Using Power More Efficiently: Dynamic Tuning 2.0

A common thread in modern microprocessor design is being able to use the power budget available. There have been many articles devoted to how to define power budgets, thermal budgets, and what the mysterious ‘TDP’ (thermal design power) actually means in relation to power consumption. Intel broadly uses TDP and power consumption simultaneously, along with a few other values, such as power limits 1 and 2 (PL1 and PL2), which apply to sustained power draw and peak power draw respectively. Most Intel processors up until this point will allow a processor to turbo, up to a peak power draw of PL2 for a fixed time, before enforcing a PL1 sustained power draw. This is all very OEM dependent as well. However, for Ice Lake, this changes a bit.

For Ice Lake, Intel has a new feature called Dynamic Tuning 2.0, which implements a layer of machine learning on top of the standard turbo mode. The idea behind DT2.0 is that the processor can predict the type of workload that is incoming, say transcode, and adjust the power budget intelligently to give a longer turbo experience.

Technically the concepts of PL1 and PL2 don’t magically disappear under this new regime – the processor ends up going below max turbo because the algorithm predicts that the user won’t need it, and this saves up ‘power budget’ in order to enable the turbo to work for longer.

This is a topic that Intel will hopefully go into more detail. We do know that it requires collaboration at the OS level, but how these algorithms are trained would be a useful trove of information. It is unclear whether Intel will allow this feature to be enabled/disabled at the user level, for testing purposes, but it should be noted that unless it is by default ‘on’ for OEM systems, we might end up with some systems enabling it while others do not.

Two Versions of Ice Lake, Two Different Power Targets Thunderbolt 3: Now on the CPU*
Comments Locked

107 Comments

View All Comments

  • Billy Tallis - Wednesday, July 31, 2019 - link

    A lot of the chemicals used in wafer processing are quite nasty. A bare wafer itself is pretty harmless unless you grind it up and inhale it. Solid elemental silicon or silicon dioxide is safe to handle.
  • repoman27 - Tuesday, July 30, 2019 - link

    I understand that there are embargos which must be respected and that Anandtech does not like to trade in unsubstantiated rumors, but much of what is presented regarding packaging and power seems a bit wishy-washy.

    Wikichip has had photos of both sides of the Ice Lake U and Y packages posted for some time now. Furthermore, Intel's product briefs are very clear on the power for each series:

    Ice Lake Y: Nominal TDP 9 W, cTDP Down 8 W on Core i3 only, cTDP Up 12 W but N/A on Core i3,
    Ice Lake U: Nominal TDP 15 W, cTDP Down 12 W (13 W for some UHD parts), cTDP Up 25 W
    AFAIK, no 28 W Ice Lake-U parts have been announced by Intel yet, but they most likely are in the works.

    And you can cite whatever reasons you care to, but by all reports Intel was initially targeting a 5.2 W TDP for Ice Lake-Y 4+2, and that entire platform has been solidly shifted into the 8-12 W range.

    Also, it should be noted that the 14nm 300 Series chipsets that Intel has been shipping for some time now are all Cannon Point, which was originally designed to complement Cannon Lake, and are almost identical in terms of capabilities to the 400 Series. And the particular designation for the Ice Lake PCH-LP according to Intel is "495 Series".
  • James5mith - Tuesday, July 30, 2019 - link

    Ian,

    You either have your graph or your paragraph about the store/load performance increases reversed.

    Graph says 72 -> 128 stores, 56 -> 72 loads. The paragraph below it says 72-128 loads, 56-> 72 stores.
  • ksec - Tuesday, July 30, 2019 - link

    While I do enjoy and mostly want to read Dr Ian Cutress article, I seriously don't want to read Intel's marketing hype. Action and Results speaks louder than Powerpoint slides. Ship it, let Anandtech test it. And we make an opinion on it.
  • Targon - Tuesday, July 30, 2019 - link

    The Ryzen 7 3700U is a Zen+ part on 12nm, without the big IPC plus clock speed improvements seen with the desktop CPUs. As a result, Intel is doing a comparison against the previous generation products for laptops.

    In laptops, getting max turbo or boost for more than one second is rare. Yea, Intel can put a laptop chip on a board on a bench without any enclosure to show the chip, but real world speeds will be quite a bit lower. That is true for both AMD as well as Intel, and it is up to the OEMs to come up with a design to keep the chips cool enough to run faster than the competition.

    AMD knows what is going on, so if I am correct, AMD will move up the release of the next generation of laptop chips to November. If AMD does the right thing, AMD will call the new chips the 3250U, 3400U, 3600U, and 3800U to bring consistency with desktop naming conventions. These new chips would be 7nm with either Vega or Navi, for an APU it is less important than going 7nm for both.
  • Drumsticks - Tuesday, July 30, 2019 - link

    Keep in mind, the only comparison they did with Ryzen (I think) was Graphics, not CPU. I'd imagine the Icelake chips have a solid CPU lead against quad core Ryzen based on Zen/Zen+. Zen 2 will certainly help close that gap, but it should still be roughly 15-20% behind Icelake in IPC, and I certainly won't be ahead that much on frequency.
  • Fulljack - Wednesday, July 31, 2019 - link

    I think in Q4 19 they'd release Ryzen 4000 series (based on Zen 2) and call it day, like last year or two year ago.

    they'll be 4300U, 4500U, and 4700U for U-series and 4350H, 4550H, and 4750H for high-performance part with integrated graphics based on Navi.

    but since Zen 2 has 8 cores now on each CCX, they'd probably also sell 6-core and 8-core part, but I don't know if they'll release it on U-series, though.
  • Apple Worshipper - Tuesday, July 30, 2019 - link

    Thanks Ian ! So how does Ice Lake purportedly stand next to Apple’s A12x in iPad Pro based on the Spec scores ?
  • PeachNCream - Tuesday, July 30, 2019 - link

    More importantly, how does Ice Lake taste?
  • HStewart - Tuesday, July 30, 2019 - link

    I have seen anything that was successfully comparing x86 based cpus with AEM based cpus

    But one things - that makes all this MacBook ARM stuff meaningless to me is one sheer fact - Apple has yet to release development tools for iOS on actual iOS. It might be Apple trying force Macs for development but Apples own development tools don't run on iOS

Log in

Don't have an account? Sign up now