AMD offers a power plan in its third Ryzen Community Update

It's a happy day for Ryzen-equipped gerbils. AMD has put out its third Ryzen Community Update, and today's blog post is all about power. The company created its own Ryzen Balanced power plan for Windows 10, and it's now asking gamers with Ryzen CPUs to try it out.

In case you don't remember, AMD previously recommended that users looking for every possible slice of performance use Windows' High Performance power plan. That move is meant to curb the OS' aggressive core parking, which can result in slightly reduced performance in some games. However, keeping a system on this plan disables quite a lot of Windows' power-saving mojo and can lead to higher system power consumption, so AMD felt it needed some sort of compromise. Enter the AMD Ryzen Balanced plan, downloadable here. Gerbils can install it with a double-click after decompressing the archive, and they'll be greeted with a new plan under Windows' power options.

AMD says that gaming performance with the Ryzen Balanced plan should be pretty close to that of the High Performance setting. To illustrate that fact, the company made a graph comparing the performance of its new plan against High Performance, using the stock Balanced setting as a baseline. According to AMD's data, the difference between Ryzen Balanced and High Performance should be borderline negligible, barely ever reaching a 1% delta.

The company said that it intends to make this power plan part of its chipset drivers for Ryzen CPUs. AMD also notes that while the change in power plans doesn't work out to a performance improvement in many titles, it believes that "there are enough [affected] games to warrant a change." Those looking for the skinny on what exactly the plan does and why can dig into the P-state and C-state talk in the community update.

That's not the only news today. Total War: Warhammer should now be a little faster on Ryzen CPUs after the recent "Bretonnia" patch, at least at 1920×1080 and Ultra settings. The Ryzen Master monitoring and overclocking utility also got a small version bump to 1.0.1. The utility now reports junction temperature instead of tCTL (more details here), and no longer requires High Precision Event Timer (HPET) to be enabled on machines with an AGESA 1.0.0.4-based BIOS.

Comments closed
    • Bensam123
    • 3 years ago

    Core parking was never about overall fps, it’s all about reducing microstuttering. That’s specifically something gaming communities talk about when they approach this topic. Someone needs to take fcat to this.

    • derFunkenstein
    • 3 years ago

    OK so I’ve installed this power plan. It’s basically something very close to the High Performance Power Plan. Min CPU state is 90, max is 100 (instead of 100 on both for High Performance or 5/100 for Balanced). CPU-Z says the CPU spins down to 2.7GHz under this power plan and 1.55GHz under the Balanced plan. High Performance (obviously) doesn’t drop down at all – you’re at max clocks all the time.

    Very Unscientific Kill-A-Watt P3 idle power consumption results for default Balanced, High Performance, and AMD:

    Balanced: 49W
    High Performance: 61W
    AMD: 52W

    So it gets you ALMOST down to “balanced” power consumption, at least on my system. A Prime95 stress test is obviously the same on all three settings, which is between 175-225W with my relatively mild overclock, depending on whether small FFTs are active.

    So I can really crank up the power consumption. πŸ˜†

      • cegras
      • 3 years ago

      Sounds like it’s working as intended.

      [quote<]What about power? Now that you know a little more about the performance of our new plan, let’s talk power. The AMD Ryzen Balanced power plan does not change how our processor handles low-power idle states called β€œCC-States.” These CC-States number cc1 through cc6, representing increasingly aggressive clock and power gating. In fact, cc6 represents a core that is essentially turned off. The core is sleeping so deeply that only its voltage can be detected by software. The sophisticated power management technology in the β€œZen” core can autonomously enter and exit these CC-States as quickly as 1ms. Software tools, unable to see through the sleep, will simply report the last P-state known to the OS before the core entered a CC-state. Don’t be alarmed! The effective frequency of a sleeping core is much lower (generally sub-1GHz).[/quote<]

    • freebird
    • 3 years ago

    I’ve been running a Modified High Performance Power Plan for the last 2 weeks. Allows my Ryzen to run all cores at 4GHz when stressed, or only one+x at 4GHz and the rest as low at 1550 Mhz when just the background stuff and a browser are running…

    Voltage also drops with the core speeds even though I have a FIXED setting of 1.3825 in the BIOS on my AsRock Fatal1ty X370 Pro Gaming Motherboard. Seen it down around .9v at times.

