Upcoming Ryzen BIOS updates could improve memory performance

As we and other sites have demonstrated, using fast memory can result in a significant performance boost on systems with AMD's Ryzen processors. It can be a trying task to actually get your Ryzen rig to make use of fancy overclocker memory, though. AMD's 1.0.0.4 AGESA microcode update seemed to help matters somewhat, and some users reported improved memory speeds after that update. Now we have word—by way of a Gigabyte employee's forum posts—that AGESA update 1.0.0.5 is on its way with further improvements for memory compatibility.

Gigabyte rep Matt posted in the company's forums last Wednesday confirming that the company is working hard to use the new AGESA microcode in its BIOS updates to help improve its AM4-platform motherboards. The new updates should include fixes to IOMMU support, "soft brick" issues, and the ability for the CPU to enter lower power states when it's overclocked. Matt stated that the firmware updates would include AMD's latest AGESA microcode version 1.0.0.6. However, last Friday he posted in the same thread correcting himself and stating that the AGESA update would be version 1.0.0.5.

Either way, he says that the microcode update should enable access to "20+ memory registers" in the Ryzen CPUs' memory controller, and that the change should improve memory compatibility. It seems unlikely to us that these registers were actually disabled, but perhaps they were locked down and unavailable for modification. Matt says the BIOS updates for Gigabyte's boards should be available this week.

It's fairly safe to assume that Gigabyte isn't the only company readying BIOS updates with the new AGESA microcode. Asus, MSI, and all the other manufacturers are certainly working hard on new firmware as we speak.

Comments closed
    • Meadows
    • 2 years ago

    What the hell is a “soft brick”? Besides a very unreliable construction material.

    • albundy
    • 2 years ago

    hopefully we will see new and better am4 boards at computex.

    • Laykun
    • 2 years ago

    Is the Catalyst driver team working on the CPU side of things now?

      • ermo
      • 2 years ago

      Yeah, poor guy.

    • DPete27
    • 2 years ago

    I mean…. >DDR4-2400MHz dual rank RAM has been around for how long? They couldn’t squeeze this support in before launch? C’mon guyz.

      • derFunkenstein
      • 2 years ago

      If your RAM has DDR4-2400MHz in the SPD, the CPU could use those at launch. The problem is that at least some faster kits, like my Corsair DDR4-3000 kit, only had SPD profiles up to 2133 and relied on XMP (which did not work well for me) for higher speeds. Gigabyte’s current BIOS with 1.0.0.4 enabled my memory to run at 2933, resulting in a nice performance boost.

        • DoomGuy64
        • 2 years ago

        One thing I’ve noticed about XMP is that it doesn’t always give correct timings for lower speeds. You pretty much have to run it at the exact speed for it to work correctly. I have this issue with a 3600 kit running @ 3200. The auto timings set by the motherboard are worse than the 3600 timings.

          • derFunkenstein
          • 2 years ago

          The AB350 Gaming 3 picks the right timings for 2933 (same as 3000)

          • morphine
          • 2 years ago

          That’s going to vary more by motherboard implementation than by RAM spec.

        • Chrispy_
        • 2 years ago

        Indeed. The problem with XMP is that it’s an Intel-specific timing standard. It “massages” the true speeds of the RAM at standard JEDEC timings by shortening the phases of a clock cycle that Intel memory controllers are optimised for.

        Just because RAM has an XMP profile for DDR4-3000 doesn’t mean it can run at that speed with JEDEC timings. Blame Intel’s near-monopolisation of the processor market for manufacturer’s behaviour. It’s now really difficult to find the genuine JEDEC timings on memory and it’s almost all sold by it’s XMP speed :\

      • TwistedKestrel
      • 2 years ago

      At some point you have to launch with what you have. I’m pretty happy they can tune the memory controller post launch at all

      • stefem
      • 2 years ago

      But then you couldn’t go around saying your products age like fine wine 🙂

      • blastdoor
      • 2 years ago

      The Death Star is knocking at your door. Do you try to nail down every last issue with your X-Wing, or do you launch and hope that R2 can fix any glitches in flight?

