Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×
AMD IT

AMD Improves Zen 5 CPU Latency and Performance With BIOS Updates 17

AMD has released BIOS updates to boost performance and reduce latency for its Ryzen 9600X and 9700X processors. The updates come a month after disappointing Zen 5 desktop CPU reviews and coincide with Windows 11 optimizations for AMD chips. The new AGESA PI 1.2.0.2 firmware addresses high core-to-core latency issues and introduces a 105-watt cTDP option, promising up to 10% performance gains for multithreaded workloads.

AMD Improves Zen 5 CPU Latency and Performance With BIOS Updates

Comments Filter:
  • by Targon ( 17348 ) on Monday September 30, 2024 @10:56AM (#64828373)

    AMD released the new AGESA version weeks ago, but it has taken motherboard makers time to prepare the new BIOS versions and test them before making these updates available. Now, the problem with Windows 11, and how things were actually reducing even Zen4 performance before the latest Windows Insider builds(as of two months or so ago, maybe longer), makes me wonder if someone at Microsoft was paid off by Intel to hurt AMD performance. The updates to Windows did improve Zen4 performance as well, even without the new BIOS, so you have to wonder if it was deliberate.

    On top of that, Zen5 launched with lower TDP ratings than Zen4, so "efficiency" was better, but the result was that performance wasn't as good as it should have been. The new AGESA allows users to get the "proper" TDP rating compared to Zen4, which is why performance will be higher. There is no problem with CPU failures due to this change though, but it will increase power use a fair bit.

    • Any idea whether these AGESA updates will be released for AM4 boards as well and will improve Zen3 and Zen4 performance on Windows 10?

      • I would think the performance issues with Windows 11 are related to relying on all the security features in modern processors. Except those security features had never really had that much real world use, so they didn't know how badly their implementation would make everything perform.

    • by higuita ( 129722 )

      It is more simply justified by devs using intel machines and optimizing for intel... and not noticing that they affected badly the amd performance...
      or maybe they noticed, but let it to be fixed in the next release cycle, that was now

  • https://cdm.link/amd-ilok-issu... [cdm.link] from TFA : " Musicians updating to AMD’s latest-and-greatest Ryzen 9000 Series / Ryzen AI 300 Series CPUs have found they’re unable to use PACE’s iLok License Manager and plug-ins that use its copy protection scheme. Fortunately, a fix from AMD is inbound that should resolve the issue."
    • ```
      Musicians ... found theyâ(TM)re unable to use PACEâ(TM)s iLok License Manager
      ```

      Wow, we never would have had Jimi Hendrix without spyware DRM.

  • Rather suspicious they suddenly have a firmware fix that improves performance for a newly-released CPU that was poorly received by the market. I guess they figured out they couldn't hold off on these gains until Zen 6 so decided to unlock them on Zen 5.
    • by Guspaz ( 556486 )

      More like this is just further evidence that Zen 5 was rushed to market before it was ready. AMD bungled this launch in just about every way possible.

      • by Targon ( 17348 )

        On the flip side, Intel 13th and 14th gen had that whole degradation issue, plus Big.little in 12th gen and newer has been screwing up performance and had to get fixed after release as well. Honestly, those on Zen4 wouldn't bother going to Zen5 anyway, even 20% performance isn't worth it for most people.

    • by Targon ( 17348 )

      There are two issues going on, the BIOS updates being talked about, and then you have the Windows issue where performance for Zen4 and Zen5 processors were lower than they should be(and fixed with recent Windows 11 insider edition builds). For the BIOS, much of the changes are about the TDP ratings on the Zen5 chips. Notice that except for the 9950X, all Zen5 chips have a lower TDP rating than the previous generation had. Much of this is probably due to a lot of talk about needing chips that don't tak

      • by higuita ( 129722 )

        Correct, people forget that lab conditions aren't the real conditions on all over the planet, so experience show that you need to be conservative when releasing as not everyone have a AC controlled environment and that too much power can lead to problems. If you already have the performance crown, why risk it! That is also why Intel had that problem, they were trying to catch up AMD and had to risk it more and push the power to gain more performance.

        After a CPU (or GPU) is released, they can finetune the po

      • It wouldnâ(TM)t be too surprising if AMDâ(TM)s data center focus plays a part as well. Very solid efficiency gains alongside tepid performance improvements did not impress the desktop(especially the sort of customers who might actually buy a CPU to upgrade a system, rather than just OEMs who will release new systems with whatever CPUs are current); but AMD shares CCDs between ryzen and epyc; and efficiency gains get a lot more interesting in the âoehow many cores can you cram into one socket
  • by slaker ( 53818 ) on Monday September 30, 2024 @01:11PM (#64828709)

    The vast majority of people whining about Zen 5 are whining about gaming performance. In fact, if you go look at tech reviews and Youtube content, you'll come away with the idea that it's a failed platform because it doesn't compare favorably with Zen 4 X3D systems, which have an ungodly amount of on-die cache purpose built to improve gaming performance.

    But I'm not particularly interested in playing games and I AM interested in having a fast video editing workstation. So here's an alternative take: My 9950X is around 33% faster than my previous editing system, a Threadripper 3960X, which has 50% more cores than the 9950X. It also has about half the TDP of the Threadripper and it also tolerates undervolting the CPU in a pretty substantial way, and all on a platform that cost less than my last motherboard.

    If you're coming from a Zen 4 system, yes, there's zero reason to upgrade, but why the fuck are you looking at upgrading from a two year old platform in the first place? Zen 5 is still a safe harbor for people using latter-day Intel, people upgrading from Zen 3 and those looking for a moderately future proof upgrade path.

    • Anyone who does care about gaming can just wait for the Zen 5X3D chips that will come out sometime in the next 3-6 months anyways and get a much better oranges to oranges comparison. Even if the existing Zen 5 CPUs offered a better upgrade what gamers would have switched as opposed to holding out for half a year to get an even better upgrade when the v-cache CPUs were released? Maybe someone cares about small bumps in performance enough to do that, but I suspect most people posting here have gaming rigs wit
    • by Slayer ( 6656 )

      Zen 5 is still a safe harbor for people using latter-day Intel, people upgrading from Zen 3 and those looking for a moderately future proof upgrade path.

      As a matter of fact, I've been looking upgrade options for my close to five year old Zen 3 Ryzen 5600X, and to be honest: I've decided to sit this one out. I'm supposed to toss out my DDR4, yet MicroATX motherboards for AM5 offer little else. It's still PCIe 4 on these, and for a potential 30% CPU speed boost I certainly won't yank out a working system. Intel would support my DDR4, but also wouldn't offer much of a performance improvement.

      This is not what Moore's law predicted, and certainly not worth my mo

If you fail to plan, plan to fail.

Working...