Skip to main content

AMD A8-3850 Review: Llano Rocks Entry-Level Desktops

Power Consumption

The A8-3850 APU hits a 100 W TDP. In comparison, the Phenom II X4 965 Black Edition is a 125 W part, and the Core i3-2105 is a 65 W component.

Of course, the A8-3850 and Core i3-2105 include graphics. The 890GX chipset adds up to 25 W to the Phenom platform’s maximum power use.

It’s no surprise, then, that the Phenom II X4-based machine uses the most power across a run of PCMark 7 and the common usage performance tests. We’re also not shocked to see the 100 W A8-3850 fall into second place in this lower-is-better measurement.

More interesting is the fact that AMD’s Llano-based APU finishes the test before the Core i3 (despite the fact that Futuremark awards Intel a higher score), and only averages 4 W higher average power use. See? Those TDP numbers can be a little tricky when vendors don’t use the same methodology for rating their parts.

If you chart out the entire run, you get this gem of a line graph. Hard though it might be to decipher, I’ll help out as much as possible.

Most obvious is that the Phenom II X4 and its Radeon HD 4290 take forever to wrap this test up, and the only reason I can come up with is the weak graphics core that stumbles over the 3D components of the benchmark.

Beyond that, it’s clear that the Dual Graphics-equipped setup uses the most power. And while it’d seem AMD’s Phenom II would be the second-most severe offender, it’s actually the Intel Core i3 configuration armed with discrete graphics with the second-highest average.

What’s nice to see is how low the A8-3850 with Radeon HD 6550D graphics and Core i3-2105 with HD Graphics 3000 dip (41-42 W total platform power at idle). 

  • whatisupthere
    Great review! Thanks Toms
    Reply
  • Tamz_msc
    Another win for AMD!
    Reply
  • SteelCity1981
    So then what's the point of getting the Turbo Core versions when they are going to be Turbo Clocked slower then the none Turbo Clocked versions...
    Reply
  • cangelini
    SteelCity1981So then what's the point of getting the Turbo Core versions when they are going to be Turbo Clocked slower then the none Turbo Clocked versions...
    They don't want you to see better performance from a cheaper APU in single-threaded apps by pushing Turbo Core further ;-)
    Reply
  • Known2Bone
    i really wanted see some amazing gains in the content creation department what with all that gpu power on chip... oh well games are fun too!
    Reply
  • ivan_chess
    I think this would be good for a young kid's PC. It would be enough to run educational software and a web browser. When he grows up to be a gamer it would be time to replace the whole machine anyway.
    Reply
  • DjEaZy
    ... it's may be not the greatest APU for desktop... but it will be a powerful thingy in a laptop... the review was nice... but in the gaming department... would be nice to see a standard 15,x'' laptop resolution tests @ 1366x768... or something like that...
    Reply
  • Mathos
    Actually if you want good DDR3 1600 with aggressive timings, the Ripjaws X series memory that I have does DDR3 1600 at 7-8-7-24 at 1.5v, not all that expensive when it comes down to it either.
    Reply
  • Stardude82
    This makes little sense. An Athlon II X3 445 ($75) and a HD 5570 ($60, on a good day you can get a 5670 for the same price) would provide better performance for the same price ($135) and not have to worry about the RAM you use.

    So is AM3+ going to be retired in favor of FM1 in the near future? Why are there chipset at all? Why isn't everything SOC by now?

    Otherwise this is a very good CPU. If AMD has used 1 MB level 2 caches in their quads when they came out with the Deneb Propus die, they would be much more competitive.
    Reply
  • crisan_tiberiu
    stardude82This makes little sense. An Athlon II X3 445 ($75) and a HD 5570 ($60, on a good day you can get a 5670 for the same price) would provide better performance for the same price ($135) and not have to worry about the RAM you use. what about power consumption?
    Reply