      Make the wrong choice and you’re dead. Make the right choice and you’re still probably dead.

        • Whispre
        • 2 years ago

        Since R2 is the defacto “go to guy” for saving the day in the Star Wars universe… I’d say stick with R2.

          • BurntMyBacon
          • 2 years ago

          Why did they have to go with R3, R5, and R7? If only they had stuck with R2, they wouldn’t be in this mess. ;’)

      • Anonymous Coward
      • 2 years ago

      Give AMD credit for what they did do. (A lot.)

      • ikeke1
      • 2 years ago

      I’ve been running 4x8GB of dual rank memory @2400 since i got my Ryzen setup.

      [url<]http://valid.x86.fr/njbcn7[/url<]

      • ptsant
      • 2 years ago

      The XMP profile that configures this memory for the Intel chips may not work for Ryzen. That doesn’t mean that Ryzen itself can’t handle this memory, simply that it needs an appropriate profile because it is a different architecture. If you actually buy memory that is validated for Ryzen (read the MB support/compatibility page) it will almost certainly work as advertised.

      Anyway, the 2933MHz is probably the sweet spot for Ryzen and any obscene overclocks after that will need to decouple the internal fabric from memory (which is the major source of improvement as frequency increases from 2133 to 2933 MHz). Gains are usually minor in that case.

      • ptsant
      • 2 years ago

      By the way, I am running 2x16GB dual rank at 2667MHz and I could almost certainly push it to 2933 MHz with some tweaking. I might try this after the next BIOS update.

    • K-L-Waster
    • 2 years ago

    This is exhibit A in why buying a new platform early is risky.

    I think it may actually be worthwhile for TR to do a “RyZen Revisited” review around mid-summer to see how much performance and stability have improved since launch. The BIOS fixes may actually move some of the points on the scatter plots….

      • derFunkenstein
      • 2 years ago

      Once TR gets around to “part 2” of their Ryzen 5 review with the extra CPUs Jeff received, I’m hoping that updated Ryzen 7 results are included. Part 1 already used faster RAM speeds than the launch-day Ryzen 7 review.

        • Shobai
        • 2 years ago

        If they do get to publishing ‘Ryzen 5 part 2’, perhaps they could do the “what we’re not testing today” part of the Ryzen 7 review?

          • derFunkenstein
          • 2 years ago

          also hoping for that, too

      • just brew it!
      • 2 years ago

      Yup. I’m letting all you early adopters deal with the buggy BIOSes/EFIs, CPU scheduler patches, etc… I’ll probably build a Ryzen system this fall at the earliest. Maybe RAM prices will have come back down by then too.

        • K-L-Waster
        • 2 years ago

        I’m thinking the same, although I want to see what (Coffee Lake / Cannon Lake / whatever obscure body of water it is) does before deciding whether to go AMD or Intel.

          • sreams
          • 2 years ago

          And then, once Coffee Lake comes out, you can wait and see what Ryzen 2 does. And then once that comes out, you can wait to see what Intel does next. And then wait for AMD again. An excellent way to save money.

            • K-L-Waster
            • 2 years ago

            Lol.

            In this case it’s more like “AMD has hexa-core and octa-core systems with good performance, Intel is s’posed to be coming out with hexa-core systems that won’t cost you your first born plus an arm and a leg, let’s see which one turns out to be the better choice.”

            • just brew it!
            • 2 years ago

            Skip all that and just wait for the Singularity.

          • jts888
          • 2 years ago

          Coffee Lake is still supposedly soon-ish, but I wouldn’t hold my breath for Cannon Lake, which will be stuck in 2c mobile chips until 2019, assuming zero further roadmap slippage.

          Intel’s 10nm is having a rather bumpy deployment.

