL testing we tested with Photoshop CS6 which by default uses the graphics processor to boost performance. Therefore we did test Photoshop CS6 using OpenCL.

ET3D, TechSpot Paladin, said:

Sorry I am a bit late to the party but for all those complaining about the lack of OpenCL testing we tested with Photoshop CS6 which by default uses the graphics processor to boost performance. Therefore we did test Photoshop CS6 using OpenCL.

The fact that Photoshop CS6 can take advantage of OpenCL doesn't mean that the workload used in the test does. It stands to reason that if OpenCL was used then the FX 4350 with its Radeon 7970 would have had much better results than the A10-6800K. That's not guaranteed, of course, but the benchmark looks consistent with the relative CPU power, and I would have expected results using OpenCL to be somewhat different.

Staff
Steve Steve said:

Actually all platforms used the Radeon HD 7970GHz in that test. The only tests we dumped the 7970 for were the IGP tests. What workload do you suggest we use? Have you got something hand picked?

I see the need to test Photoshop CS6 without the dedicated GPU and that is something we will add in future reviews with these kinds of CPUs.

PCMark 8 does have a pretty detailed Photoshop CS6 workload so we will probably use that in future.

JC713 JC713 said:

The fact that Photoshop CS6 can take advantage of OpenCL doesn't mean that the workload used in the test does. It stands to reason that if OpenCL was used then the FX 4350 with its Radeon 7970 would have had much better results than the A10-6800K. That's not guaranteed, of course, but the benchmark looks consistent with the relative CPU power, and I would have expected results using OpenCL to be somewhat different.

AMD just signed a deal with Adobe to further integrate OpenCL. It probably wont be another few months before it is fully integrated.

Guest said:

Please stop using BF3 in single player as a CPU benchmark, it is totally irrelevant. In multiplayer you can see a big difference between CPUs, especially on 48/64 player servers.

Guest said:

I think the A10 Trinity improves drastically at 1866 and then the curve is gradual to 2133. From here... [link]

Load all comments...

Add New Comment

TechSpot Members
Login or sign up for free,
it takes about 30 seconds.
You may also...
Get complete access to the TechSpot community. Join thousands of technology enthusiasts that contribute and share knowledge in our forum. Get a private inbox, upload your own photo gallery and more.