    Sleep function works fine also, but I think I disabled Hybrid sleep along with Hibernation, since I don’t need it burning out my evo 960 with 30-40GB of writes every time my 64GB PC “Hibernates”… ugh.

    • Questar
    • 3 years ago

    If this decreases the time it takes Windows to change power states, would this work on Intel CPUs also?

      • colinstu12
      • 3 years ago

      no

    • wingless
    • 3 years ago

    People are so unrealistic! They expect AMD to make an identical copy of Intel’s Core i7 or something. It’s not 1991 and AMD can’t do that anymore folks. A different micro architecture requires a different approach and some tweaking by OS and software developers. Heck, when Intel makes big changes to their architecture, Microsoft and devs have to make tweaks and changes for it just the same!

    People have short memories.

      • DancinJack
      • 3 years ago

      I think the criticism, at least mine, comes from the fact that AMD rushed out the release of Zen for some reason. Most of these issues could have been avoided, but they brought them on themselves for some odd reason.

      I don’t think people are thinking like you said tbh. If they are, they’re confused.

      • etana
      • 3 years ago

      i loved my 386DX40

    • Concupiscence
    • 3 years ago

    Hey, the link to the third Ryzen community update’s broken! Just letting you know.

      • morphine
      • 3 years ago

      That’s what happens when you copy/paste under the High Performance power plan.

      Fixed, thanks πŸ™‚

    • barich
    • 3 years ago

    This unfavorably reminds me of the early days of SSDs when companies like OCZ would give you a huge list of “tweaks” that you shouldn’t have to do in order to mask deficiencies in their products.

    Not to put AMD on the same level as OCZ. But I’d be inclined to hold off on a Ryzen-based system for a bit until they’ve worked out this sort of thing.

      • swaaye
      • 3 years ago

      Bulldozer needed scheduler tweaks.
      Phenom needed power management tweaks.
      Athlon 64 X2 needed a TSC fix (dual core optimizer)

      Things that were eventually forgotten and built into the next OS. I don’t understand why AMD doesn’t get these things addressed before launch though. Companies do a massive amount of validation for CPUs but we always see these seemingly obvious quirks shortly after a new AMD architecture launch.

      Though one interesting benefit is the continuous press they get as they address these issues. Hmmm.

        • tipoo
        • 3 years ago

        Intel has all the clout with Microsoft, AMD has very little, I assume.

          • swaaye
          • 3 years ago

          I don’t know. This processor has been in the works for years, across 2 Windows generations. I don’t even believe that they could be unaware of the problems exposed by reviewers. This silly power plan should’ve been ready for launch day. Etc.

          Though on the other hand, we have no idea of what they know that we don’t know. πŸ™‚ There may be loads of ugly quirks that have been worked around to get this chip out the door.

        • Krogoth
        • 3 years ago

        Intel also had issues with its new CPU features in the past.

        Haswell’s promised “TSX” on Haswell was broken and disabled in a micro-code patch and several of their previous CPUs had power management issues that need to be tweak.

    • TwistedKestrel
    • 3 years ago

    I find the level of communication and speed of action from AMD very encouraging. If I had the budget for a new build, I’d be Ryzen to the occasion – it feels like AMD is taking care of the early adopters

    • Rza79
    • 3 years ago

    Windows 10 Creator Update also brings in a small boost to some games.

    [url<]https://www.computerbase.de/2017-04/ryzen-windows-10-creators-update-benchmarks/#abschnitt_benchmarks_von_ryzen_mit_dem_creators_update[/url<]

      • derFunkenstein
      • 3 years ago

      For me it brought around a 10% or so uptick in the built-in GTA5 benchmark.

    • Jigar
    • 3 years ago

    My gaming PC is always on High Performance Mode, ain’t got no time to save power when i am playing game.

    • sophisticles
    • 3 years ago

    A perfect illustration of just how pathetic Ryzen actually is; AMD, instead of putting out a compelling product put out an “ok” product and then put together a “power plan” in order to get an extra 1% performance improvement.

    I don’t know what’s sadder, the fact that AMD did this or the fact that invariably there will be a bunch of AMD “fans” that make it seem like this is a great thing.

    I know I won’t be buying an AMD processor in a while.

