AMD's Trinity APU Efficiency: Undervolted And Overclocked

We've been playing with AMD's Trinity APUs for four months, and they're just now being rolled out to the channel. This time, we take a look at the architecture's efficiency compared to a pair of Ivy Bridge-based Core i3s. Can A10 and A8 stand up to Intel?

By now, you’ve read all about AMD’s unconventional approach to introducing its Trinity APU architecture, peeling back the embargo on gaming-oriented performance first, and then granting permission to talk about pricing, overclocking, and alacrity in x86-based apps via benchmarks a few days later.

I warned the company that splitting Trinity’s debut into two days made it look like AMD wasn’t particularly proud of its showing in productivity and content creation apps—and it had no reason to do this, in my mind. As far back as June, we had already shown that a Piledriver module was performing about 15% better than Bulldozer in single- and multi-threaded apps at the same clock rate. We knew that Trinity would be faster than Llano in some situations, but slower in others.

AMD went ahead with its plan. But because we got our hands on Trinity-based A10, A8, and A6 processors right after they were announced at Computex, the performance data that AMD wanted to keep under wraps until now was already available in AMD Trinity On The Desktop: A10, A8, And A6 Get Benchmarked! and AMD Desktop Trinity Update: Now With Core i3 And A8-3870K. If you want to know how Piledriver does at the same clock as Bulldozer, how memory bandwidth scaling affects graphics performance, how effective Dual Graphics is, and how power changes from one generation to the next, all of that information is available between those two links, and was up in our top carousel last week.

But back in June, I was missing performance data from Intel’s Ivy Bridge-based Core i3s. They weren’t available yet. So, as I was trying to come up with an interesting angle for today’s story, I knew we needed results from the Core i3-3220 and -3225, at least.


Radeon HD
GPU (MHz)
Shaders
TDP
Cores
Base CPU
Turbo Core
L2 Cache
Price
A10-5800K
7660D
800
384
100 W
4
3.8 GHz
4.2 GHz4 MB
$122
A10-5700
7660D
760
384
65 W
4
3.4 GHz4.0 GHz4 MB
$122
A8-5600K
7560D
760
256
100 W
4
3.6 GHz3.9 GHz4 MB
$101
A8-5500
7560D
760
256
65 W
4
3.2 GHz3.7 GHz4 MB
$101
A6-5400K
7540D
760
192
65 W
2
3.6 GHz3.8 GHz1 MB
$67
A4-5300
7480D
724
128
65 W
2
3.4 GHz
3.6 GHz
1 MB
$53


Also, it was apparent that a lot of work went into improving Trinity’s idle power consumption compared to Llano. But the APU’s 100 W TDP is still significantly higher than Core i3’s 55 W ceiling. Almost certainly, Trinity can’t be made to outperform and under-consume Ivy Bridge. We're nevertheless testing our A10-5800K at lower voltages and higher clock rates to see if it can be made to shine more brightly against the backdrop of Intel’s entry-level chip, which is far less flexible.

As a result, we get to add overclocked and undervolted performance to our existing library of Trinity-based data, along with a comparison to Intel’s Core i3 competition. And with pricing information in our hands, it becomes pretty easy to set AMD and Intel up head-to-head and declare one company’s solution better than the other’s. Are you ready for the big judgement?

This thread is closed for comments
138 comments
    Your comment
  • esrever
    Most PC are idle or semi idle when people have them on. 90% of the time I use my PC, I do web surfing or watch video or a text editor for work, my pc is not loaded with benchmarks 24/7. If you look at idle power consumption, the trinity APUs are amazing. They easily beat out intels offerings. If you are looking at the power consumption over a month, the trinity will be much more energy efficient than the i3 for most people.
  • tacoslave
    man getting this in a 17inch laptop with a 12 cell battery would make it an instabuy
  • mayankleoboy1
    Quote:
    In the end, then, both Intel and AMD are offering you an experience. Which one do you pick?


    At this price point, i would choose AMD Trinity.
  • DjEaZy
    ... i like the WinZip with OpenCL acceleration benchmark... it shows...
  • cangelini
    esreverMost PC are idle or semi idle when people have them on. 90% of the time I use my PC, I do web surfing or watch video or a text editor for work, my pc is not loaded with benchmarks 24/7. If you look at idle power consumption, the trinity APUs are amazing. They easily beat out intels offerings. If you are looking at the power consumption over a month, the trinity will be much more energy efficient than the i3 for most people.

    Happy to set a couple of systems up and let you know what I find.
  • mayankleoboy1
    AMD should team up with developer of 7zip to accelerate it on APU's. That will make Trinity look better. A lot of people use 7zip. And most of the installation setup exe files are compressed using LZMA algorithm.
  • mayankleoboy1
    Chris, it would be great to see some benchmarks of applications that uses the new FMA3 instructions of the Piledriver.
  • cangelini
    esreverMost PC are idle or semi idle when people have them on. 90% of the time I use my PC, I do web surfing or watch video or a text editor for work, my pc is not loaded with benchmarks 24/7. If you look at idle power consumption, the trinity APUs are amazing. They easily beat out intels offerings. If you are looking at the power consumption over a month, the trinity will be much more energy efficient than the i3 for most people.

    So, it's probable that we're seeing a difference in configuration. It looks like Anand is using the Gigabyte A85X board and perhaps an older driver version. I'm on the MSI board and Cat 12.8, with a different Intel setup as well. On the Windows desktop, after 10 minutes on each config, I get 59 W for Intel and 67 W for AMD at idle.
  • techcurious
    Chris, for the sake of completeness, any chance you could undervolt the i3-3225 at stock speeds and run the power consumption/efficiency tests on it? ;) ...to reveal how low the i3 can be pushed with some tweaking as well, and create the opportunity for a more fair comparison with the undervolted Trinity results.
  • sarinaide
    Thanks Chris, another great article to pass time over. You really need to comment on the forums more and more so to help out against the blatent belligerence against what AMD are trying to achieve and how they are looking to achieve it.