      • UberGerbil
      • 2 years ago

      Even with Intel systems — and even with as little difference there is between their CPU “generations” at this point — there’s always a flurry of BIOS updates in the first few months, and you frequently hit weird “did I screw something up, or is it the BIOS” errors — especially if you’re trying to push it at all. Unless you absolutely need the latest product for some reason (and simultaneously don’t need it to be 100% reliable) it’s always best to wait about a calendar quarter while the people who can’t sit on their wallets endure the grief of being unpaid beta testers on your behalf.

        • K-L-Waster
        • 2 years ago

        Exactly. The price of being an early adopter is you get to become an intern in their QA department for a few months.

      • odizzido
      • 2 years ago

      I would like this, actually I would like to see ryzen used in the review of AMD’s upcoming GPUs.

      My dream review would be something like:
      intel CPU + nvidia/AMD GPU tests
      AMD CPU + nvidia/AMD GPU tests

      there has been some funky stuff going on with AMD CPUs + Nvidia GPUs and I want to know more.

        • BurntMyBacon
        • 2 years ago

        I’m not seeing any “funky stuff” going on with my Phenom II x4 955 (4.0GHz) paired with a GeForce 780Ti. I built a Ryzen R7 1700 with a GeForce 1070 for a friend about 3 weeks ago and as of last night, he is having no issues either.

        Please enlighten me. Is there a particular game or program that is having issues? Is it motherboard or manufacturer specific? Perhaps it only occurs with XFR, specific video cards, or sync capable monitors. Is it perhaps related to the driver version?

        When you talk about “funky stuff”, it is helpful if you at least describe what “funky stuff” means. Otherwise, we don’t have a starting point to research it. Better still would be adding some information that might help prospective buyers avoid the issues.

      • kn00tcn
      • 2 years ago

      how is it risky if the boards/cpu/ram are usable & only get better over time? where’s the data loss or throwing away parts?

        • K-L-Waster
        • 2 years ago

        In this case there hasn’t been data loss — but it’s not impossible that there could have been, especially if the memory controller had been buggered badly enough.

      • ronch
      • 2 years ago

      That’s why I’m waiting for later iterations of Zen. Maybe Zen 2.0 or 3.0. The other day my FX was going bonkers and I was afraid it was the mobo. While the prospect of buying a set of parts for Ryzen now may seem exciting, I really hoped my mobo isn’t done for just yet because I want to get a (much) more refined and mature Ryzen platform. Now just isn’t the time and I only use my PC for downloading YouTube videos and playing old games like Thief so it’s not like I really need more performance. Yes, I could use some more energy efficiency but spending ~$800 just because Ryzen is more efficient is hardly justifiable. Hopefully my current PC lasts long enough to get to next year or so.

        • K-L-Waster
        • 2 years ago

        Even a Zen 1.0 with a Rev 1.1 motherboard and a 1.0.0.9 BIOS could make a world of difference.

        • MOSFET
        • 2 years ago

        It was probably just Windows 7 going bonkers. Reboot that old OS 🙂

      • Anonymous Coward
      • 2 years ago

      I think exhibit A was the TLB bug.

        • raddude9
        • 2 years ago

        For me the pentium FDIV bug would be exhibit A for CPU’s, with phenom TLB as exhibit B.

      • ET3D
      • 2 years ago

      On the contrary, those who buy early get the most “free upgrades”. If they were happy with what they bought, they’d be even happier later.

        • K-L-Waster
        • 2 years ago

        Call me old fashioned, but I don’t consider “now it does what we said it would do when you bought it” to be a “free upgrade.”

      • DoomGuy64
      • 2 years ago

      Not if you didn’t need OC’d ram. This is more like platform maturation instead of early adopter problems. Early adopter problems would be the windows power profile issue that requires a driver update.

      • albundy
      • 2 years ago

      but is it really? DDR4, usb 3, m.2, all the tech besides the chipset has been out for quite some time. there was no excuse for this.

      • Geonerd
      • 2 years ago

      Absolutely! The lingering “memory issue” is the one thing I’m waiting on before boarding the Zeppelin.

Pin It on Pinterest

Share This