    Edit: ROTFLMAO @ all the AMD faithful that are claiming that I’m a Intel shill or an “idiot”.

    Let’s examine that for a second. Intel earns billions per year, AMD released an 8 core 16 thread processor that loses in many cases to the midrange Intel 4 core 8 thread offering, AMD’s offering has a poor AVX2 implementation, it can’t use the fastest ram (even if it could that wouldn’t change thinks all that much) but I’m a shill and an idiot. OK.

    For the record I also consider Intel’s high end 8C/16T offerings to be over-priced and not worth the money but don’t let facts get in the way of your vitriol.

      • Krogoth
      • 3 years ago

      Go away shill

      • Mat3
      • 3 years ago

      Intel fanboy or Intel marketing, I wonder which you are…

        • thx1138r
        • 3 years ago

        This guy is a fanboi/troll. Intels marketing on the other hand is done by Chuckula πŸ™‚

        • Redocbew
        • 3 years ago

        Intel is actually good at marketing. I see no evidence of that here.

        • Kretschmer
        • 3 years ago

        Intel wouldn’t pay for that.

      • Jigar
      • 3 years ago

      This is not WCCF, go take your garbage to that website.

      • Amiga500+
      • 3 years ago

      AH. You tried to disguise it there the other day.

      But now its clear to all. Your an idiot or shill (possibly even both).

      • ptsant
      • 3 years ago

      It appears to me that you underestimate the launch of a completely new architecture.

      There is a ton of work to be done, including small tweaks like this specific power plan, the new BIOS-core that corrects the FMA3 bug and apparently improves memory latency by 6ns, the validation and launch of Ryzen-specific, the distribution of Ryzen-specific development tools (compiler tuning, math libraries like BLAS/LAPACK) and even the creation of Ryzen specific game and application updates.

      All that cannot be done on day 0, especially when we’re talking about a brand new architecture (not a tick or a tock….).

      You may not like Ryzen, which is understandable if you have a different use case (pure gaming?), but you can’t really think that any major CPU launch can happen without this sort of difficulties.

        • w76
        • 3 years ago

        Well, I think Intel considers a tock a new architecture, but this is a radical departure from the one prior, I get what you mean. Similar to Netburst -> Core.

        Anyway, I think the trolls are a little harsh, it was somewhat disappointing but I think the first refresh of Ryzen will put it in pretty good shape, assuming Intel remains somewhat stagnant.

      • Welch
      • 3 years ago

      Only a true fan denies AMD’s divinity!!!

      • PrincipalSkinner
      • 3 years ago

      We’re not impressed.

      • GrimDanfango
      • 3 years ago

      [quote<]Edit: ROTFLMAO @ all the AMD faithful that are claiming that I'm a Intel shill or an "idiot".[/quote<] You realise that refering to the normal, mostly-unbiased regulars of this site as "the AMD faithful" does little to make you look less like a shill-and/or-""idiot"", right?

    • Shobai
    • 3 years ago

    Speaking of which, Jeff, how’s progress on the “What we’re not testing today” article?

      • Peldor
      • 3 years ago

      It would be particularly interesting to get his thoughts on this as TR usually (always?) tests at the “Balanced” power profile.

      Of course it would be even more interesting to get his test results on this.

    • Bumper
    • 3 years ago

    I’m impressed they followed through with their promises on schedule. AMD’s cpu division is a prideful bunch.
    The graphics division seems to be getting better at driver releases too. It’s too bad they fud up Polaris’s launch with that power issue. I don’t think raja lasts past navi if it isn’t a success. Though I do wonder if those mobile gpus with HBM2 that I’ve heard about will have the HBCC and how that will play out with vega…

    • Dizzytaz00
    • 3 years ago

    Untill AMD gets all the bugs out. I’m not putting a 1 cent into a AMD Build. Just craziness on Amd cpu front!

      • just brew it!
      • 3 years ago

      Something that results in a 1% to 8% performance difference depending on the power plan being used hardly qualifies as a “bug”. It’s a “tweak”. Aggressive power management almost always involves performance tradeoffs, otherwise there would be no reason to have a way to dial it back or disable it.

        • DPete27
        • 3 years ago

        Typical AMD. They can’t compete with the other guys so they pump their chips full of power to try and close the gap.