    Hopefully this articale can start to filter around particularly for the budget users which A-series is premised to target.
  • So now that we got that out of the way...............where is the hybrid xfire chart so we know what's the max discrete card that will be supported? And while you're at it, when you find that out can you check to see if there are any significant gains when setting up a discrete + discrete + 7660 triple hybrid xfire set up, or even a quad hybrid xfire set up (3 discrete cards + 7660) of if either of those are even possible? After seeing that write up on how the dual 7750's performed, I'd love to see what trinity's version of hybrid xfire can pull off.
  • m32
    I could get rid of my family computer with a dedicated gpu and just slap an A10k in there. Most of the time it is just used for web browsing and such, so it would be an killer for my family and friends that don't need an lot.
  • The_Trutherizer
    Nice article, but I must say that much as I enjoy the over clocking stats from AMD; To be fair to Intel their part should also be over-clocked to make this a sporting comparison. I believe the two would be more or less equal, except for AMD's APU being considerably more capable at handling gfx tasks. And yes... As some people have stated we really need benchmarks where a discrete gfx card is used in conjunction with the APUs and HD CPUs as I believe this is what most people will do currently.
  • bulldozer83
    The_TrutherizerNice article, but I must say that much as I enjoy the over clocking stats from AMD; To be fair to Intel their part should also be over-clocked to make this a sporting comparison. I believe the two would be more or less equal, except for AMD's APU being considerably more capable at handling gfx tasks. And yes... As some people have stated we really need benchmarks where a discrete gfx card is used in conjunction with the APUs and HD CPUs as I believe this is what most people will do currently.


    overclock the locked Intel chips? how do you suppose they do that? they weren't testing against Intel K series unlocked chips.
  • Nintendo Maniac 64
    Wouldn't it be more fair to compare the i3 power consumption to the 65w Trinity APUs (such as the A10-5700) rather than the 100w ones?
  • americanbrian
    Umm, WHY DIDN'T YOU SHOW THE GAME BENCHMARKS WITH THE OVERCLOCKED GPU SETTINGS!!!

    I can't be the only one who was waiting for the money shot of what is the difference in performance when you clock up from 800Mhz to >1000Mhz.

    SUCH AN OVERSIGHT. UNFORGIVABLE!
  • americanbrian
    I mean really, why not show people what they want to know? I WANT TO KNOW.
  • theconsolegamer
    Where's the gaming benches of the retail APU?
  • chesteracorgi
    Given the results of head to head comparison in gaming, I'm interested in seeing them compete in transcoding, and comparisons when paired with discrete GPUs. Presently AMD Trinity seems to be the runaway winner for laptops, but a poor option for desktops.
  • abitoms
    Chris, and team, a few things I - and probably others- would like to see here;

    1. overclocked/undervolted benchmarks for the i3 parts
    2. dedicated gpu game benchmarks at 1440, 1680, 1920 for the A10 and the A8
    3. More OpenCl benchmarks with and without dedicated GPUs for the i3 parts as well as the A10 parts

    p.s. I realised I was getting thumbed up and down for this. do these seem like too many requests? nobody has covered trinity like toms and that too with superb writing quality. is it wrong for me to get greedy to read more of their stuff? :-) i'm addicted to this stuff is all. now if you'd excuse me, I have an F5 button to press.
  • ojas
    ChesteracorgiGiven the results of head to head comparison in gaming, I'm interested in seeing them compete in transcoding, and comparisons when paired with discrete GPUs. Presently AMD Trinity seems to be the runaway winner for laptops, but a poor option for desktops.

    Well, unless you're going to look at battery life, i think.
  • Could you simulate the exact clocks of the 65 W A10 somehow and see how much power it really takes?
  • dscudella
    I would definitely pick up a Trinity A10 laptop. If they could get them down to around $500 where the i3 laptops sit, they would sell like crazy
  • ojas
    My conclusion is this:
    For regular folk (internet, word, excel, movies, some transcoding, etc without an OC), a core i3 still wins. Why? No Quick Sync on the Pentiums, afaik, otherwise i'd agree with Chris on the G2120.

    For gaming with a discrete GPU: Core i3
    For gaming with an IGP: Trinity

    For cheap multitasking without an OC: Core i3

    For exploiting as much work done as possible and exploiting all available resources with or without an OC: Trinity.

    The thing is, as evidenced by your benchmarks, if you throw an average non-gaming workload at an A10 and an IVB i3, the i3 still completes tasks within the same time or in slightly more time than an OC A10 (except winzip with OpenCL, but that's because Intel/Nvidia are locked out).

    So unless one's using a lot of OpenCL applications, or very heavily threaded ones, it doesn't make much sense to go the Trinity route. I think with OCing on the A10, the power consumption should off-set the price difference in some time (don't live in the US so don't know how much time). Without an OC they were pretty much the same in most cases.

    As far as power consumption is concerned, even with a Intel Q8400 and a GTX560 + 2 sticks of DDR3 and 2 drives along with 4 fans (+3 if you include the CPU and GPU coolers) on an ATX mobo, i get 64W active idle. So you'll have to try and select as similar platforms as possible because there will be considerable variances.

    Or, you could use CoreTemp, i saw the new version reporting the power consumption of SNB chips. Don't know if it'll work with Trinity chips though.