          • Krogoth
          • 3 years ago

          Intel does the same bloody thing with their highest-binned chips. i7-6950X and i7-7700K aren’t exactly easy on power and sacrifice power efficiency for more performance.

        • POLAR
        • 3 years ago

        1% to 8% performance difference qualifies as “next generation” among cyan buffalo pastures.

      • Ryhadar
      • 3 years ago

      I have a feeling you’ll get downvoted for your comment but I kind of agree with you. Ryzen still has a very beta smell and I’m saying that as a fan and shareholder. If my PC died I’d probably go with a Ryzen build but I’m in no rush to replace my (mostly) stable 2600K system.

        • Bumper
        • 3 years ago

        It’s not about what he said, it’s about how he said it. it read like he’s trolling. “just craziness on Amd cpu front!” I picture him trolol after that.
        the only big bug I’ve heard of is the FMA3 issue, which has a fix incoming through bios update.

        Any new adopter deals with new issues. I always wait a few months before jumping in if I do. Some people love new technology though and just getting their hands on it. I can catch that fever, but try not to.

          • just brew it!
          • 3 years ago

          Yeah, the FMA3 thing is definitely a bug, and AMD really should’ve caught that in validation. Sloppy. Good thing bugs like this in modern CPUs can be fixed with microcode patches; it’s not like the bad old days of the FDIV bug where the only fix was to physically swap the CPU chip.

            • Amiga500+
            • 3 years ago

            Actually, from what I gather the FMA3 thing isn’t really a bug — the string of operations you need to call to trip the bug are essentially NULL operations.

            i.e. the equivalent of
            a+b = c
            c-a = b
            a+b = c
            c-a = b
            repeated ad naseum

            The idea of these calls was to stress the FPU – to ascertain power consumption.

            I don’t know if you can actually trigger the bug while doing actual work.

            • just brew it!
            • 3 years ago

            Even if that’s true (first I’ve heard of it, but it seems plausible) it’s still a bug, with some potentially nasty implications.

            Arbitrary sequences of instructions in user space should not be able to cause the entire system to hard-lock, ever. Worst case they should result in an application crash.

            At the very least, this leaves the system more vulnerable to DOS exploits.

            • freebird
            • 3 years ago

            Did they fix it?
            How long did it take to fix it? 2-3 weeks and BIOS update should be available shortly if not already.
            Do you run FMA3 on a Ryzen?
            IF NOT, Y do you give a Ratz Azzz.

            I own a Ryzen and haven’t encountered the FMA3 bug and I doubt 99% of other owners have either unless they ran the 5-year old FPU stress test that found the bug…

            I’m just glad the WILL fix it and quickly as any good manufacturer should do…

        • brucethemoose
        • 3 years ago

        [quote<]2600k[/quote<] Speaking of buggy releases... I remember Sandy Bridge chipsets having some significant launch issues. Not so say that it's a bad platform. But launch issues do seem to iron themselves out in the long run.

          • Goty
          • 3 years ago

          The X58 chipset had its own issues as well, such as the S3 sleep bug, which required a hardware revision to fix.

            • the
            • 3 years ago

            There were tons of bugs around virtualization on X58 and the server equivalent 5500 series.

          • Krogoth
          • 3 years ago

          The PCH on the first stepping of 6 series chipset boards had a defective SATA controller which caused performance issues and data corruption.

          There was an industry-wide recall on affected motherboards.

          • Welch
          • 3 years ago

          Yup and I was one of those laughing about Sandybridge SATA bug all the while telling people to wait for the amaaaazing Bulldozer.

          Long story short, I’ve been using my i5-2500k since 2012, LOL.

        • sreams
        • 3 years ago

        “I’m in no rush to replace my (mostly) stable 2600K system”

        Which speaks just as much to Intel as it does to AMD. Without competition, Intel has moved very slowly to offer anything more compelling than your 6 year old CPU. Intel and AMD fans alike (and those who don’t care either way) can be thankful for AMD’s return to the game.

        • terranup16
        • 3 years ago

        Coming to an R7 1700X and AM4 X370 Prime from an i7-950 and ROG Formula III, feels pretty mature. I run my 3200MHz sticks at 2933MHz is the most beta I’ve found it to be so far (and honestly I had been looking at 3000MHz before Ryzen compatibility came out and the price and frequency difference is pretty much nothing there), but CAS timings are rock solid and the single thread gains even in 1080p gaming compared to Nehalem are appreciably mind blowing.

        But yeah, 2600k, as long as it’s chugging along, there aren’t many things that won’t do well today, especially if you have an OC on it. I suspect that may change as higher core counts with Haswell+ IPC become more affordable and apps take advantage of that better.

        • just brew it!
        • 3 years ago

        My original plan was to wait about 6 months after launch to give some time for the early adopters to find the issues, for UEFI implementations to settle down, and for support in the Linux kernel to mature.

        Given that I’ve got an Intel build in the pipeline now (due to winning some gear in the holiday giveaway) my Ryzen plans have been pushed out. The new plan is to put Windows on the Intel-based rig, and use that to get back into a little gaming (it has been too long). The Linux rig (which will remain my daily driver for non-gaming use) will stay on the FX-8350 for a while yet…

          • Krogoth
          • 3 years ago

          You also forgot to mention having a selection of motherboards and CPUs with confirmed ECC support. πŸ˜‰

            • just brew it!
            • 3 years ago

            I think motherboard ECC support is more or less confirmed at this point, at least for ASRock. They’re not normally my go-to brand, but I’m willing to consider them if Asus drops the ball.

            • Amien
            • 3 years ago

            I bought an Asus b350 prime plus with 16GBx2 of ECC 2400 mhz RAM(Kingston). The motherboard was a bit of a compromise because all x370 boards were sold out. Asrock support is particularly bad where I live so that wasn’t an option either. I’ll be over the moon if ECC works.

        • derFunkenstein
        • 3 years ago

        As an owner, I do feel like a beta tester sometimes. We’re almost to something worth releasing. Too bad it’s a month AFTER release.

      • albundy
      • 3 years ago

      bugs out? we cant even get good motherboards for them! i feel like i will be waiting months or probably more for some decent motherboards to roll out.

        • JustAnEngineer
        • 3 years ago

        The [url=http://www.gigabyte.us/Motherboard/GA-AB350M-Gaming-3-rev-10#kf<]GA-AB350M-Gaming 3[/url<] looks like a decent motherboard.

      • ultima_trev
      • 3 years ago

      X99 and Z170 had their own teething issues as well, this isn’t just an AMD thing.

      • Dizzytaz00
      • 3 years ago

      As system/ enthusiast builder perspective , What i’m getting at is the Amd Ryzen Cpu’s aren’t the greatest @ this point & that being said, By review after reviews. The ( my way of i’m seeing it!) as JBI said (TWEAKS!)
      I’ve waited until ryzen come out & not impressed at all, on any level of cpu performance. ( that’s being of my last build was a phenom II 955 black).

      I’m seriously think on buying intel I-7 7700K!

      I’m only 4 Major components away, Motherboard, Ram, CPU, & Graphic card, from finishing my build.

      BTW: Down votes welcome! Doesn’t bother me one bit!

        • derFunkenstein
        • 3 years ago

        well, I don’t think you really earned them and I tried to undo a few of them, because it’s really feeling kinda beta-y right now. Maybe if board makers can incorporate the coming AGESA fixes in the next couple of weeks things will start to look up.

        • ultima_trev
        • 3 years ago

        So that (at best) 10-12% improvement in gaming performance is enough for you to support a monopoly, ignoring the fact that many games these days (and this trend will likely continue) support more than 8 threads? Ignoring the fact that Ryzen scales obscenely well with higher speed DDR4?

        Well, if Intel’s monopoly means than much to you, go ahead however don’t complain when AMD goes under and Intel starts charging $1,000 for Pentiums and $5,000 for i7s.

          • derFunkenstein
          • 3 years ago

          This is dopey. People should buy what’s best for them. If he thinks an i7 is best for him, he should buy it. It’s not his fault AMD hasn’t met his expectations. AMD hasn’t met a lot of expectations up to and including Ryzen 7.

            • Redocbew
            • 3 years ago

            Agreed. In terms of performance most of the fuss since launch has been about details which are relatively minor. Neither Ryzen nor an i7 can be called a bad choice in general. It’s not the same as it was with Bulldozer. We’re comparing oranges and tangerines not oranges and pork chops.

            • derFunkenstein
            • 3 years ago

            All of these things that have been relatively minor can add up, though. So this power profile gives folks on average a 5.5% boost, let’s say. And AMD’s coming AGESA update gives another moderate boost since it’ll deal better with faster memory. And I saw another ~5% boost or so when I got my RAM up from 2400MHz to 2933. Well now there’s a 15-17% boost that AMD could have had at launch. Launching early like they [url=http://www.legitreviews.com/one-motherboard-maker-explains-why-amd-am4-boards-are-missing_192470<]apparently did[/url<] really did short-term damage to Ryzen's performance perception. They might claim some of that back on the 11th when some of this stuff is in effect for the Ryzen 5 reviews. But let's face it. We have no choice but to think AMD believed it had no choice. The company's CPU division has been revenue-starved for years. It hadn't produced a new FX CPU since like 2013. AMD's powers that be clearly felt it was now or never when waiting for May 1 might have been the right call. It's unfortunate. I say all that to say this: you're right that neither Baby Lake or Ryzen is a bad choice, depending on what you're doing. With that in mind, Ryzen could have looked like a FAR BETTER choice in non-optimal workloads than it was at launch if that launch had been delayed.

            • cegras
            • 3 years ago

            Why put so much emphasis on past Ryzen? It sounds like you are making an argument not to purchase Ryzen in its current form based on its past performance.

            • Flying Fox
            • 3 years ago

            I do see the bright side of this early rushed launch. It puts pressure on everyone, including AMD, to fix the kinks. Come back to school time, meaning mass production by [i<]July the latest[/i<], we should be on good footing. They may have missed some tax refund sales of March and April, but back-to-school is still a bigger season. We'll see if they score any OEM design wins soon enough.

          • Dizzytaz00
          • 3 years ago

          I’m not for Intel by anyway shape or form. I glad to see Competition & Amd trying to make a come back!

          As a consumer your choice for enthusiast for gaming are marginal to be the lease.

          • Meadows
          • 3 years ago

          Couldn’t name a single game off the top of my head that meaningfully uses more than 8 cores, or even just 8 cores flat.

        • Krogoth
        • 3 years ago

        I7-7700K doesn’t make sense unless you need as much single-threaded performance as much possible and aren’t interested in overclocking. The HT on it only helps in multi-threaded workloads but Ryzen R7 1600 and 1700 are better at such loads.

        You should opt for the more reasonable i5-7600K which can easily overclock to i7-7700K levels.

          • Flying Fox
          • 3 years ago

          I thought the 7700K OCs just fine if you win the TIM lottery, or if you dare to delid?

            • Krogoth
            • 3 years ago

            You are spending an extra ~$100 for [b<]Hyper-Threading[/b<] and a minor bump with Turbo and base clockspeed over an i5-7600K. πŸ˜‰

        • Chrispy_
        • 3 years ago

        Aren’t the greatest at what? They’re not the greatest at running Intel-optimised games, no. They’re Haswell-era performance which is okay and within a few percent of current Kaby-lake. Certainly, the high-clock, high-IPC advantage of Intel’s current architecture is all that matters to current games on current engines.

        Everything else, especially multithreaded workloads? OMG they’re 200% better than Intel in the price/performance department when Intel have been making single-digit improvements for almost a decade.

      • ptsant
      • 3 years ago

      Well, I have to say I expected a lot of bugginess when I built my new Ryzen PC, but this was the very first time I had a machine boot correctly the first time and directly into my old windows installation without any error whatsoever. I couldn’t believe it: windows booted without even asking for drivers, all USB peripherals were working, my tri-display setup was working and not a single “unrecognized” device.

      I am aware of the difficulties that several people have reported (especially MB bugginess), but for me the experience has been incredibly smooth. My guess is that by early May, the things will be reasonably stable.

        • LostCat
        • 3 years ago

        My onboard audio has been buggy. I may or may not have caused it but I doubt it.

        Usually I use HDMI audio so I don’t actually care, but headphones are plugged into Xbox controller atm.

        • derFunkenstein
        • 3 years ago

        I couldn’t boot a Windows install from an Intel system, but the motherboard I started with was flaky as hell. Maybe this Gigabyte would have worked.

Pin It on Pinterest

Share This